Closed Circuit Television, Third edition

Closed Circuit Television, Third edition
Closed Circuit Television
This page intentionally left blank
Closed Circuit Television
Third edition
Joe Cieszynski
IEng MIET Cert. Ed. LCGI
AMSTERDAM • BOSTON • HEIDELBERG • LONDON • NEW YORK • OXFORD
PARIS • SAN DIEGO • SAN FRANCISCO • SINGAPORE • SYDNEY • TOKYO
Newnes is an imprint of Elsevier
Newnes is an imprint of Elsevier
Linacre House, Jordan Hill, Oxford, OX2 8DP
30 Corporate Drive, Burlington, MA 01803
First edition 2001
Reprinted 2002
Second edition 2004
Reprinted 2004, 2005
Third edition 2007
Copyright © 2001, 2004, 2007, Joe Cieszynski. Published by Elsevier Ltd. All rights reserved
The right of Joe Cieszynski to be identified as the author of this work has been asserted in
accordance with the Copyright, Designs and Patents Act 1988
No part of this publication may be reproduced, stored in a retrieval system or transmitted
in any form or by any means electronic, mechanical, photocopying, recording or otherwise
without the prior written permission of the publisher
Permission may be sought directly from Elsevier’s Science & Technology Rights
Department in Oxford, UK: phone (44) (0) 1865 843830; fax (44) (0) 1865 853333; email:
[email protected] Alternatively you can submit your request online by visiting
the Elsevier web site at http://elsevier.com/locate/permissions, and selecting Obtaining
permission to use Elsevier material
Notice
No responsibility is assumed by the publisher for any injury and/or damage to persons
or property as a matter of products liability, negligence or otherwise, or from any use or
operation of any methods, products, instructions or ideas contained in the material herein.
Because of rapid advances in the medical sciences, in particular, independent verification
of diagnoses and drug dosages should be made
British Library Cataloguing in Publication Data
A catalogue record for this book is available from the British Library
Library of Congress Cataloging-in-Publication Data
A catalog record for this book is available from the Library of Congress
ISBN-13: 978-0-7506-8162-9
ISBN-10: 0-7506-8162-4
For information on all Newnes publications
visit our website www.books.elsevier.com
Typeset by Charon Tec Ltd (A Macmillan Company), Chennai, India
www.charontec.com
Printed and bound in Great Britain
07 08 09 10 11
10 9 8 7 6 5 4 3 2 1
Contents
Preface
Acknowledgements
ix
xi
1
The CCTV industry
1
The role of CCTV
The CCTV industry
2
4
Signal transmission
7
2
3
4
CCTV signals
Co-axial cable
Ground loops
Twisted pair cable
Structured cabling
Power over Ethernet
Ribbon cable
Fibre-optic cable
Infrared beam
Microwave link
UHF RF transmission
CCTV via the telephone network
Connectors
Cable test equipment
7
9
17
21
24
29
31
32
36
37
40
40
41
43
Light and lighting
47
Light and the human eye
Measuring light
Light characteristics
Artificial lighting
48
50
52
53
Lenses
59
Lens theory
Lens parameters
Zoom lenses
Electrical connections
Lens mounts
Filters
Lens adjustment
Lens finding
59
61
77
79
83
84
85
87
vi
Contents
5
Fundamentals of television
6
7
8
90
Producing a raster
Picture resolution
Synchronization
The luminance signal
The chrominance signal
Television signals
Digital video signals
Video compression
MPEG-2 compression
MPEG-4 compression
Wavelet compression
Common interchange format (CIF)
ITU-T recommendations
90
93
96
99
101
103
107
110
113
116
120
123
123
The CCTV camera
125
Charge coupled device
CCD chip operation
Electronic iris
IR filters
Colour imaging
Camera operation
White balance
Back light compensation
Colour/mono cameras
Camera sensitivity
Camera resolution
Camera operating voltages
Specialized cameras
Covert cameras
360° cameras
Number plate recognition cameras
125
126
133
133
134
137
141
142
142
144
145
146
147
148
150
152
Video display equipment
156
The cathode ray tube
The colour CRT
CRT monitors
Monitor safety
Liquid crystal displays (LCDs)
Plasma display panels (PDPs)
Projection systems
Termination switching
Resolution
Ergonomics
156
160
161
166
168
172
175
178
181
181
Video recording equipment
183
Digital video recorders (DVRs)
DVR principle
184
185
Contents
Effects of compression
Recording capacity
RAID disk recording
Digital video information extraction
VHS recording
Time-lapse recording
VCR maintenance
Video head cleaning
Tape management and care
Digital video tape
9 Camera switching and multiplexing
Sequential switching
Matrix switching
The quad splitter
Video multiplexers
Video motion detection (VMD)
10 Telemetry control
Control data transmission
Pan/tilt (P/T) control
Receiver unit
Dome systems
Data communications
11 CCTV over networks
Network topology
Network hardware
Network communications
IPv4 classes
Reserved addresses
Subnetting
Assigning IP addresses
Manually assigned IP addresses
Address resolution protocol (ARP)
Autoconfiguration
Domain name service (DNS)
Ports
Other network protocols
IPv6
Network diagnostics
CCTV over a network
Network CCTV example
Integrating analogue cameras
Summary
12 Ancillary equipment
Camera mountings
Towers and columns
vii
187
188
191
194
196
197
199
200
201
202
204
204
209
213
215
219
222
223
225
227
229
230
234
234
237
241
243
245
246
248
250
252
255
255
256
257
259
260
264
266
268
270
271
271
276
viii
Contents
Pan/tilt units
Monitor brackets
Power supplies
Voltage drop
13 Commissioning and maintenance
Commissioning
Measuring resolution
System handover
Preventative maintenance
Corrective maintenance
Fault location
Oscilloscope default settings
279
285
285
287
290
290
290
294
296
297
298
300
Glossary of CCTV terms
303
Index
321
Preface
In the preface to the first edition I wrote that closed circuit television (CCTV)
was a growth industry, the growth being very much a result of the impact
of new technology. As I write the pr eface to this thir d edition of Closed
Circuit Television, growth in the industry has continued, not only as a ersult
of technological advances that continue to bring clearer images, more intelligent systems and lower equipment costs, but also because of the heightened awareness of risk that is prevalent in Western society today. There is
a demand for everything from small, inexpensive systems to highly sophisticated systems covering many square miles.
And yet, like any high-technology installation, these systems will only
function correctly if they are properly specified, installed, commissioned
and maintained. Consequently, in addition to having an in-depth knowledge of CCTV principles and technology , the modern CCTV engineer is
expected to be conversant with electrical and electronics principles, the latest
digital and microprocessor principles, electrical installation practice, health
and safety regulations, and telecommunications and network technologies.
Clearly no single textbook could pr ovide a detailed coverage of all of
these subjects, and it is the aim of this book to concentrate on CCTV principles and technology in or der to provide the underpinning knowledge
required by CCTV practitioners. Like the first two editions befor e it, this
text will prove invaluable for those who are studying towards the City &
Guilds Knowledge of Security and Emer gency Alarm Systems (course
1852) and/or those who are working towards the NVQ level II or level III
in CCTV installation and maintenance. On the other hand, this book is
really intended for anyone who is involved with video signal pr ocessing
and transmission, which naturally includes those who ar e practising in
the industry and who wish to further their technical knowledge and
understanding, but also includes anyone who uses closed circuit television
for other applications such as surveying, medical, theatre production, etc.
As well as bringing the content of the second edition up to date, this
third edition includes much new material on subjects such as the most
recent (at the time of writing) video compr ession techniques, flat panel
display technologies and str uctured (CAT 5/6) cable principles. A complete new chapter has been included to help engineers grasp the principles of modern networks and ther efore have a better understanding of
how to specify, set up and troubleshoot network CCTV systems.
It is my continued hope and wish that trainees and engineers alike will
find this textbook a useful aid towards their personal development.
Joe Cieszynski
This page intentionally left blank
Acknowledgements
A text such as this would not be possible without the help and support of
people in the industry and, since scripting the first edition ofClosed Circuit
Television, I have been assisted by a number of people, many of whom are
specialists in their field. The people listed below have of fered both their
technical expertise and their time, for which I am very grateful.
Andrew Holmes of Data Compliance Ltd, with whom I have worked
on numerous occasions, is a constant sour ce of information. I am also
indebted to Simon Nash of Sony , and Martin Kane, who have on many
occasions provided me with information, help and guidance.
A thank you also to David Grant of ACT Meters and Gar Ning of NG
Systems. Although their services were not called upon during the writing
of this third edition, the marks that they made on the two pr evious editions remain.
I must also acknowledge the manufacturers who went out of their way
to provide photographs and information for use in this book. Such support
only helps in my quest to incr ease the level of knowledge and understanding of engineers in the industry . These manufacturers are acknowledged alongside their individual contributions.
A thank you to David Close for his sterling efforts in producing some of
the photographic work which is used to illustrate video compression, and
to Tim Morris of the University of Manchester for his much appr eciated
input into the video compression content in this book.
I would also like to thank my colleagues at P AC International Ltd for
their support. In particular , Graeme Ashcroft for his pr oofreading of a
number of portions of text, and Graham Morris and Steve Pilling who
both spent much time pr oofreading the network theory, providing much
appreciated feedback and suggested content.
As always, I am greatly indebted to my friend Ian Fowler for his input,
which spans all thr ee editions of this book. Once again he made himself
available to discuss aspects of theory and technology and gave a lot of
time to proofreading.
Finally, thanks again to David, Hannah, John and Ruth, my four (gr
own
up) children, for their patience and support during the writing of this edition, and to Linda, my terrific wife, for her continued support.
This page intentionally left blank
1 The CCTV industry
The term ‘closed circuit’ refers to the fact that the system is self-contained,
the signals only being accessible by equipment within the system. This is
in contrast to ‘broadcast television’, where the signals may be accessed by
anyone with the correct receiving equipment.
The initial development of television took place during the 1930s, and a
number of test transmissions took place in Europe and America. In the UK
these were from the Alexandra Palace transmitter in London. The outbreak of World War II brought an abrupt end to much of the television
development, although interestingly transmissions continued to be made
from occupied Paris using an experimental system operating fr om the
Eiffel Tower; The Nazi pr opaganda machine was very inter ested in this
new form of media.
Ironically the war was to give television the boost it needed in terms of
technology development because in the UK it seemed as if every scientist
who knew anything about radio transmission and signalling was pressed
into the accelerated development programme for radar and radio. Following
the war many of these men found themselves in gr eat demand from companies eager to renew the development of television.
Early black and white pictur es were of poor r esolution; however, the
success of the medium meant that the money became available to develop
new and better equipment, and to experiment with new ideas.
At the
same time the idea of using cameras and monitors as a means of monitoring an area began to take a hold but, owing to the high cost of equipment,
these early CCTV systems wer e restricted to specialized activity, and to
organizations that had the money to invest in such security. These systems
were of limited use because an operator had to be watching the scr een
constantly. There was no means of r ecording video images in the 1950s,
and motion detection connected to some form of alarm was the stuf f of
James Bond (only even he did not arrive until the 1960s!).
Throughout the 1960s and 1970s CCTV technology pr ogressed slowly,
following in the footsteps of the broadcast industry, which had the money
to finance new developments. The main stumbling block lay in the camera
technology, which depended completely on vacuum tubes as a pick-up
device. Tubes were large, required high voltages to operate, wer e generally
useless in low light conditions (although special types were developed – for
a price), and wer e expensive. Furthermor e, an early colour camera
required three of these tubes. For this r eason, for many years CCTV
remained on the whole a low-resolution, monochrome system which was
very expensive.
2
Closed Circuit Television
By the 1980s camera technology was improving, and the cost of a reasonable colour camera fell to a sum that was affordable to smaller businesses
and organizations. Also, VHS had arrived. This had a serious impact on the
CCTV industry because for the first time it was possible to r ecord video
images on equipment that cost well below £1000. For a number of years
prior to this, CCTV could be recorded on monochrome reel-to-reel machines,
but these were expensive and were not exactly user-friendly.
From the mid-1980s onwards television technology advanced in quantum
leaps. New developments such as the CMOS microchip and charge-coupled
device (CCD) chip brought about an increase in equipment capability and
greatly improved picture quality, whilst at the same time equipment
prices plummeted. Manufacturers such as Panasonic and Sony developed
digital video recording machines, and although these wer e intended primarily for use in the br oadcast industry (at £50 000 for a basic model the
CCTV industry was not in a hurry to include one with every installation!),
these paved the way for digital video signal pr ocessing in lower-resolution
CCTV and domestic video products.
For many years, CCTV had to r ely on its big br other – the br oadcast
industry – to develop new technologies, and then wait for these technologies to be downgraded so that they became af fordable to customers
who could not afford to pay £30 000 per camera and £1000 per monitor .
However, the technology explosion that we are currently seeing is changing this. PC technology is rapidly changing our traditional ideas of viewing and r ecording video and sound, and much of this har
dware is
inexpensive. Also, whereas in the early years the CCTV industry r elied
largely on the traditional broadcast and domestic television equipment
manufacturers to design the equipment, there are now a large number of
established manufacturers that are dedicated to CCTV equipment development and production. These manufacturers are already taking both concepts and hardware from other electronics industries and integrating them
to develop CCTV equipment that not only produces high quality images,
but is versatile, allows easy system expansion, is user friendly, and can be
controlled from anywhere on the planet without having to sacrifice one of its
most valuable assets – which is that it is a closed circuit system.
The role of CCTV
So often CCTV is seen as a security tool. Well of course it is; however, it plays
equally important roles in the areas of monitoring and control. For example,
motorway camera systems are invaluable for monitoring the flow of traffic, enabling police, motoring or ganizations and local radio to be used
to warn drivers of pr oblems, and thus contr ol situations. And yet in the
case of a police chase, contr ol room operators can assist the police in
directing their r esources. This same versatility applies to town centr e
CCTV systems.
The CCTV industry
3
CCTV has become an invaluable tool for or ganizations involved in
anything to do with security, crowd control, traffic control, etc. Yet on the
other hand the pr oliferation of cameras in every public place is ringing
alarm bells among those who ar e mindful of Geor ge Orwell’s book
Nineteen Eighty-Four. Indeed, in the wr ong hands, or in the hands of the
sort of police state depicted in that book, CCTV could be used for all manner
of subversive activity. In fact the latest technology has gone beyond the predictions of Mr Orwell. Face recognition systems, which generate an alarm as
soon as it appears in a camera view , have been developed, as have systems
that track a person automatically once they have been detected. Other equipment which can see through a disguise by using parameters that make up a
human, such as scull dimensions and r elative positions of extreme features
(nose, ears, etc.), or the way that a person walks, is likewise under development. At the time of writing all such systems are still somewhat experimental and are by no means perfected, but with the current rate of technological
advancement we can only be a few years away from this equipment being
installed as standard in systems in town centres, department stores, night
clubs and anywhere else where the authorities would like early r ecognition of ‘undesirables’.
To help control the use of CCTV in the UK, the changes made to the
Data Protection Act (DPA) in 1998 meant that images from CCTV systems
were now included. Unlike the earlier 1984 Act, this had serious implications for the owners of CCTV systems as it made them legally r esponsible
for the management, operation and contr ol of the system and, per haps
more importantly, the recorded material or ‘data’ pr oduced by their system. The Data Pr otection Act 1998 requires that all non-domestic CCTV
systems are registered with the Information Commissioner . Clear signs
must be erected in areas covered by CCTV warning people that they ar e
being monitored and/or recorded. The signs must state the name of the
‘data controller’ for the system, and have contact details. When er gistering
a system, the data controller must state its specific uses and the length of
time that material will be retained. Recorded material must be stored in a
secure fashion and must not be passed into the public domain unless it is
deemed to be in the public inter est or in the inter ests of criminal investigations (i.e., the display of images on police-orientated programmes).
In 2004 the Information Commissioner ’s Office published a r evised
document in the light of a court case where the definition of the ‘information relating to an individual’ was challenged. Although the case did not
directly involve CCTV ‘information’, nevertheless there were implications
for smaller CCTV systems in the UK. The document advised that some
smaller CCTV systems are not covered by the DPA because the information contained in their recordings cannot be considered to relate to an individual. By definition, if the cameras are fixed (i.e., no PTZ capability), are
not used to monitor staf f members to observe their behaviour , and
recorded information is only passed to a law enfor cement body such as
the police, then the system does not have to be registered under the DPA.
4
Closed Circuit Television
On 2 October 1998 the Human Rights
Act became ef fective in the
United Kingdom. The emphasis on the right to privacy (among other
things) has strong implications for CCTV used by ‘public authorities’ as
defined by the Act and system designers and installers should take note of
these implications. Cameras that ar e capable of targeting private dwellings
or grounds (even if that is not their real intention) may be found to be in contravention of the rights of the people living there. As such, those people may
take legal action to have the cameras disabled or r emoved – an expensive
undertaking for the owner or, perhaps, the installing company who specified
the camera system and/or locations.
In relation to CCTV, the intention of both the Data Pr otection and
Human Rights Acts is to ensur e that CCTV is itself pr operly managed,
monitored and policed, thus pr otecting against it becoming a law unto
itself in the future.
The arguments surrounding the uses and abuses of CCTV will no doubt
continue; however, it is a well-proven fact that CCTV has made a huge, positive impact on the lives of people who live under its watchful eye. It has
been proven time and again that both people and their possessions are more
secure where CCTV is in operation, that people ar e much safer in cr owded
public places because the crowd can be better monitored and controlled, and
that possessions and premises are more secure because they can be watched
24 hours per day.
The CCTV industry
Despite what we have said about CCTV being used for operations other
than security, it can never fully escape its potential for security applications because, whatever its intended use, if the police or any other public
security organization suspect that vital evidence may have been captured
on a video recording system, they will inspect the recorded material. This
applies all the way down to a member of the public who, whilst innocently using a camcorder or a video recorder on a mobile phone, happens
to capture either an incident or something relating to an incident. For this
reason it is per haps not surprising to hear that the CCTV industry is
largely regulated and monitored by the same people and or ganizations
that monitor the security industry as a whole.
The British Security Industry Association (BSIA) Ltd is the only UK
trade association for the security industry that r equires its members to
undergo independent inspection to ensur e they meet relevant standards.
The BSIA’s primary role is to pr omote and encourage high standar ds of
products and services throughout the industry for the benefit of customers.
This includes working with its members to produce codes of practice, which
regularly go on to become full British/Eur opean standards. The BSIA also
lobbies government on legislation that may impact on the industry and
actively liaises with other r elevant organizations, for example the Office of
The CCTV industry
5
the Information Commissioner (in r elation to the Data Pr otection Act) and
the Home Office Scientific Development Branch (HOSDB). The BSIA also
provides an invaluable service in producing technical literature and training
materials for its members and their customers.
Inspectorate bodies are charged with the role of policing the installation
companies, making sure that they are conforming to the Codes of Practice.
Of course, a company has to agr ee to place itself under the canopy of an
Inspectorate, but in doing so it is able to advertise this fact, and gives it
immediate recognition with insurance companies and police authorities.
To become an approved installer a company must submit to a rigorous
inspection by its elected Inspectorate. This inspection includes not only
the quality of the physical installation, but every part of the organization.
Typically the inspector will wish to see how documentation r elating to
every stage of an installation is pr ocessed and stored, how maintenance
and service records are kept, how material and equipment is ordered, etc.
In addition the inspector will wish to see evidence that the or ganization
has sufficient personnel, vehicles and equipment to meet maintenance
requirements and breakdown response times.
In some cases the or ganization is expected to obtain BS EN ISO 9002
quality assurance (QA) accr editation within two years of becoming an
approved installer. At the time of writing there is no specific requirement
that engineers working for an appr oved installation company hold a
National Vocational Qualification (NVQ) in security and emergency systems
engineering; however, this may well become the case in the future.
Another significant body is Skills for Security , the Standar ds Setting
Body for the security business sector. Skills for Security incorporates many
of the functions formerly undertaken by SITO (Security Industry Training
Organization) as well as adopting a wider r emit similar to Sector Skills
Councils. In the UK, SITO were responsible for the development of training
standards for the security industry and did much to raise those standards
throughout the 1990s. They developed the NVQ levels II and III for electronic security systems, plus many other awards covering all sectors of the
security industry. Skills for Security came into being in January 2005 and
work closely with the industry to identify the training needs (both present
and future) and develop pr ogrammes and qualifications that will meet
these needs.
Awarding bodies such as City & Guilds and Edexcel play an important
role in the security industry because it is they who devise the course syllabus and assessment criteria for the training and education of personnel
working in the industry. The UK qualification for CCTV engineers is the
City & Guilds NVQ level II or level III in Security and Emer gency Alarm
Systems. The City & Guilds also of fer the underpinning knowledge test
papers (course 1852) for the four disciplines relating to security and emergency systems engineering, these being CCTV, intruder alarm, access control and fir e alarm systems. These awar ds are intended to contribute
towards the underpinning knowledge testing for the NVQ level III award,
6
Closed Circuit Television
although a candidate may elect to sit these tests without pursuing an
NVQ. It must be stressed, however, that the 1852 award is not an alternative
qualification to an NVQ, and a person holding only the 1852 certificates
would not be deemed to be qualified until they had proven their competence
in security systems engineering.
The awarding bodies appoint external verifiers whose role it is to check
that NVQ assessment centres, be these colleges, training organizations or
installation companies, are carrying out the assessments to the recognized
standards.
The Home Office Scientific Development Branch (formerly the Police
Scientific Development Branch – PSDB) plays a most significant r ole in
CCTV. For many years the CCTV industry had no set means of measuring the
performance of its systems in terms of picture quality, resolution and the size
of images as they appear on a monitor scr een. This meant that in the
absence of any benchmarks to work to, each surveyor or installer would
simply do what they considered best. This situation was not only unsatisfactory for the industry; potential customers were in a position where they
had no way of knowing what they could expect fr om a system and, once
it was installed, had no r eal redress if they wer e unhappy, because there
was nothing for them to measure the system performance against.
The PSDB set about devising practical methods of defining and measuring
such things as pictur e resolution and image size and, for example, in 1989
introduced the Rotakin method of testing the resolution and size of displayed
images (see Chapter 13). They also developed methods of analysing and
documenting the needs of customers prior to designing a CCTV system. This
is known as an Operational Requirement (OR). HOSDB continue this work,
providing much practical guidance on issues r elating to the latest CCTV
technologies such as watermarking of r ecorded video images, methods of
archive retrieval, measurement of resolution of digital images, etc.
CCTV is a gr owth industry. It has pr oven its ef fectiveness beyond all
doubt, and the availability of high-quality, versatile equipment at a relatively
low cost has resulted in a huge demand for systems of all sizes. W ithin the
industry there is a genuine need for engineers who tr uly understand the
technology they are dealing with, and who have the level of underpinning
knowledge in both CCTV and electronics principles that will enable them to
learn and understand new technologies as they appear.
2 Signal transmission
A CCTV video signal contains a wide range of a.c. components with
frequencies varying fr om 0 Hz up to anything in the or der of 10 MHz.
Furthermore, in addition to the a.c. components ther e is also an essential
d.c. component which must be preserved throughout the signal transmission process if accurate brightness levels ar e to be maintained. Pr oblems
occur when engineers consider video signal transmission in the same
terms as transmitting low-voltage d.c. or low-fr equency mains voltage.
When you consider that domestic medium wave radio is transmitted
around 1 MHz, then it becomes clear that the 0–10 MHz video signal is
actually going to behave in a similar manner to radio signals.
In this chapter we shall examine the peculiar behaviour of highfrequency signals when they are passed along various types of cables, and
therefore explain the need for special cables when transmitting video signals, and the reasons for the limitations in each transmission medium.
CCTV signals
An electronically produced square wave signal is actually built up from a
sinusoidal wave (known as the fundamental) and an infinite number of
odd harmonics (odd multiples of the fundamental fr equency). This basic
idea is illustrated in Figure 2.1 where it can be seen that the addition of just
Fundamental
Steeper
sides
Third harmonic
Flatter top
Resultant
Figure 2.1 Effect of the addition of odd harmonics to a sinusoidal waveshape
8
Closed Circuit Television
the third harmonic component changes the appearance of the fundamental sine wave, moving it towar ds a squar e shape. Adding the fifth harmonic would have the effect of steepening the sides and flattening the top.
In other words the waveshape becomes mor e square. Taking this to the
extreme, adding an infinite number of odd harmonics would pr oduce a
waveshape that has perfectly vertical sides and a perfectly flat top.
If we reverse this process, i.e., begin with a squar e wave and r emove
some of the harmonic components using filters, then the corners of the
square wave become rounded, and the rise time becomes longer. In other
words, the squar e wave begins to r eturn to its sinusoidal fundamental.
This effect is illustrated in Figure 2.2.
If signal path
hf signal path
Low pass filter
Figure 2.2 Removal of high-frequency harmonic components increases the rise
time and rounds the corners
In Chapter 5 we shall be looking at the make-up of the video signal
(Figure 5.13), and we will see that it contains square wave components. It
is the sharp rise times and right-angled corners in the video signal waveform
which produce the high-definition edges and high-resolution areas of the picture.
If for any reason the signal is subjected to a filtering action resulting in the
loss of harmonics, the r eproduced picture will be of poor r esolution and
may have a smeared appearance. Now one may wonder how a video signal could be ‘accidentally’ filtered, and yet it actually occurs all of the time
because all cables contain elements of r esistance, capacitance and inductance, the three most commonly used components in the constr uction of
electronic filter circuits. When a signal is passed along a length of cable it
is exposed to the effects of these R, C, L components.
The actual effect the cable has on a signal is dependent on a number of
factors which include the type and construction of cable, the cable length,
the way in which bends have been formed, the type and quality of connectors and the range of fr equencies (bandwidth) contained within the
signal. This means that, with respect to CCTV installations, it is important
that correct cable types are used, that the correct connectors are used for a
given cable type, that the cable is installed in the correct specified manner
and that maximum run lengths are not exceeded without suitable means
of compensation for signal loss.
Signal transmission
9
Different cable types are used for the transmission of CCTV video signals,
and indeed methods other than copper cable transmission ar e employed.
Both the surveyor and the installing engineer need to be aware of the performance and limitations of the various transmission media, as well as the
installation methods that must be employed for each medium.
Co-axial cable
As stated earlier, the behaviour of high-frequency signals in a copper conductor is not same as that of d.c. or low fr equencies such as 50/60 Hz
mains, or audio, and specially constr ucted cables are required to ensure
constant impedance acr oss a range of fr equencies. Furthermore, radiofrequency signals have a tendency to see every copper conductor as a
potential receiving aerial, meaning that a conductor carrying an RF signal
is prone to picking up stray RF fr om any number of sources; for example
emissions from such things as electric motors, fluor escent lights, etc., or
even legitimate radio transmissions. Co-axial cable is designed to meet the
unique propagation requirements of radio-fr equency signals, of fering a
reasonably constant impedance over a range of frequencies and some protection against unwanted noise pick-up.
There are many types of co-axial cable, all manifesting different figures
for signal loss, impedance, screening capability and cost. The construction
of a co-axial cable determines the characteristics for a particular cable type,
the basic physical construction being illustrated in Figure 2.3.
Inner insulating sleeve
Copper core
Copper
braid
Insulating outer sleeve
Figure 2.3 Co-axial cable construction
The signal-carrying conductor is the copper central core, which may be
a solid copper conductor or stranded wire. The signal return path could be
considered to be along the braided screen; however, as this is connected to
the earth of a system, the signal may in practice return to its source via any
10
Closed Circuit Television
number of paths. However , the scr een plays a far mor e important r ole
than simply to serve as a signal return path. It provides protection against
radio-frequency interference (RFI). The way that it achieves this is illustrated
in Figure 2.4, where it can be seen that external RF sour ces in close proximity of the cable ar e attracted to the copper braided scr een, from where
they pass to earth via the equipment at either end of the cable. Pr ovided
that the integrity of the screen is maintained at every point along the cable
run from the camera to the monitor, there is no way that unwanted RF signals can enter either the inner cor e of the co-axial cable or the signal pr ocessing circuits in the equipment, which will themselves be scr eened,
usually by the metal equipment casing.
Radio transmitters
RFI
Electric motors
Car ignition
Striplights
Figure 2.4 RFI is contained by the copper screen, preventing it from entering
the signal processing circuits
Integrity of the screen is maintained by ensuring that there are no breaks
in the screen at any point along the cable length, and that all connectors are
of the correct type for the cable and have been fitted correctly. We shall consider connectors later in this chapter, but the issue of breaks in the screen is
one which we need to consider. Co-axial cable is more than a simple piece of
wire, and only functions correctly when certain criteria have been met in
relation to terminations and joints. Under no circumstances should a joint
be made by simply twisting a pair of cor es together and taping them up
before twisting and taping the two screens. Although this might appear to
be electrically sound, it breaks all the rules of RF theory and, among other
things, can alter the dynamic impedance and expose the inner core to RFI.
All joins should be made using correctly a fitted connector (usually BNC)
on each cable end, with a coupling piece inserted in between.
Where RFI is pr esent in a video signal, it usually manifests itself as a
faint, moving patterning ef fect superimposed onto the pictur e. The size
Signal transmission
11
and speed of movement of the pattern depends on the fr equency of the
interfering signal.
The inner sleeve of the co-axial cable performs a much more important
function than simply insulation between the two conductors: it forms a
dielectric between the conductors which intr oduces a capacitive element
into the cable. This cable capacitance works in conjunction with the natural
d.c. resistance and cable inductance to pr oduce a characteristic impedance
(Zo) for the cable. One of the factors which governs the value of a capacitor
is the type of dielectric (insulator) used between the plates, and co-axial
cables of differing impedances are produced by using different materials
for the inner core. This is why not all co-axial cables are suitable for CCTV
applications, and why a connector designed for one cable type will not fit
onto certain other types; the cable diameter varies depending upon the
dielectric. The equivalent circuit of a co-axial cable is shown in Figure 2.5.
Figure 2.5 Equivalent circuit of a co-axial cable, also known as a transmission line
The characteristic impedance for a cable of infinite length can be found
from the equation Zo L/C. However, this concept is somewhat theoretical as we do not have cables of infinite length. On the other hand, for
a co-axial cable to function as a transmission line with minimum signal
loss and r eflection (we will look at this is a moment), the termination
impedance at both ends must equal the calculated characteristic impedance for an infinite length. Thus, if the characteristic impedance, Zo, for a
cable is quoted as being 75, then the equipment at both ends of the cable
must have a termination impedance of 75 .
If this is not the case a number of problems can occur. First of all signal
loss may be apparent because of power losses in the transfer both to and
from the cable. It can be shown that for maximum power transfer to occur
between two electrical cir cuits, the output impedance of the first cir cuit
must be equal to the input impedance of the second (Figur e 2.6). If this is
not the case, some power loss will occur. In our case the co-axial cable can
be considered to be an electrical unit, and this is why all equipment connected to the cable must have a matching impedance.
Another problem associated with incorr ect termination is one of
reflected waves. Where a cable is not terminated at its characteristic impedance, not all of the energy sent down the line is absorbed by the load, and
12
Closed Circuit Television
Unit A
Zout
Unit B
Zin
Figure 2.6 Maximum power transfer only occurs when Zout in Unit A is equal to
Zin in Unit B. (Assume that the connecting cables have zero impedance)
because the unabsorbed energy must go somewhere, it travels back along
the line towards its source. We now have a situation where there are two signals in the cable, the forward wave and the reflected wave. In CCTV, reflected
waves can cause ghosting, pictur e roll, and loss of telemetry signals . However,
these symptoms may not be consistent and may alter sporadically, leaving
the unsuspecting service engineer chasing from one end of the installation
to the other looking for what appears to be a number of shifting faults –
and perhaps for no other reason than because a careless installation engineer
has made a Sellotape-style cable connection in a roof space!
CCTV equipment is designed to have 75 input and output impedances. This
means that 75 co-axial cable must always be used. Her e again the
installing engineer must be awar e that not all co-axial cable has 75 impedance, and 50 and 300 versions are common. For example, cable
type RG-59 is a common 75 co-axial cable used in CCTV installations.
Cable type RG-58 looks very similar, but it is designed for different applications and has a characteristic impedance of 50 . A CCTV installation
using this cable would never perform to its optimum capability, if indeed
it were able to perform at all.
Termination switches are included in CCTV equipment to ensur e that
there is a 75 impedance at both ends of any co-axial cable network. This
topic will be discussed in more detail in Chapter 7.
Up to now we have not taken into consideration the length of the
co-axial cable. Over short distances the effects of C and R on the signal are
small and can be ignor ed. However, as the cable length is incr eased, these
components have an effect on the signal which is similar to a voltage dr op
along a d.c. supply cable, the main difference being that the filtering action of
the cable results in greater losses at the higher signal frequencies. Figure 2.7
illustrates a typical co-axial cable fr equency response. Cable losses ar e
usually quoted in terms of dB per 100 m, at a given fr equency. Manufacturers may quote figures for a range of frequencies; however, those quoted
for around 5 MHz are the most significant to the CCTV engineer because,
Signal transmission
13
as seen from Figure 2.7, it is at the top end of the video signal fr equency
spectrum where the most significant losses occur.
Signal loss per 100 m
0 dB
1
2
3
4
0
1
2
3
4
5
6
f MHz
Figure 2.7 Frequency losses in a typical co-axial cable
Every cable employed in CCTV signal transmission has a specified
maximum length, beyond which optimum system performance will only
be maintained if additional equipment is installed. T ypical specifications
for the three most common co-axial cables employed in the CCTV industry are given in Table 2.1. The figures quoted for the maximum cable r un
length are those quoted in the BSIA
Code Of Practice for Planning,
Installation and Maintenance of CCTV Systems, document 109 issue 2
1991, and some variance with these figur es may be noted when comparing different manufacturers’ data; however, the installer will do well to
heed the guidelines laid down in the BSIA document.
Table 2.1 Co-axial cables commonly employed in the CCTV industry
Cable type
Max run length
Impedance
Loss/100 m at 5 MHz
URM-70
RG-59
RG-11
250 m
350 m
700 m
75 75 75 3.31 dB
2.25 dB
1.4 dB
As a rule, monochrome signals tend to cope better with long cable runs
than do colour signals. This is because a PAL colour signal contains a highfrequency 4.43 MHz colour subcarrier which is af fected by the filtering
action of the cable. However , even a monochr ome signal contains h.f.
14
Closed Circuit Television
components and, bearing in mind the ef fects of h.f. filtering on a squar e
wave (Figure 2.2), will therefore suffer some loss of resolution and signal
level where the cable run length is excessive.
To illustrate the pr oblem of signal loss, consider the cable illustrated
in Figure 2.7. At 3 MHz the loss per 100 m is approximately 1 dB. Thus,
over a distance of 350m the loss will be in the order of 3.5 dB. In terms of
voltage, assuming that a standard 1 Vpp video signal was injected into the
cable, 3.5 dB represents an output voltage at the end of the cable of
around 0.7 Vpp; a signal loss of 0.3 V. At 5 MHz the loss is in the or der
of 1.75 dB per 100 m; therefore over 350 m the loss in dBs will be approximately 6 dB. Thus, it can be shown that at 5 MHz the signal output will
be approximately 0.5 V. Now consider what would happen if an installer
were to ignore these figures and fit a 700m length of this cable. The output
figures become 0.45 V at 3 MHz, and 0.25 V at 5 MHz. At best such a signal
will produce a low-contrast picture, more than likely with a loss of colour,
and perhaps with picture roll due to the loss of sync pulses.
Where runs in excess of the maximum specified length for a particular
cable are unavoidable, launch amplifiers and/or cable equalizers can be
installed. The use of these can at least double the length of a cable run.
A launch amplifier is usually installed at the camera end of the cable
where there is an available source of power, although there is a sound argument for installing it half way along a length of cable if a means of supplying power can be found. A typical launch amplifier r esponse is shown in
Figure 2.8 where it can be seen that the level of amplification is not uniform
across the 0–5.5 MHz video signal bandwidth. The amplifier is designed to
give extra lift to the higher frequencies where the greater losses occur.
Signal
level
Launch amp response
Cable response
f (MHz)
0
1
2
3
4
5
6
Figure 2.8 A launch amplifier compensates for the filter action of the cable
The amplifier usually has an adjustment to allow the gain to be set to
suit the length of cable; the longer the cable the higher the gain setting.
The idea is to set the output voltage level such that, after losses, a uniform
Signal transmission
15
1 Vpp signal appears at the other end. In some cases the gain control is calibrated in cable lengths, and it is ther efore necessary to have an appr oximate idea of the length of the r un. Do not simply turn the contr ol until a
‘good, strong picture’ appears on the monitor . This practice can lead to
problems in relation to vertical hold stability wher e switchers or multiplexers are involved, and possibly a loss of picture resolution.
A cable equalizer is a form of amplifier, but it is designed to be installed at
the output end of the cable (Figure 2.9). The problem with this is that the unit
is having to pr ocess the signal once the losses have been incurr ed, and in
boosting the signal levels it will also boost the background noise level, which
will have risen in the absence of a str ong signal. The advantage of using a
cable equalizer is that it can be installed in the control room, which can be a
real plus in cases where the camera is inaccessible. If the installer has a choice
of which to use, a launch amplifier is usually preferable as it lifts the signal
before losses occur, thus maintaining a better signal-to-noise ratio.
Camera
Launch amplifier
Monitor
Camera
Cable equalizer
Monitor
Figure 2.9 Use of launch amplifiers and cable equalizers
It is possible to employ mor e than one amplifier in cases wher e very
lengthy cable runs are required. The idea is that these ar e placed at even
distances along the cable such that, just as the signal would begin to deteriorate, another amplifier lifts it once again. This principle is shown in
Figure 2.10 wher e it can be seen that the total cable loss is
33.75 dB,
which is compensated for by the overall gain in the system of 36 dB.
All this sounds well and good, but it takes a highly experienced engineer with the correct equipment to be able to adjust the gain and r esponse
of all of these units to a point wher e a perfect, uniform 1 Vpp, 0–5.5 MHz
video signal is obtained at the other end without any incr ease in noise
level. And remember, once noise has been intr oduced into the signal, it
will simply be boosted along with the signal in each subsequent amplifier.
16
Closed Circuit Television
RG-59 cable loss = 2.25 dB/100 m
500 m
11.25 dB
Camera
Launch amp 1
1000 m
22.5 dB
Launch amp 2
Cable
equalizer
Monitor
Figure 2.10 Launch amp 1 gain 12 dB. This compensates for the first 500 m
of cable. Launch amp 2 gain 12 dB. This compensates for 50% of the losses in
the following l km of cable. Cable equalizer gain 12 dB correcting for losses in
the 1 km cable run
Still on the subject of losses, it should be noted that every BNC (or other
type) connector introduces an element of signal attenuation and reflection,
and it is good practice to keep the number of joins in a cable to a minimum.
In the UK, all CCTV signal cable installation should comply with current codes
of practice as laid down in BS 7671; Requir ements for Electrical Installations ,
especially in r elation to electrical segr egation of low- and high-voltage
cables. However, apart from the electrical safety issues surr ounding segregation, installers should pay particular attention to the pr oximity of
co-axial cables with mains power cables, in particular those carrying a
high current, or supplying lar ge numbers of fluor escent lights, heavy
machinery, etc. Any current-carrying conductor produces an electromagnetic field around its length. Furthermore, high-frequency spikes passing
along a cable can pr oduce large electromagnetic pulses (EMP). Therefore
it follows that both of these energy fields must surround all mains supply
cables, because they are carrying high-frequency noise spikes in addition
to the high current 50 Hz mains supply. Where co-axial cables are run parallel to mains cables, there is a good chance of the electromagnetic interference (EMI) penetrating the screen and superimposing a noise signal onto
the video signal. Where this occurs, the displayed or recorded picture will
suffer such ef fects as horizontal ripples r olling up or down, or random
flashes when lights are switched or machinery operated.
Naturally the co-axial scr een provides much pr otection against such
noise ingression, but at best the scr een will be no mor e than 95% ef fective; and many cables may have a much lower figur e. To prevent noise
ingression it is good practice to avoid long, close-pr
oximity, parallel
co-axial/mains supply cable runs wherever possible, maintaining at least
30 cm (12) between cables. This may r ule out using plastic segr egated
trunking because, although it offers electrical segregation, it does nothing
to prevent the problems we have just outlined. Metal tr unking provides
screening against interfer ence, and in cases wher e co-axial video cable
must run through areas of high electrical noise, it is good practice to use
Signal transmission
17
steel trunking or conduit to minimize the chances of EMI compr omising
system performance.
Having looked at the constr uction of co-axial cable we know that the
characteristic impedance depends, among other things, upon the capacitance of the cable, which is determined by the type and thickness of the
inner insulating material. Therefore, should the inner sleeve become damaged by the cable being cr ushed, kinked, or filled with water, the characteristic impedance will alter, opening the system up to the inherent problems
of signal loss and r eflected waves. Putting this another way , installers
should take care not to damage the cable during installation, and should
not lay cables in places where they may easily be damaged at a later time.
BNC connectors are not waterproof and were never intended for external
use. Therefore where external connections ar e necessary, they should
always be enclosed in a weatherpr oof housing. Once water enters a coaxial cable the capillary action may allow it to travel many metr es along
the cable, introducing all manner of undesirable pictur e effects, and very
often these can be intermittent.
In order to prevent damage to the inner sleeve, co-axial cable should
not have any severe bends. A rule of thumb is to ensure that the radius of all
bends is no tighter than five times the diameter of the cable. For example, if the
cable diameter is 6.5 mm, the radius of a bend should be at least 32.5 mm.
Ground loops
These occur when the earth (voltage) potential dif fers across the site.
Because every item of mains power ed equipment must be connected to
earth, where the earth potentials dif fer, an a.c. 50 Hz current will flow
through the low-impedance screen. The problem is illustrated in Figure 2.11
where a length of co-axial cable has a potential dif
ference of 40 V
between its ends. It naturally follows that a curr ent will flow through the
low-impedance co-axial scr een which is bypassing the much higher
impedance of the ground, which was the cause of the potential difference
in the first place.
Differing ground potentials are very common, especially over long distances, and the problem can be further compounded when equipment at one
end of a cable is connected to a different phase of the mains supply than that
at the other end. The example in Figur e 2.11 indicates a potential difference
of 40 V; however, a difference of just 2–3 V is sufficient to cause problems.
When a gr ound loop curr ent flows along a co-axial cable scr een,
because the centre core is referenced to the screen, a 50 Hz ripple is superimposed onto the video signal. This means that the brightness levels in the
signal information are constantly moving at a rate of 50Hz, and the effect on
the monitor display is either a dark shadow or a ripple r olling vertically
through the picture. This effect, known often as a hum bar, can also upset
the synchronizing pulses, resulting in vertical picture roll.
18
Closed Circuit Television
Co-axial screen earthed
at both ends
30 V
10 V
Earth potentials
Figure 2.11a A CCTV system where earth potentials differ
50 Hz a.c. current
30 V
Z
10 V
Earth potentials
Figure 2.11b Equivalent electrical circuit. The high-impedance earth path (Z) is
by-passed by the low impedance of the co-axial screen
It is possible to test for an earth potential pr oblem during installation
by taking an a.c. voltage measurement between the co-axial screen and the
earth of the equipment to which it is to be connected. Under perfect earthing conditions, the reading should be 0 V. In practice it is usual to obtain a
reading of at least a few hundred millivolts; however, in severe conditions
potentials of 50 V or even greater are possible. In such cases it is not safe to
assume that the problem is simply caused by differences in earth potential
as there might actually be a serious fault in the earth circuit of the electrical
supply, and if the CCTV installer himself is not a qualified electrician, he
should report the potential fault to the appropriate persons, in writing, in
order that a full inspection of the supply can be carried out.
Signal transmission
19
There are various ways of avoiding or over coming the pr oblem of
ground loops in a CCTV system. Avoidance is always the best policy, but
is not always practical. Remember that gr ound loops occur because the
system has more than one earth point, and these are at differing potentials.
Therefore if 12 Vd.c. or 24 Va.c. cameras can be used, the only earth connection to the co-axial cable is at the contr ol room end, and ground loops
will not occur. This principle is illustrated in Figure 2.12. Other methods of
avoidance are to employ twisted pair or fibre-optic cables, which we shall
be looking at later in this chapter . However, fibre-optic cables are more
expensive to install, and besides, the pr oblem may not be identified until
after co-axial cables have already been installed.
Monitor
Camera power
supply
24 Va.c.
N
P
230 Va.c. mains
outlet board
Figure 2.12 In a low-voltage camera supply, the co-axial cable is only earthed at
the monitor end
Ground loop corr ection equipment is available. Ther e are two types:
transformer and optical. T ransformer types ar e usually contained in a
sealed metal enclosure which acts as scr eening. In order to provide ideal
coupling of the broadband video signal, the internal circuits may contain
more than just a transformer . Nevertheless, the principle behind these
units is to break the co-axial cable earth circuit but still provide video signal transmission without af fecting the integrity of the cable scr een. The
basic circuit operation is shown in Figur e 2.13. In practice a single unit
may contain two transformers, allowing two separate video circuits to be
corrected. The unit can be installed at either end of the cable, although it is
usually more convenient to locate it at the control room end.
It is worth noting that not all correction transformers perform to the same
standard when it comes to broadband video signal coupling, and sometimes
20
Closed Circuit Television
Ground loop
correction
transformer
30 V
10 V
Figure 2.13 Inclusion of a transformer breaks the 50 Hz current path through the
co-axial screen
a loss of resolution may be evident. Furthermor e, where a transformer is
not capable of coupling high frequencies, this can pose problems for certain
types of telemetry contr ol signal, resulting in a loss of telemetry to cameras which have a gr ound loop correction transformer included. As with
any type of CCTV equipment, car eful selection is important, and when
you have found a product which performs satisfactorily, stay with it.
Optical correctors rely on opto-couplers to br eak the co-axial scr een
(Figure 2.14). The video signal is applied to a light-emitting diode which
converts the varying voltage levels in the video signal into variations in
light level. These in turn are picked up by a photodiode which converts the
light signal back into a variable voltage. Units containing a number of individual inputs (typically 8 or 16) are available, and can be included with the
Optical isolator
Figure 2.14 Principle of a single channel opto-isolator
Signal transmission
21
control room equipment, acting as a buffer for each camera input. These
are ideal for installations where it is anticipated at the planning stage that
ground loops may pose a pr oblem because it is known that cameras will
either be connected acr oss different phases of the mains supply , or will
span a lar ge geographical ar ea. A multiple input gr ound loop corr ector
can be included in the initial quotation, thereby removing the problems of
additional costs once the installation is underway.
Twisted pair cable
As the name implies, this cable comprises two cor es which ar e twisted
around each other. The number of twists per metre varies depending upon
the quality of the cable, but a minimum of 10 turns per metr e is recommended for CCTV video signal transmission applications – the more turns
there are, the better the quality of the cable in terms of noise rejection.
This type of cable pr ovides balanced signal transmission (as opposed
to unbalanced, which is how co-axial cable functions). As illustrated in
Figure 2.15, in a balanced transmission system, because the two conductors are twisted together, they are evenly exposed to any sources of electrical
or magnetic interference present. Furthermore, the induced noise signals
travel in the same direction along both conductors, whereas the video signal is travelling in opposite dir ections along each conductor (signal send
and return).
RFI
EMI
Figure 2.15 Noise is induced equally into both conductors in a twisted pair
A receiver unit containing an operational amplifier (op-amp) circuit is
inserted at the output end of the cable for the purposes of noise cancellation. An op-amp has two inputs, and the conductors in the twisted pair ar
e
22
Closed Circuit Television
connected to these inputs. Because the noise signals ar e travelling in the
same direction on both conductors, they ar e effectively applied to both
op-amp inputs in the same phase. However , the action of the op-amp is
such that the noise signals ar e added in antiphase and they ar e thus cancelled out. This noise-cancelling action is illustrated in Figur e 2.16. The
video signal, on the other hand, is only present on the ‘send’ conductor and
is therefore only applied to the non-inverting input on the op-amp. Thus,
the only signal present at the op-amp output will be the video signal.
Video signal
Non-inverting input
Noise
signal
ⴙ
Op-amp
ⴚ
Inverting input
Figure 2.16 Noise at the inverting input is added to that at the non-inverting
input, resulting in cancellation
Because of the noise-cancelling action, a twisted pair cable need not
(in theory) be scr eened. This type of cable is commonly r eferred to as
unshielded twisted pair (UTP). However, in cases wher e large amounts of
RFI are anticipated, a screen is recommended as it provides added protection against induced noise. This cable type is known asshielded twisted pair
(STP). Note that because of the action of the twisted pair,mains hum introduced into the pair via gr ound loops flowing thr ough the screen is cancelled in the same manner as any other noise signal, and thus the inclusion
of the screen poses no problems in this area.
There are two practical issues that must be addressed when employing
twisted pair cable for CCTV signal transmission. First, it is not possible to
fit a BNC connector onto a twisted pair cable. Second, twisted pair cable
has an impedance in the order of 100–150 , making it incompatible with
the 75 impedance of all CCTV equipment. To overcome these issues, the
signal output from the BNC connector on the camera is fed immediately to
a twisted pair transmitter, which both isolates the twisted pair from earth,
and places the video signal acr oss the two conductors. The transmitter
also provides impedance matching between the 75 co-axial cable and
the 100–150 twisted pair cable. At the other end of the twisted pair , an
impedance-matching receiver places the video signal back onto a 75 Signal transmission
23
co-axial output to facilitate connection to the following item of CCTV
equipment. The equipment arrangement for a twisted pair installation is
shown in Figure 2.17.
Co-axial links
Twisted pair
transmitter
Optional screen Twisted pair
receiver
Figure 2.17 Twisted pair transmission arrangement
Because the twisted pair r eceiver contains an active electr onic circuit
(i.e., the op-amp signal pr ocessor), a power supply is r equired. For the
twisted pair transmitter, both passive and active devices ar e available.
Clearly, the passive transmitters do not require a power supply; however,
the active devices generally offer much greater cable distances.
There is no reason why twisted pair and co-axial cabling cannot co-exist
in a CCTV installation. Shorter cable r uns may be made using co-axial
cable, and longer r uns, where signal loss and gr ound loops might pr ove
problematic, may be made using twisted pair cable. In some CCTV
telemetry control systems, it is necessary to run a twisted pair cable alongside the co-axial video signal cable to carry the telemetry data to the
pan/tilt/zoom (PTZ) units or dome assemblies. Wher e the installer has
chosen to use twisted pair for both video and telemetry signals, most
transmission systems will permit both the video and data signals to be
transmitted along a single, four -core cable containing two separate twisted pairs, without interference or cross-talk.
The primary advantage of employing twisted pair video signal transmission in CCTV (compar ed to co-axial cable) is the much longer distances possible owing to the lower signal attenuation in the cable. Wher e
a high-grade cable (i.e., CA T 2 or better) is used along with active transmitters and receivers, a distance of 1000m for a colour signal transmission
is easily possible, and manufacturers frequently quote figures in excess of
2000 m for a monochrome signal.
The main drawback with using twisted pair is the need for the transmitter and r eceiver at each end of every cable r un, which inevitably
increases the cost of the installation. Multiple channel receiver units are
available which reduce both the installation cost and the number of separate boxes scattered behind the control console.
24
Closed Circuit Television
A cost-saving alternative may be to employ equipment at the contr ol
room end that offers direct twisted pair inputs as well as 75 co-axial connection. A typical example is shown in Figure 2.18. To make effective use
of this feature, the installer may still employ transmission equipment at
each camera, or alternatively they may employ cameras that offer a direct
twisted pair output. The advantages are clear: cost saving in receiver (and
possibly transmission) hardware, reduced losses (there are always some
losses when converting from one transmission format to another), and no
need to locate the receive (and possibly transmit) units.
Figure 2.18 A DVR/MUX which offers direct twisted pair video signal input.
(Photo courtesy of Tecton Ltd)
Structured cabling
In Chapter 11 we will look at computer networking and its applications to
CCTV. Here in this chapter we will consider the str uctured cable systems
employed in modern networks; i.e., CAT 5, CAT 5e and CAT 6 (at the time
of writing CA T 7 is still under discussion). The EIA/TIA
(Electronics
Industries Association/Telecommunications Industry Association) have
provided standards for the categories of twisted pair cable systems for
commercial buildings. This is the
EIA/TIA-568 Standard, which was
adopted by the American National Standards Institute (ANSI), although
in Europe these same standards can be found in BSEN 50173. The 568 standard can be divided into two: 568A and 568B. However, for the purposes
of this text, the main dif ference can be taken to be the accommodation of
Tx/Rx crossover connection in the RJ 45 pinout wiring convention (see
Table 2.3). The standar ds are constantly being r evised and updated to
keep pace with the rapidly advancing technology that is associated with
data transmission, which is appar ent when one looks at the number of
revisions to date there are for the 568 standards. An outline of categories 1 to
7 can be seen in Table 2.2. The specifications for each category encompass
not just the cable but the complete data transmission system including data
transmission rates, system topologies, cable specifications, maximum cable
and patch lead lengths, termination impedance, har dware (for example,
network cards, hubs, etc.) specifications and installation practice.
Signal transmission
25
Table 2.2 Maximum test frequency and maximum data rate for categories 1 to 7
from the IEA/TIA-568 Standard
Category
Max test frequency
Max data rate (bps)
1
none
2
3
4
5
5e
6
7
1 MHz
16 MHz
20 MHz
100 MHz
100 MHz
250 MHz
600 MHz
Voice only – not
used for data
4M
10 M
16 M
100 M
1G
1G
1 G (at time of writing, a
10 Gbps option is being
discussed)
In the networking world, the pr essure is on to achieve ever -increasing
data transmission rates and this has seen the industry pr ogress from the
very early CA T 1 (1 Mbit per second (bps)) to the curr
ent CAT 6
(1–10 Gbps). But what determines the maximum data rate thr
ough a
cable? Well primarily, the cable bandwidth. Referring again to Figures 2.1
and 2.2, we know that squar e-wave signals are effectively filtered by the
capacitive and inductive effects of the transmission cable. Therefore, cable
manufacturers are constantly being challenged to develop cables having
properties that offer minimum filter action and therefore maximum bandwidth at high transmission frequencies.
CAT 5 and CAT 6 networks all employ UTP cables, designed to have an
impedance of 100 . CAT 7 will employ STP, where each pair will be individually screened, with an overall outer scr een/shield. The cable specification for the original CA T 5 of fered a vast impr ovement in network
bandwidth over the earlier CA T 3 that it was to supersede. W e see from
Table 2.2 that CAT 3 offered a maximum bit rate of 10 Mbits per second
whereas, when CA T 5 was agr eed in October 1995, speeds of up to
100 Mbits per second became available. As each new specification is introduced, to date, it has always been conditional that the new networks ar e
backwards-compatible.
The CAT 5 standard became obsolete in May 2001 when the EIA/TIA
agreed new standards for an enhanced CAT 5 cable – CAT 5e. This cable
type is simply a more refined version of the original, offering more reliable
data transmission at higher data rates. Both CAT 5 and CAT 5e have been
quoted as being suitable for gigabit Ethernet, of fering speeds of up to
1000 Mbps. However, reliable data transmission at such high rates only
really became possible with the intr oduction of CAT 5e, and in r eality
many networks continued to operate at 100 Mbps.
In June 2002 the EIA/TIAagreed CAT 6. In spite of the greatly improved
bandwidth figure of 200 MHz (which is what is normally quoted – the figure
26
Closed Circuit Television
of 250 MHz in Table 2.2 is the maximum test figure) at first glance it is difficult to see any significant difference between the new standard and CAT
5e, both being capable of 1Gbps data transmission. However, a closer look
at the specification reveals that the crosstalk and noise figures for CAT 6
are far superior to those for CAT 5e. Furthermore, for CAT 6, it is not only
the cable specifications that have been refined. The specifications for connectors, patch cords and network device input/output chip sets ar e also
included in the specification and ar e far more stringent. The vastly superior CAT 6 specification means a much mor e reliable data transmission
when compared with any of the earlier standar ds. Finally, CAT 5/5e networks only use two of the four available cable pairs (leaving the other two
free for other applications such as telephone connection, or even a second
network), whereas CAT 6 utilizes all four pairs (see Table 2.3).
For Ethernet communications, you will often see terms such as
100BaseT being used. The first figure (100) indicates the data rate – in this
example 100 Mbits per second. ‘Base’ means that baseband signalling is
employed, and ‘T’ indicates that the system uses twisted pair cable.
Connection to CA T 3, CA T 5 and CA T 6 networks is via the RJ 45
(Registered Jack) plug/socket (see Figure 2.19). This requirement is a part of
the backwards compatibility which was mentioned earlier , but it must
be remembered that CAT 6 networks require RJ 45 connectors which meet
the higher CA T 6 specification. This is because a network will always
function at the rate of the slowest device in the data chain. Therefore if, for
example, a CAT 6 network were to include CAT 5 connectors or a CAT 5
hub, that network would only be capable of data transmission at the old
1995 CAT 5 standard. This point has important ramifications for the CCTV
engineer who has been tasked to install IP cameras onto an existing network. The customer may be confident that they have a high-performance
CAT 6 network which will easily handle the added load imposed by the IP
cameras. However, if that customer has unwittingly included CA T 5 or
CAT 5e components on their network, the CCTV system may well not perform to standard. Cameras may go off line, picture frames may be continuously lost, or other effects may be noticed on the network such as printers
running very slow, email taking much longer , etc. The poor CCTV engineer will often get the blame for ‘br eaking the network’, but in tr uth the
problem was potentially always there – it just required the added load on
the network to bring it to the surface.
The four twisted cable pairs ar e identified by their colours – br own,
blue, orange and gr een – and it is important during installation that the
pairs are maintained throughout. Using, say, one green wire and one blue
wire as a pair for data transmission would render the noise-cancelling and
crossover-cancelling properties of the cable inef fective, and it is unlikely
that the network would function. T able 2.3 shows the EIA/TIA wiring
conventions for 568A and 568B, for both CA T 5 and CAT 6 installations.
Note that for CAT 6, pins 1, 2, 3 and 6 are the same convention as for CAT
3/5, enabling backwards compatibility.
Signal transmission
27
Figure 2.19 RJ 45 plug and socket connectors. Note that although physically
the same, connectors for CAT 3, CAT 5, CAT 5e and CAT 6 differ in their
specifications. For example, fitting CAT 5 sockets into a CAT 5e network will
reduce the performance bandwidth to that of CAT 5
Table 2.3 EIA/TIA-568 Standard wiring conventions
RJ 45
connector
pin
Wire
colour
(T568A)
Wire
colour
(T568B)
CAT 3
CAT 5
CAT 5e
CAT 6
PoE
connections
1
white/
green
green
white/
orange
blue
white/
blue
orange
white/
blue
brown
white/
orange
orange
white/
green
blue
white/
blue
green
white/
blue
brown
Transmit
Tx_D1
Mode A
Transmit
Receive
Tx_D1
Rx_D2
Mode A
Mode A
Unused
Unused
BI_D3
BI_D3
Mode B
Mode B
Receive
Unused
Rx_D2
BI_D4
Mode A
Mode B
Unused
BI_D4
Mode B
2
3
4
5
6
7
8
28
Closed Circuit Television
In any transmission system, it is necessary to connect the transmit link
(Tx) to the receive link (Rx). In an Ethernet network, devices are normally
connected to each other via hubs, switches, bridges, etc., and crossover of
the Tx and Rx links takes place at these points. Wher e it is necessary to
directly connect, say, two PCs, a crossover patch cable must be used, otherwise communication will not be possible because Tx will connect to Tx
and Rx to Rx. Practical connection arrangements for both straight and
crossover cables are shown in Figure 2.20.
RJ 45 socket
connections on PC
RJ 45 socket
connections on hub
Tx
Tx
1 Tx+
1 Tx+
2 Tx
2 Tx
Rx
Rx
Rx
3 Rx
3 Rx
Tx
4 spare
5 spare
4 spare
5 spare
6 Rx
6 Rx
7 spare
7 spare
Rx
8 spare
Tx
8 spare
Straight cable connection.
Note that Tx and Rx are reversed in the hub
RJ 45 socket
connections on PC
RJ 45 socket
connections on PC
Tx
Tx
1 Tx+
1 Rx+
2 Tx
2 Rx
Tx
Tx
Rx
3 Rx
3 Tx
Rx
4 spare
5 spare
4 spare
6 Rx
6 Tx
7 spare
7 spare
8 spare
8 spare
Rx
5 spare
Rx
Crossover cable connection.
Used to connect two like devices such as two PCs
Figure 2.20 Wiring conventions for EIA/TIA cables and sockets. Note
that the colours used will differ for 568A and 568B network connectors
(see Table 2.3)
Another point to be aware of is that CAT 3 and CAT 5 cables physically
look very similar, so it is always a good idea to verify that the correct cable
has been used for the network installation. Also look out for excessive
bending, stretching or crushing of the cable as all of these will alter the
cable properties and can r esult in excessive data err ors and subsequent
system failure.
Signal transmission
29
Power over Ethernet
A recent development in the field of networking is the IEEE 802.3af Power
over Ethernet (PoE) standard which was adopted in June 2003. The principle is very simple: use the existing Ethernet cables to carry power to network devices, rather than r un separate cables or go to the expense of
installing 230 Va.c. supply outlets at the location of every Ethernet device.
The system is designed to accommodate power transmission using either
the spare pairs in a CAT 3/CAT 5 Ethernet cables, or the actual data lines.
With PoE, the available power on any network segment is 48 Vd.c. at
350 mA maximum, or 16.8 W. The limitation of 350mA is necessary to prevent overload of the r elatively small cross-sectional area network cables.
A PoE power supply is r equired, and this unit must incorporate over current and over-voltage protection. The protection is two-way: it ensures
that network devices are not damaged in the event of a power supply failure, and it also ensures that the power supply itself is not damaged due to
a faulty network device or cable short circuit.
The power supply is normally located in the main wiring cabinet containing the network r outers/switches. For existing installations wher e
PoE is to be included, a PoE midspan hub is added, and the network segments that are to employ PoE ar e patched to the existing switch. Power
would now be available to all network devices on these segments. The
arrangement is illustrated in Figur e 2.21. For new installations, Ethernet
switches incorporating a PoE power supply ar e becoming incr easingly
available, negating the need for a midspan hub.
Network segments not requiring PoE
230 Va.c.
PoE power
supply
Existing Ethernet
router/switch
Patch cables
48 Vd.c
PoE midspan hub
Network segments requiring PoE
Figure 2.21 Connecting a Power over Ethernet (PoE) power supply to an
existing network using a midspan hub
30
Closed Circuit Television
Figure 2.22a shows how PoE is deliver ed using the data cables. Power
is injected via a centre-tapped transformer, which means that, from a d.c.
point of view, both conductors are in parallel. However, from an a.c. (data
signal) point of view, there is no d.c. of fset between each conductor. The
centre-tapped pickup transformer in the network device is often r eferred
to as the picker or splitter. This extracts the 48 Vd.c. supply from the two
conductors and also, via transformer action, recovers the data signal.
Figure 2.22b shows the circuit arrangement where the spare conductors
in a CAT 3 or CAT 5 cable are being used for PoE. Note that both conductors in each pair are being used in order to increase the cross-sectional area
and so reduce power loss in the cable. A voltage regulator or d.c.-to-d.c.
converter in the network device is usually r equired to reduce the supply
voltage from 48 V to, usually, 12 V or 5 V.
The standard colour code for PoE wiring will obviously have to follow
the same convention as the data wiring, because they are in effect the same
thing. However, to ease identification, the colours are included in Table 2.3,
where Mode A denotes the colours used when the data lines ar e to be
employed, and Mode B denotes the colours for when PoE is to be delivered
using the spare conductors.
Midspan hub
PoE network device
4
Spare
4
5
1
Data conductor
5
1
Rx
Tx
2
3
48 Vd.c.
PSU
2
Data conductor
3
Rx
8
Tx
6
6
7
Voltage
regulator
12/5 Vd.c.
to device
7
Spare
8
Figure 2.22a PoE distribution using data conductors in the CAT 5 cable
Midspan hub
4
Spare
4
5
1
Data conductor
5
1
Data conductor
2
3
Rx
Tx
2
3
48 Vd.c.
PSU
PoE network device
Rx
6
7
6
7
8
Spare
Voltage
regulator
Tx
8
Figure 2.22b PoE distribution using spare conductors in the CAT 5 cable
12/5 Vd.c.
to device
Signal transmission
31
The question must now be asked: what happens if 48 Vd.c. is applied to
a network device that does not accommodate PoE? The answer is simple:
the device network input cir cuits will be destr oyed! Clearly this issue had
to be taken into consideration when the IEEE 802.3af standar d was being
developed, and a plug and play solution had to be found because ther e are
many non-technical persons who r outinely connect network devices such
as laptops, etc., to the near est available network point. Every PoE-enabled
device must have a 25 k resistor connected acr oss its input, causing it to
draw a curr ent of 2 mA. Whenever a device is connected to a PoE-enabled
network link, the midspan hub does not immediately apply the full
48 Vd.c. Instead, it applies a lower test voltage and measur es the load cur rent. In ef fect it is pr obing for the 25 k resistor. If it detects a r esistor, the
hub will apply the full 48 Vd.c., otherwise the pr obe voltage is r emoved
and the network link is tr eated as an or dinary data link. This method both
prevents accidental damage to non-PoE devices, and permits devices on a
network segment to be exchanged.
The implications of PoE for CCTV ar e obvious: network IP cameras
are becoming incr easingly popular for a number of r easons, including
the ease of wiring (the other r easons will be consider ed in Chapter 1 1).
However, until the introduction of PoE, a d.c. supply voltage still had to
be provided for each camera. PoE-enabled IP cameras are becoming increasingly available, and as they do so, the industry will lean ever mor
e towards
them for many applications.
Ribbon cable
Also known as ‘flat twin’ cable, this has two parallel conductors and functions on the balanced transmission principle we have just been discussing.
Because the conductors are not twisted it cannot be guaranteed that they
will both be subjected to identical amounts of noise ener gy, although in
practice over short distances this will usually be the case. A typical ribbon
cable construction is shown in Figure 2.23.
Figure 2.23 Ribbon cable construction
32
Closed Circuit Television
This type of cable is useful for interconnection between equipment in a
control room, especially for desk-mounted units wher e a lar ger, more
rigid cable type can prove cumbersome.
Fibre-optic cable
Fibre-optic signal transmission was largely pioneered by the telecommunications industry, and for many years it remained very much within that
industry. Perhaps this was because of the specialized skills and equipment
required to install fibr e-optic cables, particularly in r elation to joining
(splicing) and terminating. Or per haps it was due to the r elative higher
cost of the cable compar ed with co-axial or other copper transmission
medium. Whatever the reason, the CCTV industry was slow to pick up on
this very effective method of sending CCTV signals over any distance.
Because the signal travelling thr ough a fibre-optic cable is in the form of
light, the medium is not prone to any of the problems associated with copper
transmission systems such as RFI, EMI, lightning, etc. Yet fibre-optic transmission has a much wider bandwidth and much lower signal attenuation
figures, which means that signals can be sent over far greater distances without the need for any line-corr ection equipment. Fibre-optic cable also pr ovides complete electrical isolation between equipment so ther e is never any
chance of a ground loop, and from a security point of view it is almost impossible to tap into without it being obvious at the receiving end. The construction of both single fibre and fibre bundle is illustrated in Figure 2.24.
One of the gr eatest problems associated with signal transmission
through optical cable is that of modal distortion, which is caused by the
light energy finding a number of dif ferent paths thr ough the cable.
Because the path lengths ar e not all the same, a single light pulse with a
duration of, say, 1 ns applied at the input arrives at the output over a
period of around 2 ns. In other words, the information becomes distorted.
The longer the cable run, the more acute the problem.
The degree of modal distortion per unit length is determined by the
construction of the fibr e-optic material, and ther e are a number of cable
types available, each having differing characteristics. In order to minimize
modal distortion specially engineer ed cable must be used; however , the
manufacturing of these cables is very expensive. For CCTV systems the
cable runs are relatively short (compared with something like a transatlantic undersea telephone cable!), and ther efore the effects of modal distortion are minimal and cheaper cable designs are adequate.
Three forms of fibr e-optic transmission ar e illustrated in Figur e 2.25.
Monomode cable is the most expensive of the three, owing to its very small
core diameter (typically 5 m) but it of fers the greatest transmission distances with minimal distortion. Step index multimode cable employs two
different materials in the cor e, each having a dif ferent refractive index.
This results in multiple light paths of differing path lengths, which are added
Signal transmission
33
Outer jacket
Kevlar strengthener
Tight buffer
Fibre coating
Optical fibre
Single fibre construction
Outer sheath
Single fibres
Solid central core
Ripcord
24 Way fibre bundle
Figure 2.24 Construction of individual fibre cable (top), and construction of a
fibre bundle (bottom)
at the output to pr oduce a resultant signal. This cable is r elatively inexpensive; however, it offers the greatest degree of signal distortion. Graded
index multimode cable also employs two different core materials; however,
in this case they ar e diffused into each other, resulting in a multitude of
refractive indexes which causes the light to effectively ‘bounce’ from side
to side as it passes along the cable. Graded index (GI) cable of fers much
better distortion figures than step index cable and is much less expensive
than monomode cable, making it the prime choice for CCTV applications.
For both step and graded index cables the cor e diameter is much lar ger
than for monomode cable (typically 50 m).
With this particular type of fibre-optic cable, runs can easily reach 2 km
and, depending on the fr equency of the light sour ce, may extend much
further without the need of any signal booster equipment.
34
Closed Circuit Television
Light source
Inner material
Outer material
Light pick-up
Light input
pulse
Outer protective
sleeve
Light output
pulse with
minimal distortion
Figure 2.25a Monomode cable. Light travels in a straight line through the inner
material
Light input
pulse
Light output
pulse with
high degree of distortion
Figure 2.25b Step index multimode cable. Refraction between two different
materials results in multiple light paths
Inner material
Outer material
Light input
pulse
Multiple light paths. Light rays refract as they enter
the outer material
Light output
pulse with
some distortion
Figure 2.25c Graded index multimode cable. The constantly changing refractive
index results in numerous light paths
The light source may be either a light-emitting diode (LED) or a laser
diode. Of the two, the LED is far less expensive, but it has a much lower
operating frequency. Having said this, an LED can respond at frequencies
of up to 100 MHz, which is more than adequate for CCTV applications. The
frequency of light emitted by the LED/laser diode varies between devices;
however, some fr equencies suffer less attenuation than others ther efore
offering much longer transmission distances, a factor which manufacturers cannot afford to ignore. The video signal is applied to the light source,
which translates the variations in signal voltage into brightness variations.
Signal transmission
35
The light pick-up is a photodiode, which is a device that converts light
levels into corresponding voltage levels. This forms part of a receiver unit
containing the necessary electr onic signal processing to produce a 1 Vpp
standard video signal on a 75 impedance co-axial output.
As discussed earlier, perhaps the main reason why optical fibre cables
are not used mor e widely on smaller CCTV installations is the cost of
installation – and servicing when things go wrong. The cables themselves
are expensive; however, on top of this is the cost of employing specialist
contractors to perform the critical operations of cutting and splicing the
cables, and terminating the ends. Cutting and splicing is critical because
any microscopic flaws or inaccuracies in splices will r esult in the loss of
some of the light modes as they deflect into the cladding. An alignment
error caused by contamination in the alignment jig is illustrated in Figure
2.26. Cutting the cable requires special tools because, as shown in Figure 2.24,
most cables have a Kevlar strengthener running through them and simply
chopping through this with a pair of side cutters would r esult in damage
to the optical fibre core.
There are a number of methods for splicing (joining) optical fibr
e
cables, all of them r equiring specialized and expensive equipment – and
a degree of training. Mechanical splicing involves aligning the two cables
in a jig before being either clamped or fixed using an adhesive. Before the
two ends can be joined, they must be cut accurately and then polished so
that they meet at precisely 90° in order to prevent losses. Alignment is also
critical because, when aligning cor es with a diameter of just 5 m, an
alignment error of anything more than this will result in a 100% light loss.
Another common method for splicing optical fibr e cables is fusion splicing, where the two ends are melted together in a process similar to electrical arc welding. This produces a far more accurate splice than mechanical
methods and removes the need for such accurate cutting and polishing.
There are a number of ways of multiplexing CCTV analogue video signals so that more than one can be transmitted thr ough a single fibre-optic
cable. The actual multiplexing techniques ar e complex and beyond the
scope of this textbook; however, it is useful for the CCTV installer who is
considering fibre-optic video transmission to be awar e that these multiplexers are available. The alternative to using a multiplexer is to use multicore fibre-optic cables, but these may prove to be a more expensive option.
It is important to point out the safety issues surrounding fibre-optic cable.
Although the cable is quite substantial, the optical fibre itself is thinner than
a human hair and particles of fibr e can present an almost invisible health
hazard to engineers. Small fragments can easily puncture the skin and break
off, causing irritation and even infection which can be dif ficult to deal with
medically. Furthermore, broken particles can be inhaled under the right
conditions and could remain lodged in the lung indefinitely. Even for engineers who do not deal with fibre-optic cable directly, where their work is following fibre-optic installation, they should check that the ar ea has been
properly cleaned, and should avoid allowing their skin to come into dir ect
36
Closed Circuit Television
Alignment jig
Contamination
Poorly cut
fibre face
Accurately cut
fibre face
losses resulting from refractions
at badly cut fibre end faces
Figure 2.26 Alignment loss due to microscopic contamination in the alignment jig
contact with the floor or other surface ar eas where they suspect that fibr es
may still be present.
Infrared beam
In essence this is a variation of fibre-optic signal transmission. An infrared
light source is modulated by a video signal, and the light is focused by an
optical assembly onto a r eceiver unit which may be a kilometr e or more
distant. Both the transmitter and receiver must be able to ‘see’ each other
in a straight line of sight.
As with fibre-optic, there are two types of light sour ce: LED and laser
diode. Comparing the two, the LED transmitter is far less expensive, and
it produces a wide, diver ging beam in the or der of 10–20°. However ,
Signal transmission
37
the diverging beam limits the range to, generally , a few hundred metres.
Naturally the limited range can be a pr oblem; however, the wide beam
makes for fairly simple system alignment, and the str uctures onto which
the transmitter and r eceiver units ar e mounted do not have to be completely stable. The laser diode transmitter, on the other hand, pr oduces a
very tight beam of light (ar ound 0.2º) which can travel much gr eater distances. However, the tight beam r equires very accurate alignment, and
equipment really needs to be mounted onto solid structures such as buildings to avoid signal loss caused by movement. Also, beware of the effects
of direct sunlight, which can cause movement of the metal mountings,
throwing the light beam off target.
When locating these units it must be remembered that any break in the
light path will r esult in immediate loss of signal, and although infrar ed
light can penetrate fog and rain to some extent, the range of the equipment
will be reduced, and severe weather conditions may result in a loss of signal.
Therefore beware of operating these at their specified limits: always allow
for some amount of signal loss.Also be aware of other changes which may
occur such as leaves appearing on tr ees during spring, tr ees that may
grow after the equipment has been installed, str uctures which might be
erected, etc.
Infrared links can be very useful for bridging gaps in a CCTV system
which would otherwise prove difficult to deal with. For example, it might
be decided that an additional camera is required in a town centre system,
but its sighting would mean a lot of expensive civil work and much inconvenience to traffic. In these circumstances an infrared link could prove to
be a much more cost-effective and desirable solution. A typical application
is shown in Figure 2.27.
The equipment usually resembles a pair of small camera housings, and
a good quality pr oduct will include many of the essential ‘extras’ which
come with camera housings such as wash/wipe, heater, and in some cases
an optional fan to cool the equipment during the summer – remembering
that extreme heat could cause temporary misalignment of the optics.
Microwave link
The term ‘microwave’ refers to a band of fr equencies in the radio spectrum which extends fr om 3 GHz to 30 GHz (G giga, which is 10 9, or
1000000000). These fr equencies are above the domestic UHF television
channel frequencies and incorporate many of the domestic satellite TV
transmissions and mobile phone channels.
In order to prevent interference between signals the air waves are regulated, and no-one is permitted to operate a radio signal transmitter without proper authority, unless it is on one of the fr equency bands that have
been allocated for free, unlicensed use, and even then the equipment must
comply with certain r egulations regarding its transmission power and
38
Closed Circuit Television
IR beam
IR receiver
IR transmitter
Camera
Figure 2.27 A wide beam infrared link across a main road
bandwidth. A typical example of one of these ‘fr ee for all’ channels is the
CB radio band. These r estrictions mean that CCTV signal transmitting
equipment must operate within specified bands, and must not exceed certain power output levels. In the UK, micr owave CCTV equipment operates on frequencies located around either 3 GHz or 10 GHz.
Directional (dish) aerials are normally used (Figure 2.28). There are two
advantages in this: firstly the transmission range is gr eatly increased if the
power is channelled in one direction, and secondly the signal is more secure
against interception by someone operating r eceiving equipment tuned to
the same frequency. However, dish aerials require careful alignment, as just
a few degrees of error can result in a loss of signal. This also means that the
dishes must be stable. Note that if the dish alignment is only appr oximate,
the signal may produce a perfect picture in good weather conditions; however, as soon as rain or snow settles on the dishes the signal attenuation
causes the picture to degrade, with the familiar ‘sparkles’ appearing. The
technology is identical to satellite TV , and this phenomenon is familiar to
anyone who has a misaligned or incorrectly sized dish on their home!
Microwave energy does not penetrate solid objects, and thus ther
e
must be a clear line of sight between the transmitter and the receiver. This
factor tends to limit the use of micr owave to short range, unless it is possible to locate the equipment at both ends of the system on top of high
structures. Alternatively, the signal would pr opagate well acr oss flat
expanses although there are not many of these in the UK!
In a CCTV application, the transmitter is located close to the camera,
connected by a co-axial link. The r eceiver is located somewher e within
line of sight, but in a place wher e the signal can be sent via cable to the
Signal transmission
39
A dipole (rod) aerial allows the RF energy
to spread in all directions
A dish aerial directs the RF energy in one direction,
increasing the transmission range
Figure 2.28 Comparison of non-directional and directional radio transmission
control room, ideally as close to the control room as possible, or perhaps at
an optical fibre collector point.
Microwave-linked camera installations ar e proving to be incr easingly
popular where cameras need to be located in remote areas, and the cost of
40
Closed Circuit Television
providing both a mains power supply and a video cable is prohibitive. By
employing a camera/housing assembly that incorporates a solar panel
and rechargeable battery plus a micr owave transmitter, the installation
often becomes financially viable.
Two-way microwave links are available, allowing for transmission of
video in one dir ection, and telemetry in the other . The advantage of this
can be seen; however , the equipment cost somewhat r estricts the use of
this technology. Some transmitters also include a sound channel.
UHF RF transmission
In recent years a number of domestic DIY CCTV kits have become available but many DIY householders are unhappy at having to r un cables
around their homes, and when you see the way some of these have been
installed it is not dif ficult to understand why! Thus the pr oposition of a
wireless CCTV system can be very appealing to the householder.
These wireless systems have been adapted from a technology that was
originally developed for domestic TV and VCRs. The idea is to connect the
composite video output from a VCR/DVD player into a local UHF transmitter. Because UHF will penetrate solid objects to a limited degr ee, the
signal can be picked up in every r oom in the house. However, it must be
noted that in many cases the signal may also be picked up by the neighbours. This same equipment has been adopted for local CCTV signal
transmission, but it must be noted that the range is very poor, and there is
no security of the signal whatsoever. On the other hand there are circumstances where a UHF video transmitter designed for domestic VCR use
can prove to be a cost-effective solution to a difficult problem.
CCTV via the telephone network
This idea is not new – it has been around for a few decades – but the problem for many years was the very narr ow bandwidth of the PSTN (public
switched telephone network), which is in the or der of 4 kHz. The idea of
sending a 5.5 MHz video signal along such a cable might at first appear
impossible to achieve, but this was successfully done for many years with
specially developed slow scan equipment.
The principle behind slow scan was to grab a single TV pictur e frame,
digitize it, send it along the PSTN line in the form of a dual-tone audible
signal, and then decode this signal at the receiving end. The problem was
the time that it took to send the digital information for just one frame,
which was in the order of 32 seconds, not including dialling time.
Developments in digital video signal compr ession brought about an
increase in the transmission rate, and fast scan was born. Reductions in the
amount of data, coupled with incr eased modem speeds for PSTN lines,
led to picture transmission rates of up to 1 TV picture frame per second.
Signal transmission
41
On PSTN the bit rate is in the or der of 14 kilobits per second (kb/ s);
however, this figur e was dwarfed by ISDN (Integrated Services Digital
Network). Developed during the 1970s, this was br ought in to cope with
the rapidly incr easing demand on the telephone network. As the name
implies, it is a digital system and ther efore offers much higher data rates:
64 kb/s. This can be incr eased by paralleling lines up to of fer 2 64 128 kb/s. In the UK, ISDN lines ar e installed to order by British Telecom,
and a customer wishing to use this for CCTV fast scan must be pr epared
to pay for the installation and use of this facility.
For many years ISDN proved to be a very effective medium for sending
CCTV over long distances without the need for complex and expensive
dedicated cable installation. However , more recent developments in
Ethernet network technology is rendering this form of CCTV transmission
redundant. And, of course, Ethernet itself may be transmitted over the telephone network using ADSL, so CCTV via telephone network is still very
much alive.
Connectors
The most common co-axial cable connectors in CCTV ar
e the BNC
(Bayonet–Neil–Concelman), phono (also known as RCA) and SCAR
T
(Syndicat des Constructeurs d’Appareils Radio Recepteurs et Televiseurs).
Another type of connector that is very ef fective and r obust is the UHF
(PL259) type, but this has never really been adopted extensively in CCTV,
and tends to be looked upon as something of a nuisance when it does
appear because often engineers do not carry r eplacements, couplers or
converters. A range of connectors is shown in Figure 2.29.
Of the other three types, BNC is by far the most robust, and the locking
action of the bayonet fitting means that it does not easily come adrift.
Furthermore, its construction and method of termination of the screen and
core means that it maintains the characteristic impedance of the cable and
does not introduce a significant amount of signal attenuation, pr ovided
that good quality connectors are used. The best types have a gold-plated
inner pin; or gold-plated inner ferrule in the case of the female connector.
BNC connectors for cable mounting are available as crimp, twist-on and
solder fitting. Crimp types are by far the most suitable for CCTV applications, as long as the correct crimping tool is used. Flattening the flanges with
pliers, wire cutters, hammers, etc., does not pr oduce a low-impedance,
robust fixing, and wher e this practice is carried out installers can expect
to have problems such as intermittent signal loss, ghosting, poor contrast,
noisy picture, and loss of telemetry, to name just a few. Twist-on types are
very quick and simple to fit, and no special tools are required, but care must
be taken to make certain that the screen is compressed tight against the metal
body. If this is not the case, then a high-impedance connection exists, if not
immediately, then perhaps some time afterwar ds if the copper scr een is
42
Closed Circuit Television
Figure 2.29 Range of connectors. From left to right: BNC crimp, BNC T coupler,
BNC straight coupler, two variations of BNC/phono adapter (top), Phono, UHF,
SCART.
exposed to a damp or corr osive atmosphere. Soldered types make a very
strong, low-impedance connection, but they can be extremely messy to fit
and the operation is not one which an engineer would wish to perform on
site, especially whilst working outside, 30’ high in a snowstorm!
As their name implies, phono connectors wer e not originally intended
to be used with video signal cables but rather for connecting between a
phonograph (record player!) and the associated radiogram (hi-fi system!).
However, some equipment uses phono sockets for input/output connections because it keeps the cost down. Wher e this is the case the installer
has no choice but to either fit a phono plug to the co-axial cable, which can
prove very difficult, or fit a BNC to phono adapter, which introduces another
set of contacts, thus incr easing the signal attenuation. Having said this,
where these sockets are used on a monitor or VCR, the chances are that the
connecting leads will only be short inter connecting cables between items
of equipment, and in this case ready-made phono–phono leads using simple screened cable will generally perform satisfactorily.
The SCART connector was originally designed for domestic TV and video
equipment. During the 1980s it became clear that the television receiver of
the future was going to have to be able to accept more than one input source.
Furthermore, it was ridiculous to employ UHF RF coupling when equipment such as VCRs wer e capable of sending a much cleaner composite
video signal. Thus a number of leading manufacturers worked to develop
a multi-pin connector which would facilitate VCRs, a satellite r eceiver,
stereo audio and RGB input and output links, all on separate scr eened
cables contained within a multicor e cable. As can be imagined, initially
there were a number of variations, but thankfully an international standard
was agreed. SCART was never intended for use in industrial applications,
Signal transmission
43
but nevertheless they ar e found on some VCRs and monitors. The
assigned pin connections for a SCART connector are given in Figure 2.30.
1
3
2
5
4
7
6
9
8
11 13 15 17 19
10 12 14 16 18
21
20
1 Audio out, right
11 Green in
2 Audio in, right
12 Data
3 Audio out, left
13 Red ground
4 Audio common (ground)
14 Data ground
5 Blue ground
15 Red in
6 Audio in, left
16 Fast RGB blanking
7 Blue in
17 Composite video ground
8 Function switching
18 Fast blanking ground
9 Green ground
19 Composite video out
10 Data
20 Composite video in
21 Socket ground
Figure 2.30 SCART connector configuration
Another common connector type is the four -pin miniature Din specially developed for S-VHS applications (Figur e 5.14). We shall look at
these in more detail in Chapter 5.
When fitting co-axial connectors, car e must be taken to ensur e that
there are no short circuits between the screen and core. A single strand of
wire between the two conductors and the signal is lost. Careful and accurate stripping of the outer and inner insulators is the key to avoiding short
circuits, and although a sharp pair of wir e cutters may be used for this
purpose, wire strippers specially designed for RG-59 cable make the task
much faster and more reliable.
Cable test equipment
Perhaps the simplest method for testing co-axial cable is to use a multimeter
and test for an open cir cuit between the cor e and screen when the cable is
open at both ends, and 75 when one end is terminated. This, however, will
only tell us that ther e are no short or open cir cuits in the cable – it tells us
nothing about the signal (pictur e) quality, and the wher eabouts of a fault,
should one exist.
44
Closed Circuit Television
One method of testing the signal quality capability is to pass a specific
test signal along the cable, and measure it at the other end. The test signal
could be a sync pulse and black and white test bar pattern derived fr om
a video signal generator. Alternatively a ‘pulse and bar’ generator may be
used. This piece of equipment generates a continuous black, white and
grey pattern without any sync pulses or vertical blanking period, and is
ideal for signal analysis because of its squar e waveshape – remembering
that a square wave is made up from an infinite number of odd harmonics,
so if the cable is intr oducing any high-frequency attenuation the pulses
will appear to have rounded edges (refer to the filtering action illustrated
in Figure 2.2). The measuring equipment at the output end of the cable
on which the waveform is monitor ed would be an oscilloscope. (For
instructions on how to set up an oscilloscope to observe a video signal, see
Chapter 13.)
When the pulse and bar test indicates that a cable is intr oducing h.f.
attenuation, the engineer has to decide what to do about it. If the cable u
rn
is only a few tens of metr es in length, there is no reason why attenuation
should occur, and the cable itself must be suspect, unless it was damaged
during installation. Either way, it would have to be r eplaced. If the cable
length is a few hundr ed metres, then it is clear that some form of corr ection equipment should be installed; e.g., a launch amplifier . When these
have been installed, the pulse and bar generator may be used to check that
the signal is corr ect. In many cases it is necessary to adjust the launch
amplifier response to obtain a corr ect waveshape, and it is not only
rounded edges that ar e a problem – in some cases ther e is excessive h.f.
gain and the waveshape appears to have spikes (called overshoots) on it.
The problem with oscilloscopes is that they ar e rather bulky to carry
around, and although there are some hand-held units available, the LCD
display makes them expensive, and it would be something of a luxury for
every engineer in a larger company to be equipped with one. However, it
is not always necessary to actually see the video signal – the engineer may
only wish to know that it is present, and that it is of the correct amplitude.
To provide this information without the need to set up an oscilloscope
there are a number of video signal str ength indicators on the market.
A typical example is illustrated in Figur e 2.31, where a signal generator
is connected to one end of the cable under test, and a strength meter with
an LED display is connected to the other . Alternatively, the meter can be
substituted with a monitor , and the black and white bar display can be
viewed, although with this arrangement the engineer would not know
whether or not the amplitude is correct.
Another item of test equipment for use with co-axial cables is the time
domain reflectometer (TDR). This takes advantage of the r eflected waves
that occur in a co-axial cable when it is either incorrectly terminated, or is
not terminated at all. Both ends of the cable are disconnected, and the TDR
is connected to one end. When switched on, the unit sends a succession of
short- and long-duration pulses which travel along the cable until they
Signal transmission
45
Figure 2.31 A typical hand-held test instrument for testing co-axial cables,
checking video signal levels, verifying camera output, etc. (Courtesy of ACT
Meters Ltd)
Figure 2.32 A time domain reflectometer is an extremely useful item of test
equipment for installation and service engineers alike. (Courtesy of ACT Meters Ltd)
46
Closed Circuit Television
come to the unterminated end, wher eupon they r eflect back down the
cable and ar e picked up by a r eceiving circuit in the TDR. Because the
speed at which the signal travels along a co-axial cable is known (approximately 200 Mm/sec), the TDR is able to calculate the cable length by
analysing the time it takes for a pulse to return once it has been sent.
The TDR is used primarily to detect the position of a fault in a cable,
which is a very useful thing to know when faced with a cable length of a
few hundred metres, installed in a building, and all you know fr om your
multimeter is that ‘somewhere’ along the cable there is a short or open circuit. Upon reaching the fault in the cable the pulses reflect, and the TDR calculates the distance to the fault, usually indicating this on a digital display
(see Figure 2.32). The TDR is r emarkably accurate, and even over a distance of a few hundred metres it can be accurate to within just two metres.
Apart from detecting cable faults, the TDR can be very useful for determining the length of cable left on a roll, rather than having to reel it all out
and measure it. The TDR can also indicate the location of an excessive
bend in the cable that is impairing its performance.
An optical TDR is available for use with fibre-optic cables, and like the
co-axial version, it is an essential part of a service kit for anyone dealing
with fibre-optic cables regularly. Although the optical unit is somewhat
more expensive, it can very soon pay for itself by making savings in both
time and materials.
3 Light and lighting
Infrared
Light
Although the majority of the equipment in a CCTV system is in the business of processing electronic images in some form or another, the lens has
the task of processing the light that is reflected off the surface of the target.
The importance of this task cannot be stressed enough because, if the lens
fails to focus a true image onto the camera pick-up, then the rest of the system will have no chance of pr oducing faithful and useful images. Yet
before we can look at the operation of the lens, we need to be clear on the
nature of the quantity that it is processing, so in this chapter we will look
at those aspects of light and lighting that appertain to CCTV system
design and operation.
The world’s leading physicists are still uncertain about the true nature of
light. However, they are generally agreed that it is made up fr om minute
particles called photons. But this photon ener gy is electr omagnetic in
nature – just like radio waves – and as such propagates sinusoidally, therefore having a frequency and thus wavelength (the higher the frequency, the
shorter the wavelength). The frequency of a light wave determines the colour.
The spectral distribution for visible light, as well as the infrar ed and
ultraviolet regions, is shown in Figure 3.1. As you can see, the human eye
Ultraviolet
Violet
Indigo
Blue
Green
Yellow
100 nm
Orange
Red
1 cm
700 nm 647 nm 585 nm 575 nm 491 nm 424 nm 400 nm 350 nm
Figure 3.1 Spectral distribution of light, with associated wavelengths
48
Closed Circuit Television
begins to respond to frequencies of light energy at wavelengths of around
700 nm (nanometres), which corresponds to the r ed region, and loses its
response in the blue region at around 350 nm. However, we must remember that the camera pick-up device can have a much wider response, particularly in the infrared region, making it capable of producing a picture in
what would appear to us to be total darkness.
Light and the human eye
The eye reacts to light, converting the incoming electromagnetic radiation
into small electrical signals which are sent to the brain. The brain converts
these signals into an image.
The eye contains four sets of cells; the cells of one set have a cylindrical
structure and are known as rods. The other three sets are conical in shape
and are known as cones. The cones are sensitive to different frequencies of
light, and it is these which enable the eye to differentiate between colours.
One set of cones responds to the range of frequencies with corresponding
wavelengths in the order of 600–700 nm. Upon receiving signals from these
cones the brain acknowledges that it has seen r ed light. The second set of
cones responds to wavelengths ar ound 500–600 nm, that is, gr een light,
and the third set responds to 400–500 nm, blue light.
The eye does not respond equally to all frequencies, nor is the response
equal for all people. As Figure 3.2 illustrates, a good eye responds best to
wavelengths of around 550 nm – that is, green light.
Green
1.0
0.75
Red
0.5
Blue
0.25
350
450
550
650
750 nm
Figure 3.2 Response of the eye to different wavelengths of light
What we perceive to be white light is actually red, green and blue light
emanating simultaneously from a source. Red, green and blue are known
as primary colours, and from these all of the colours in the visible light spectrum are produced. Mixing any two primary colours together pr oduces
a secondary colour, these being yellow , magenta and cyan. This pr ocess,
known as additive mixing, is illustrated in Figure 3.3.
Light and lighting
49
red
magenta
yellow
white
green
Primary colours
= Red
Green
Blue
Secondary colours = Yellow
Magenta
Cyan
cyan
blue
Light level
Red
light
R
White
light
B
G
In this example, adding blue and green
to red results in desaturated red (pink)
Figure 3.3 Principle of additive mixing
We use the term ‘colour’ in everyday language to distinguish between,
say, red and blue. But in practice this term actually encompasses two
aspects which describe the colour: the
hue and the saturation. Hue
describes the frequency of the light, i.e., red, blue, etc. Saturation describes
the intensity or strength of the colour. For example, desaturated red is the
correct term for pink, because it defines the hue as being red, but because
of the addition of small amounts of green and blue, there is an element of
white which desaturates it.
The rods are not frequency (and therefore colour) conscious. These cells
may be seen as broadband receivers that respond to the entire visible light
spectrum, therefore measuring the intensity of the total amount of incoming radiation. They determine if the scene is very bright, very dark, or
somewhere in between. If the eye were to contain only rods we would see
everything in black and white.
Summarizing, the r ods determine the black and white content of an
image, and the cones determine the colour content. The brain constantly
50
Closed Circuit Television
processes the information coming from the four sets of cells to determine
the brightness, hue and saturation of the image.
In a CCTV system, the camera performs the function of the human eye,
converting the incoming light into r ed, green and blue electronic signals.
At the other end of the system, the monitor (or other type of display
device) has to take these signals and convert them back into ed,
r green and
blue light outputs.
Measuring light
In reality this is a complex science but it is not necessary for the CCTV
engineer to have an in-depth understanding of the subject. However, it is
important to have a basic understanding of the quantities and units used
for light measurement as these ar e frequently referred to in both system
and manufacturers’ specifications. The unit of light most commonly
encountered is the lux (lx), so let’s see how this unit is derived and in the
process we shall also define a few other units that engineers may well
come across.
From the point of view of the eye (and ther efore the CCTV camera)
there are two sour ces of light: primary sour ces such as lamps, display
devices, etc.; and secondary sour ces, which are surfaces or objects fr om
which light is reflected.
The light radiated fr om a primary sour ce is known as the luminous
intensity. This is measur ed in candelas (cd), which is the amount of light
that is radiated in all directions from a black body that has been heated to
a temperature equal to that at which platinum changes fr om a liquid to a
solid state.
The lumen (lm) is a measure of luminous flux, which is the light contained
within an area of one radian of a solid angle. One lumen is equal to a luminous intensity of 1 cd within that given area. If this all sounds a bit much,
just imagine one lumen as being 1 cd of light within a given area! Because
the light is constantly diver ging as it moves away fr om the sour ce, the
intensity, when measured at different points away fr om the source, will
reduce. The divergent light reduces by an amount equal to the squar e of
the distance from the source. Therefore, we can say that:
Scene illumination light output
distance squared
This effect, which is known as the inverse square law, is illustrated in
Figure 3.4.
When one lumen of light falls onto an area of one square metre, the surface intensity will be one lux (see Figure 3.5).
Another measure of surface intensity is the foot-candle (or foot-candela).
This unit, which is commonly used in the USA, is derived in the same
Light and lighting
51
Source
A
d
C
B
2xd
D
3xd
4xd
If light level at point A = 1lm, then...
1
= 1 = 0.25 lm
4
22
1
light level at point C = 2 = 1 = 0.111 lm
9
3
1
1
= 0.0625 lm
light level at point D = 2 =
4
16
light level at point B =
Figure 3.4 Light illumination reduces by an amount equal to the square of the
distance from source
1m
1ⴕ
1ⴕ
1m
1ⴕ
1m
Surface intensity
= 1 lux
1 candela
Surface intensity
= 1 foot-candle
1 candela
Figure 3.5 Defining the lux and the foot-candle. As an approximation, we can
say that one lux ten foot-candles
manner as the lux; however, imperial units of feet are used in place of metres.
Conversion between lux and foot-candles is not difficult because three feet
approximates to one metre and therefore as a rule of thumb we can say that
one foot-candle one tenth of one lux or, conversely, one lux ten foot-candles.
Illumination refers to the amount of light coming from a secondary surface or object, the unit of measur ement being the lux or foot-candle. The
material covering or making up a surface will determine the amount of
light reflected off that surface, so even if the primary light sour ce has a
high intensity, if the reflective quality of a surface is poor, the level of illumination will be low. As a typical example, a white pr ojector screen may
reflect 90% of the incident light and will itself serve to illuminate other
52
Closed Circuit Television
objects in the area. However, a dark coloured coat may reflect only 5% of
the incident light, resulting in low illumination.
Thus we can see that the level of illumination is dependent on both the
source lighting and the r eflective properties of the surface ar eas. As a
guide, some typical levels of illumination are given in Figure 3.6.
Starlight
Full moon
Streetlight
0.001 lux
0.1 lux
1–10 lux
100–1000 lux
10 000–100 000 lux
>100 000 lux
Figure 3.6 Typical illumination levels for conditions encountered in CCTV
installations
Light characteristics
As far as we are concerned, we can assume that light travels in a straight
line through a vacuum. Light also travels in straight lines thr
ough a
medium, and the speed at which it travels differs according to the density
of the medium. For example, when a ray of light passes fr om air into a
solid such as glass, the change in velocity causes the ray to bend. As the
ray emerges from the glass back into the air once again, its velocity
increases, causing it bend back towar ds its original angle of incidence.
This effect, known as refraction, is illustrated in Figure 3.7.
Normal
Glass
Figure 3.7a Effect of a solid on the path of a light ray
Light and lighting
53
Air
Water
Glass
Air
Figure 3.7b Effect of differing media on a light ray
The amount by which a ray is r efracted depends upon the change of
velocity, which in turn is governed by the density of the solid. Each solid
has its own refractive index, which is found by dividing the velocity of light
in a vacuum by the velocity of light within that solid. Air and clear gases
are taken as having a refractive index of 1.
However, it is not only the r efractive index which governs the angle of
refraction. The frequency (i.e., the colour) of the light ray also has a bearing.
Experimentation proves that for any given solid, the blue end of the light
spectrum will always be refracted more than the red. This is why when we
pass white light through a prism, we observe a ‘rainbow’ ef fect. Although
this effect might look striking in a table ornament, it causes considerable
headaches for lens manufacturers, as we shall see in the next chapter.
Artificial lighting
All too frequently lighting is the one factor that the CCTV system designer ,
installer and end-user have little control over. To guarantee clear, true, colour
images from a video camera, the scene should be illuminated with a high
level of evenly distributed white light. T o illustrate this point, consider the
following scenario: a television production company wish to film a scene in
a street. Apart from the film crew, sound engineers, production people, electricians, actors, catering van, etc., there will be a dedicated lighting crew who
will erect as much lighting and reflective surfaces as they deem necessary to
ensure a high-quality pictur e, free from any shadows. Now the cr ew have
gone away, and all that is left is a CCTV camera covering that same str eet,
and its owner expects it to pr oduce the same quality images in all weather
and lighting conditions, with the only form of artificial lighting at night
being the sodium street lights. Without the lighting crew, this camera has no
chance of meeting this expectation!
Bearing in mind that visible white light actually comprises the thr ee primary colours red, green and blue, it follows that the only way (using current
technology) of achieving true colour r eproduction with either video or
54
Closed Circuit Television
photographic cameras is to illuminate the subject ar ea with white light. This
point is illustrated in Figure 3.8 where, for the sake of this illustration, we shall
assume that the person’s jacket is pur e blue. The pigment in the jacket
is such that the red and green light energy is absorbed, but the blue light
energy is reflected off the surface. Therefore the eye (and thus the brain)
discerns only blue light in that part of the image. Now consider the same
image if it is illuminated with a pur e red light. All of the light would be
absorbed, so the eye would per ceive the jacket as being black. This leads
us to a very important point regarding CCTV installation: colour cameras
are only effective where the area is illuminated with white light. There is
no point in installing a colour camera wher e infrared (IR) lighting is to be used
although, as we shall see in Chapter 6, there are some colour cameras that
will automatically switch to black and white operation when it is dark or
when IR lighting is present.
White light source
Blue jacket
Red
n
Gree
e
lu
B
Blue
Figure 3.8 Pigment in the blue jacket absorbs red and green frequency
radiation, reflecting only blue frequencies
There are a number of sour ces of ‘white’ light and most ar e not tr ue
white, but nevertheless when used with CCTV equipment many of these
will provide adequate illumination. The type of light produced by a source
is determined by the colour temperature, which is a scientific measure of the
wavelengths of light and is stated in degrees Kelvin. For example, the light
from a fluorescent tube is known to be about equivalent to that of an overcast day because in both cases their colour temperatur
e is ar ound
6000–6500 K.
The spectral output range for a number of common sour ces of artificial
lighting is given in Figure 3.9. Figure 3.9a shows the overall light output for
a number of sources; however, Figure 3.9b gives a much clearer indication of
the usefulness of some of these sour ces. Bearing in mind that monochr ome
cameras respond well to IR light (unless they have been manufactur ed with
IR filters), then we can see that for monochr ome CCTV installations almost
any of the common artificial light sources will provide adequate illumination.
However, for colour operation we must be more selective.
Light and lighting
55
As a general rule, colour cameras work best with tungsten or tungsten
halogen lighting. Although Figure 3.9a would indicate that high-pr essure
sodium lighting pr ovides suitable coverage of the visible spectr um, in
practice this type of lighting pr oduces mainly yellow illumination with
only small amounts of blue and red. Fluorescent lighting produces a somewhat irregular output across the visible spectrum, but most cameras function reasonably well under these lighting conditions as long as the tubes
are not operating at the 50Hz (60 Hz) mains frequency, which can result in an
undesirable flickering effect on the picture. Note the very narrow response
of low-pressure sodium lamps. These pr oduce mainly a yellow/orange
light, making them unsuitable for both colour and monochr ome cameras;
however, until r ecent years this type of lighting had been used as the
main form of street lighting in the UK but much of it is still in use in urban
areas.
IR LED
Filtered tungsten halogen
Mercury vapour
Xenon
Low pressure sodium
High pressure sodium
Metal halide
Fluorescent
Tungsten / Tungsten halogen
300
400
500
600
700
800
900
1000 Wavelength (nm)
Visible light spectrum
Blue
Red
Figure 3.9a Spectral output range for common sources of artificial lighting
Relative light output
Fluorescent
300
Tun
en
gst
gste
n
n
oge
hal
Tun
400
500
600
700
800
900
1000 nm
Figure 3.9b Typical spectral response of common artificial lighting devices
56
Closed Circuit Television
Another problem associated with gas dischar ge lamps (i.e., low- and
high-pressure sodium) is the time that it takes for them to strike and then
reach their optimum light output performance. The time fr om strike to
maximum light output can be a few minutes, and furthermore, if they are
turned off, it is not safe to strike them again until they have cooled, which
can also take a few minutes. For this r eason, these types of lighting ar e
unsuitable for use wher e switched security lighting is r equired. On the
other hand, although tungsten halogen lamps r each their optimum light
output almost immediately, where they are being constantly switched on
and off it is generally found that the lamp life is relatively short due to failure of the filament.
Infrared lighting units are usually tungsten halogen lamps with a filter
mounted in front. There are two common types in use, one producing illumination around 730 nm, the other having a longer wavelength in the
order of 830–860 nm. 730 nm units tend to produce a small amount of visible red light which makes them visible to the naked eye – although the
area that they are illuminating appears dark. These units are ideal for use
in locations where white lighting is undesirable (perhaps because the local
residents do not want their bedrooms illuminated like a theatre stage!) but
where overt CCTV is required to serve as a deterrent. Where the CCTV system is required to be covert, employment of IR lamps in the 830nm region
will result in a high infrar ed illumination of the cover ar ea, although the
lamps will be invisible to the eye.
Because infrared lamps are often located in places that ar e not easily
accessible, the problem of lamp unreliability is made more acute. For this
reason LED (light-emitting diode) IR lamps comprising an array of
infrared LEDs are becoming increasingly popular for applications of up to
around 100 W. The idea is illustrated in Figure 3.10. In general these lamps
are covert, operating at around either 850–880 nm or 950 nm.
Figure 3.10 The drawing illustrates the principle of the LED array.
The photograph shows a typical lamp unit. Note the heat sink fins at the rear of
the lamp which provide convection cooling
Light and lighting
57
When we think about the intensity of CCTV lighting, too often we ar e
only concerned about the ar ea being too dark. However, too much lighting can also pose problems. Just as the human eye does not cope well with
excessive bright light, neither does a CCTV camera. Consider the situation
where a fully functional camera (i.e., one mounted on a pan/tilt unit and
having full operator contr ol) with artificial lamps is r equired to cover a
large car parking area. In order for the operator to be able to discern activity anywhere in the ar ea it may well be necessary to install two 500 W
flood lamps – either white or infrared. However, problems will arise when
the camera is tilted downwards to view a target in close proximity because
the target will appear completely over exposed. This pr oblem may be
overcome by having more than one type of lamp available to the operator,
each being independently switched. For example, in the case of the car
park, a high-wattage flood lamp could be used when viewing the general
area and a low-wattage spot lamp could be switched on when the camera
is being used to monitor the area close by.
In general ther e are three types security lamps available: flood, wide
angle, or spot coverage (Figur e 3.11). The choice of distribution type is
important. A 500 W flood lamp will cover a wide ar ea, but it may not provide sufficient illumination in any particular part of the ar ea. At the other
extreme, a 500 W spot lamp will leave large areas with no illumination and
might well result in overexposure of the target area. From this we should
appreciate that, when installing lighting for CCTV , consideration should be
given to such factors as the ar eas to be covered, the amount of illumination
required, and the types of lighting – i.e., white, infrared, overt or covert.
Flood
Wide angle
Spot
Multiple angle
250 W
narrow
angle
250 W
spot
500 W flood
Figure 3.11 In some situations a mixture of lighting distributions is required to
provide adequate illumination
Consideration should also be given to the angle of the lighting. In
general, security lights have to be mounted high up both to ensur e their
safety and to place them in the proximity of the camera (where the power
source and telemetry control will be available). However, if the only lighting source is from above then facial features may be lost due to the heavy
58
Closed Circuit Television
shadowing effects around the eyes and mouth. For example, the shadow cast
by the nose could well be misinterpr eted as a moustache. Also, beware of
low-level downlighting used to illuminate some paths and walkways.
Whilst the level of illumination can be mor e than adequate to satisfy the
needs of the CCTV camera, this type of lighting tends only to pr ovide illumination of the waist down, ther efore making it impossible to identify the
persons in the shot.
We have already seen that light levels ar e measured in lux. The instr ument used for taking these measurements is a light level meter, or lux meter.
From the point of view of CCTV camera installation, wher e it is felt necessary to test the light level in an ar ea, the most practical point to take the
measurement is at the camera lens because this will give an indication of the
actual amount of r eflected light entering the lens fr om the secondary surfaces and objects (see Figure 3.12). When using a light meter, take note that
the photo sensor deteriorates with exposure to light. Manufacturers provide
a black cover for the sensor and, to ensur e accurate readings and preserve
the life of the meter, this should only be removed whilst taking a reading.
Figure 3.12 When using a light level meter, take the reading at the camera
location
4 Lenses
The performance of a CCTV system is very much r eliant on the quality
and type of lens fitted to the camera. A system will offer poor picture performance when the installer does not specify the corr ect lens during the
initial survey, and ‘corr ect lens’ does not simply mean choosing a lens
which will offer the correct field of view, although this is one important
factor. The quality of the lens, the format size and the spectral ersponse are
all-important factors relating to lens performance and thus image quality.
For example, there is no point in fitting a lens with an infrared filter when
the system is expected to perform in the dark with the assistance of artificial infrared lighting! And this has been known to occur.
In this chapter we shall begin by looking at the principle of operation of
optical lenses befor e moving on to discuss the lenses employed in the
CCTV industry.
Lens theory
An optical lens is a device which makes use of the refractive effect on light
paths. There are two types of lens: convex and concave.
A simple convex lens is shown in Figur e 4.1a. The light rays entering
the lens are refracted, but because the lens surface is curved, the angle of
emergence at each point on the lens is different. If the lens is ground accurately, then all of the rays of light will converge at a single point somewhere
behind the lens. This point is called the focal point.
Focal point
Incoming light
Figure 4.1a Convex lens. Light converges onto a focal point
60
Closed Circuit Television
The concave lens is shown in Figure 4.1b. This is known as a diverging
lens because the light rays are bent outwards. In this case the focal point is
said to be at a point on the entry side of the lens fr om where the light
appears to have originated.
Apparent focal point
Figure 4.1b Concave lens. Light diverges from an apparent focal point behind
the lens
One problem encountered with a simple lens is
chromatic aberration.
Different colours of the visible light spectr um have a dif ferent refractive
index, so for the lens in Figur e 4.1a, the red end of the spectr um will not be
bent as far inwar ds as the blue end. This means that ther e are a number of
focal points, i.e., one for each wavelength in the light spectrum, resulting in an
image that appears to have a number of colour ed haloes around it. In effect,
the lens is behaving rather like a prism. The problem of chromatic aberration
can be overcome by using a lens assembly comprising a series of converging
and diverging lenses, each made from glass with a differing refractive index.
The efficiency of a lens is reduced if the glass is highly reflective because
a considerable amount of light simply bounces of f the front face. Lenses
often have a bluish tinge because they have been coated with a filter material to reduce this effect.
Looking again at Figure 4.1a, we can see another important characteristic of the convex lens. Because the light paths cross over at the focal point,
the image is inverted . Of course, this inversion is cancelled if the light is
passed through another convex lens.
A practical lens assembly incorporates more than one optical lens , which
means that the light paths may cr oss a number of times as they pass
through. This is illustrated in Figure 4.2.
The image will only appear in correct focus on the pick-up device when all
of the lenses within the assembly are at the correct distance from each other.
We shall see later that some lenses ar e made adjustable to of fer zoom effects;
however, this means that the focus must be re-adjusted, which is done by moving
the position of the lens assembly using an adjustment called the focus ring.
Lenses
61
Focal length
Lens assembly
Camera
Pick-up
device
Primary
principal
point
Secondary
principal
(nodal) point
Figure 4.2 Light paths through a lens assembly
Lens parameters
Having looked at the basic operating principle of a lens, we can now give
a more detailed consideration to the subject. A sound understanding of the
meaning of, and the principles behind, the terms lens format, focal length,
angle of view, field of view, aperture, F-number, and depth of field, is essential for
anyone who wishes to effectively specify and/or install CCTV systems.
Let’s begin with lens format. This is directly related to camera format, which
refers to the size of pick-up device employed by any particular camera.
Pick-up devices come in a range of formats (sizes). CCTV cameras originally
employed vacuum tube pick-up devices, and the 1 tube was a common
choice, offering reasonable performance at an af fordable price. How ever,
tubes have long been superseded by the charge coupled device (CCD). We
shall look at CCDs in Chapter 6.
CCDs do not require the same pick-up size as their tube counterparts to
produce a picture of matching resolution, which is why 1⁄4, 1⁄3, 1⁄2 and 2⁄3 format cameras are common in the CCTV industry . Rapid advancements in
CCD chip technology has led to ever-increasing resolution from the smaller
chips, making the 1⁄3 and 1⁄2 format cameras the m ost popular choice for
general purpose applications. 1⁄4 CCD chips are very common, but (at the
time of writing) cameras employing such chips generally r equire reasonable lighting levels and ar e only capable of pr oducing lower-resolution
colour images. As the cost of CCD chip pr oduction continues to fall, 2⁄3
format cameras are becoming increasingly popular; however, relative to
their smaller counterparts, these cameras ar e still mor e expensive and
their use tends to be r elegated to high-security CCTV installations, highway traffic monitoring or town centr e monitoring systems. 1 format
CCDs are available, offering very high resolution and excellent low light
level performance, but at a very much higher cost.
Five formats are shown in Figure 4.3, along with the corresponding horizontal and vertical chip dimensions. The chip size complies with the industry standard ratio (aspect ratio) of 4:3, and these horizontal and vertical
dimensions are standard for all pick-up devices. Imperial units are used to
denote the camera and lens format to prevent confusion between the lens
62
Closed Circuit Television
m
4m
m
6m
2.4 mm
3.2 mm
m
3.6 mm
8m
4.8 mm
4.8 mm
6.4 mm
1/4 format
1/3 format
1/2 format
m
11
m
m
.9
6.6 mm
m
15
8.8 mm
12.7 mm
2/3 format
1 format
Format
Imperial
X
Metric
X mm
⁄4
⁄3
1
⁄2
2
⁄3
1
4
6
8
11
15.9
1
1
9.5 mm
Height
X mm
Width
X mm
2.4
3.6
4.8
6.6
9.5
3.2
4.8
6.4
8.8
12.7
Figure 4.3 Dimensions of imager size for each CCD chip format
format and focal length figures. For example, it is much less confusing to
talk about a 1⁄2 12 mm lens than it is to r efer to a 12 mm 12 mm lens, even
though these are more or less the same.
It is important to note that the actual dimensions of the CCD device ar e
smaller than the format size. Take for example the 1⁄2 format CCD. You might
expect the diagonal dimension to be a1⁄2 (12.5 mm); however, it is only 8mm.
The same r ule applies to the other thr ee image devices; their diagonal
dimensions are all smaller than their quoted format sizes. The r eason for
this is that we do not use the light output from the lens at the outer edges of
the image, as this is wher e maximum optical distortion occurs. The r elationship between lens and camera is illustrated in Figure 4.4.
Let’s now look at what happens when the lens/camera formats have not
been correctly matched. In Figure 4.5a the lens format is smaller than the camera
format, and so the image does not fill the display. When viewed on a monitor,
it would appear as if we wer e looking through a porthole! In Figure 4.5b
the lens format is larger than the camera format, so in this case the image fills the
monitor screen; however, not all of the image produced by the lens is being
Lenses
63
Figure 4.4 Proportional dimensions of a 1/2 format CCD when compared with a
1/2 diameter circle, illustrating the relationship between lens and camera formats
(a) A 1/3 format lens fitted
to a 2/3 format camera
(b) A 2/3 format lens fitted
to a 1/3 format camera
Figure 4.5 Two examples of mixed lens and camera formats, drawn to scale
used. This is not necessarily a bad thing, because by only using the centre
area of the lens there will be minimal optical distortion. However, the operator must be able to view the area specified in the operational requirement.
Moving on to focal length, looking again at Figure 4.2 we see that the focal
length is the distance, measured in millimetres, from the secondary principal point
to the final focal point . The secondary principal (or nodal) point is the last
occasion where the light paths cross in the assembly, and the camera pick-up
device is located at the focal point. For small, wide-angle lenses that have
a very short focal length, it would not be possible to fit the lens to the camera because the r ear lens assembly would be for ced against the pick-up
device. In these cases it is necessary to fit additional optical devices to
compensate for this, moving the final focal point further away fr om the
rear lens assembly, without altering the effective focal length of the lens.
The focal length has considerable effect on the performance of the lens.
The effect of changing the focal length, that is, moving the nodal point, is
shown in Figure 4.6, where it can be seen that the angle of view changes.
But moving the nodal point not only changes the angle of view; it also
affects the magnification (M). All people with normal eyesight have mor e
or less the same angle of view (which is appr oximately 30°) and magnification. Taking this value as a magnification of M 1, it can be shown that
a 1⁄3 format CCTV camera fitted with an 8 mm lens has a magnification of
approximately 1, thus producing an image close to that per ceived by the
human eye. Table 4.1 shows the focal lengths for other lens formats having
a magnification of M 1.
64
Closed Circuit Television
θ
θ
θ
Nodal point
f 8 mm
Angle of view 30°
Normal view
f 4 mm
Angle of view 30°
Wide angle
f 12 mm
Angle of view 30°
Telephoto
θ angle of view
Figure 4.6 Effect of focal length on the angle of view
Table 4.1 Lens sizes having a magnification M 1. In each case the angle of
view is approximately 30°
Format size
Focal length
1
25 mm
⁄3
16 mm
2
⁄2
12 mm
1
⁄3
8 mm
1
⁄4
6 mm
1
Increasing the magnification incr eases the apparent size of an object,
and lenses with a long focal length ar e known as telephoto lenses because they
have a narrow angle of view, but objects far away appear much larger than
they would when viewed with the naked eye. Lenses with focal lengths that
produce an angle of view wider than 30°are known as wide-angle lenses because
they cover a broader area than the eye; however, the magnification is less.
It can be shown that the magnification of a lens is determined by the
expression:
M
focal length ( f )
format size
where M is the magnification factor, and format size is the size of the lens.
For example, for a 1⁄2 format camera fitted with a 12.5 mm (1⁄2) format,
25 mm focal length lens, the magnification will be:
M
25
2 times
12.5
However, if a 100 mm focal length lens is fitted to the same camera, the
magnification becomes:
M
100
8 times
12.5
Lenses
65
So we see that increasing the focal length increases the magnification. However,
the increase in focal length gives a corr esponding reduction in the angle
of view and thus the image size, known as thefield of view, is reduced. Putting this another way, increasing the magnification (zooming in) r educes
the area that can be viewed.
Field of view is one of the more critical factors in CCTV system specification and design because it determines how lar ge an image will appear
on the monitor screen for a given distance from the camera. When we talk
about the field of view we need to define whether we ar e referring to the
vertical or horizontal dimension (Figure 4.7).
W
H
Scene
e
to
ob
j
ec
t(
D
)
Nodal point
di
st
an
c
Lens assembly
W = width of scene
fo
c
al
le
ng
th
(f
)
Camera pick-up
device
H = height of scene
w = width of format
h
h = height of format
w
Figure 4.7 Relationship between the focal length of a lens, the distance of the
object from the lens and the field of view
But field of view is not only affected by the focal length; it is also a function of the format size and the distance fr om the camera to the object in
focus. This relationship is expressed as:
f
w
h
W
H
D
Where w is the width of format, W is the width of scene or object, h is the
height of format, H is the height of scene or object, f is the focal length of
66
Closed Circuit Television
lens, and D is the object distance from camera. Note that in some instances
the letters ‘V’ and ‘H’ are used to denote the vertical height and the horizontal width. This is particularly the case with some lens calculators (see later
in this chapter).
Example
It is required that an object of 2.5m in height fills the monitor screen when
observed at a distance of 15 m. The camera format is 1⁄2. Calculate the
required lens focal length.
Solution
From the table of imager sizes (Figure 4.3), a 1⁄2 format pick-up has a vertical size of 4.8 mm. Thus:
from
f
h
H
D
f h
4.8
D 15 000
H
2500
28.8 mm focal length
Naturally this is an ideal size, and the near est available size, e.g. 32 mm,
would have to be employed.
Example
For the example above, what would be the horizontal field of view (W)?
Solution
From the table of imager sizes (Figure 4.3), a 1⁄2 format pick-up has a horizontal size of 6.4 mm. Thus:
from
f
w
W
D
W 6.4 15 000
wD
28.8
f
3333 mm or 3. 333 m wide
The above examples site the maximum height and width of the field of
view, but in some cases we have to calculate the lens size from the point of
view of the size of an image on the screen. For example, it may have been
Lenses
67
decided that in order to be able to recognize a person of average height at a
range of 15 m, the image of the person must fill at least 67% (2/3) of the vertical size of monitor scr een. Faced with such a situation, the specifying
engineer must ensure that the lens fitted to the camera will perform to this
specification. This problem may be approached in the following manner:
●
●
Assume that a 1⁄3 format camera is to be employed, then (fr om Figure
4.3) h 3.6 mm.
Assume it has been decided that a person of height 1.6 m will be the
minimum height of person we wish to target. For this image to fill 67%
of the screen, the full field of view height will be 1.6m 1.5 m 2.4 m,
or 2400 mm H.
D 15 000 mm.
Thus
f 3.6
15 000 22.5 mm
2400
In this example it was simple to find the full field of view height, because
it was a convenient number: 2/3.An alternative method of calculating this
figure when the percentage screen height is not so straightforward is:
H 100%
target height (metrees)
target height (%)
To prove this equation, let’s insert the figures from the last example…
H 100%
1.6 2.39 metres
67%
As a general rule, for any format size, doubling the focal length halves the field
of view, and doubles the magnification . In other wor ds, we only see half as
much, but it appears twice as lar ge. This is a useful r ule of thumb when
working on site.
Calculating the lens size for a particular application is very important.
Failure to get this right at the specifying stage can pr ove expensive in the
long run when a customer refuses to pay the bill because the images produced on the screen are nothing like those laid down in the specification,
or Operational Requirement (OR). Given the cost of some lenses it could
prove expensive to go around changing them afterwards. The calculations
in the previous examples are based on simple geometry; however , there
are many engineers who ar e not comfortable having to perform mathematical calculations, especially on site! For these engineers ther e is some
good news. Various lens calculators, look-up tables and lens finders have
been developed which for most applications do away with the need for
any calculations. We shall look at these at the end of this chapter.
68
Closed Circuit Television
For many fixed cameras, an alternative to pr ecision lens calculation is
to use a varifocal lens. This offers a crude 2:1 adjustment range, typically,
4–8 mm or 6–12 mm. The principle of operation is very simple. Referring
again to Figur e 4.2, a varifocal lens has an adjustment which moves the
centre lens assembly, making the focal length, and ther efore the magnification, variable. However, it must be emphasized that this is not a tr ue zoom
lens because, every time the focal length is altered, the focus ring must also be
reset. A zoom lens has a mor e complex optical assembly which compensates for the changing focal length and maintains correct focus throughout
the zoom range. Zoom lenses will be discussed later in this chapter.
Many installers employ varifocal lenses because they allow for a margin of error and/or a change of mind on the part of the customer once the
installation has been completed. On the other hand these lenses ar e more
expensive than equivalent quality and size fixed lens and, if the lens is
never going to be adjusted after the initial installation, a fixed lens is far
more cost-effective and can make a quote for a system more competitive.
Another important consideration for a CCTV camera is the amount of
light falling onto the pick-up sensor. Insufficient light results not only in a
dark picture, but it will also be lacking in contrast. Excessive light will
result in what is known as pictur e burnout. This is where lighter areas of
the picture become overexposed, resulting in bright white patches with no
detail whatsoever. In general, as long as the image is not allowed to burn
out, the picture quality will be better wher e there is a high level of light
input. The contrast ratio will be high, and the depth of field (we shall look
at this later in the chapter) will be greatest.
Because a camera is expected to function effectively across a wide range
of lighting levels, the lens must have some means of controlling the amount
of light falling onto the camera pick-up. This is done by means of a mechanical iris. The iris is a delicate mechanism comprising a number of plates
which slide as they are rotated. The principle is illustrated in Figure 4.8.
Figure 4.8 Sliding plates are used to perform the function of an iris
The size of the hole created by the iris is known as the aperture. As the diameter of the apertur e changes, the change in the amount of light passing
through follows a mathematical r ule known as the inverse squar e law. If
this sounds a bit of a mouthful, then consider it this way:doubling the diameter of the aperture results in an increase in light throughput of 4 times (22); tripling
Lenses
69
the diameter of the aperture gives an increase of 9 times (33), and so on. Similarly,
halving the diameter of the apertur e results in a light r eduction factor of
four, etc.
The light-gathering ability of a lens is known as its optical speed. The greater
the amount of light falling onto the camera pick-up, the faster the CCDs
will charge. Hence a lens with a high light throughput is said to have a fast
speed. Larger aperture lenses will obviously be capable of gathering more
light than smaller ones, and so in general they have a faster optical speed
than smaller lenses. The optical speed is stated as the F-number.
The F-number is determined from:
F
focal length
lens/aperture diameter
If you look at the F-numbers on the side of a manual iris lens, you will see
that they appear to have somewhat peculiar values. This is because each
position, known as an F-stop, is designed to give either a halving or doubling of
light throughput, and the r eason why the numbers ar e at first glance
unusual is down to simple geometry. To double the light throughput, the
aperture area is doubled; however, to double the area of a circle, the diameter is not doubled; rather, it is increased by a factor of 1.4142. Likewise, to
halve the ar ea of a cir cle, the diameter is divided by a factor of 1.4142.
Look at the following example.
For a 50 mm diameter lens with a focal length (FL) of 50
mm, the
F-number would be 50 ÷ 50 1.
●
●
In order to reduce the light-gathering area of this lens by 50%, the aperture diameter will be 50 ÷ 1.4142 36 mm.
This gives an F-number of 50 ÷ 36 1.4.
Repeating these calculations for the 36 mm aperture area would give us a
25 mm aperture size, and an F-number of 2.0. So it can be seen that moving the aperture setting from the 1.4 position to the 2.0 position, i.e., one
F-stop, reduces the light throughput by a factor of 50%. Fr om this we can
conclude that each and every F-number represents a halving or doubling of light
throughput.
To illustrate this effect further, Table 4.2 shows the effect of each F-stop
on the aperture diameter and light attenuation for a total of 13 stops on a
36 mm diameter, 50 mm focal length lens. F1.4 indicates that the apertur e
is fully open. In this example, it has been assumed that when the aperture
is open, a light level of 2 lux is passing thr ough the lens, and this level
of light falling on the CCD chip in the camera is suf ficient to produce an
acceptable clear image. If the light input level incr eases, the aperture will
have to close down accor dingly to maintain a level of 2 lux at the CCD.
Thus, for example, when the light input incr eases from 2 lux to 4 lux, the
aperture diameter must reduce from 36 mm to 25 mm to maintain an output level of 2 lux at the CCD chip.
70
Closed Circuit Television
On a manual iris (MI) lens the aperture is adjusted by rotating a ring or
collar which has the F-numbers indicated beside it, the letter ‘C’ marking
the point where the aperture is completely closed.
From Table 4.2 notice that when the light level incr eases from 2 lux to
8000 lux, the aperture diameter changes from 36 mm to 0.6 mm, a total of
Table 4.2 Effect of F-stops on aperture diameter and amount of light attenuation.
It is assumed that the light output from the lens (at the CCD pick-up device) is 2 lux
when the aperture setting is 1.4
Aperture diameter
36 mm
36 mm
25 mm
18 mm
12.5 mm
8.8 mm
6.25 mm
4.4 mm
3.1 mm
2.2 mm
1.6 mm
1.1 mm
0.8 mm
Area ÷ 2 (÷ 1.4142)
–
25 mm
18 mm
12.5 mm
8.8 mm
6.25 mm
4.4 mm
3.1 mm
2.2 mm
1.6 mm
1.1 mm
0.8 mm
0.6 mm
F-number
1.4
2.0
2.8
4.0
5.6
8.0
11
16
22
32
44
64
88
Light input (lux)
2
4
8
16
32
64
128
256
512
1024
2k
4k
8k
13 stops, in order to maintain a constant level of 2lux at the pick-up. 8000lux
is what might be expected on a dark, over cast day. For this same lens to
deliver 2 lux at the pick-up on a bright sunny day wher e the light level is
typically 500 000 lux, the apertur e would have to close down another 6
stops, and the diameter would become just 0.1mm. Such a small diameter
is simply not possible to control, either manually or automatically, because
an error of just 0.05 mm would result in a light dif ference of 25% and, in
the case of an automatic iris lens, the contr ol circuit would ‘hunt’ (see
‘Electrical connections’ later in this chapter) causing the iris mechanism to
oscillate. Clearly some other solution must be found.
The answer to the pr oblem is to add a neutral density (ND) spot filter
inside the lens. A neutral density filter is one which affects all frequencies
of visible light by the same amount, and ther efore has the effect of reducing the overall light level. The term ‘spot filter ’ in this case r efers to the
fact that the filter does not cover the entir e lens area but rather appears
like a spot in the centre of the lens. The filter is designed to have maximum
effect in its centre, reducing consistently between the centre and the outer
edge. The principle is illustrated in Figure 4.9.
When the aperture is fully open the filter has little effect on the overall
amount of light passing through. However, as the diameter of the aperture
approaches that of the filter, its effect begins to be felt. Consider the effect that
such a filter would have on the lens example we have been considering in
Lenses
71
ND filter
Lens
Figure 4.9 ND spot filter on a lens
Table 4.2. Let us suppose that the filter begins to take effect as the aperture
diameter reaches 1 mm. As the light input increases from around 2000 lux
to 4000 lux there is no need for the iris to close down as far to maintain a 2lux
output, because some of the incoming light is lost through the filter. Further
increases in the incoming light level are met with corresponding amounts of
filtering as the iris closes, and the centr e area of the filter becomes the only
part of the lens that is being used.
One option available to lens manufactur ers for obtaining lower F-stop
figures is to produce aspherical lenses. Without going into great detail, these
lenses are ground in such a way that they ar e made to be mor e of a bell
shape, which gr eatly improves the light-gathering and focusing ability .
The increase in light ef ficiency enables lower F-numbers to be achieved,
although this is at some considerable financial cost because these lenses
are far more difficult to manufacture.
In practice, lenses using spot filters ar e often labelled ‘F1.4–64’. This
means that the mechanical iris of fers 10 stops (F1.4–F32) unaided, with a
further 11 stops (F32–F64) available in the filter ring ar ea. Thus the lens
assembly is capable of 21 stops without the apertur e diameter having to
close to an impractical amount.
The F-number can have serious implications when selecting a lens for a
specific application. From Table 4.2 we see that a 50 mm lens with an aperture set at 25mm equates to F2. However, for a zoom lens with a focal length
of 150 mm, a 75 mm aperture would also give an F-number of 150/75 2.
So we see that every incr ease in focal length requires a larger aperture to
maintain a workable lens speed.
In the case of a manual iris lens, the situation can arise where the picture
quality is perfectly acceptable during the daytime, but as soon the light level
begins to fall, the lens is unable to pass suf ficient light to maintain image
72
Closed Circuit Television
quality. Clearly it is impractical to manually adjust the iris every time the
light level changes, and any camera located either outdoors, or indoors in
a situation where there will be large fluctuations in light levels, should be
fitted with an automatic iris lens.
An auto lens is one wher e the apertur e mechanism incorporates some
form of motorized drive which closes the iris when the light level is high, and
opens it as and when it falls, maintaining an optimum light thr oughput for
the camera pick-up. The principle of operation is illustrated in Figur e 4.10.
The video signal emerging from the CCD is sent to an auto iris amplifier circuit, which detects and monitors the average level of the video signal.
Motor
drive
control
Level
d.c. control
Auto iris
amplifier and
control
Peak/Average
To
camera
circuits
Gear
box
Motor
Light input
Video signal
CCD
Lens
Aperture
Average
video
signal
level
Too bright
Acceptable
Too dark
Time
Vd.c.
0
Vd.c.
Time
a
b
c
d
Figure 4.10 Concept of motorized iris control
Whilst the video level falls within a range deemed acceptable, nothing happens. However, should the light level increase, the average video level will
rise, and the circuit will produce a d.c. output to activate the motor drive circuit, and hence the motor. The greater the light input, the higher will be the
d.c. control, and the motor turns faster to close down the aperture.
Lenses
73
This action is illustrated in the time-related graphs in Figure 4.10. At point
(a) the light input becomes too high (large video signal), and the monitor circuit produces a positive d.c. output to operate the iris motor.Between (a) and
(b) the iris is closing, the light level is falling, and the d.c. output fr om the
monitor circuit falls accordingly until at (b), the apertur e setting is corr ect
and the motor stops. Between points (b) and (c) the light level, although
varying, remains within acceptable levels, and the iris remains stationary. At
(c) the light level suddenly falls, and the monitor circuit produces a negative
d.c. which will cause the motor to turn in the opposite dir ection, opening
the aperture. At (d) the light falling onto the CCD is once mor e within
acceptable levels, and the monitor cir cuit removes the negative d.c. control, and the motor stops.
With the arrangement shown in Figur e 4.10, the servomotor, gearbox,
and auto iris amplifier circuit are all incorporated within the lens, which is
commonly known as an AI lens. A connection is made from the camera to
the lens to pr ovide a video signal input to the monitor cir cuit. This link
may use terminal connectors, or it may be made using a plug and socket
arrangement. Some cameras have just one single socket with a selector
switch labelled something like ‘Video/DD Iris’. We shall come to DD in a
moment, but wher e the auto iris amplifier cir cuit is located in the lens
assembly, the selector must be set to the ‘Video’ position.
Although AI lenses incorporating a motor and monitor cir cuit are used,
they are very large and expensive. A less expensive alternative to the motorized drive is the galvanometric drive. The galvanometer is by no means a new
invention; it was used as a movement in very sensitive moving coil measuring meters many years ago, and operates on very similar electromagnetic
principles to the analogue meter movement which is still in use today
.
However, instead of moving a pointer needle, the galvanometer moves the
plates of an iris mechanism. The principle is shown in Figur e 4.11. The d.c.
output is fed to the galvanometer plates, which in turn hold the aperture at
the correct position; however, the d.c. must be maintained otherwise the
plates will fall fully closed. Therefore the signal monitor circuit produces a
varying d.c. as shown. In addition to the r educed cost, two advantages of
employing galvanometric drive in preference to d.c. motors are a reduction
in the physical size of the lens, and a much lower current consumption.
The lens size can be further r educed by incorporating the AI amplifier
circuit within the camera. These compact lens types ar e known as direct
drive (DD) lenses and are widely used in the CCTV industry. Where a DD
lens is employed, the Video/DD selector switch on the camera must be set
to the ‘DD’ position.
The two controls labelled ‘Level’ and ‘Peak/Average’ in the arrangement
in Figure 4.11 give the installer some control over the iris action and thus the
image on the screen. These controls may be found in either the lens or the
camera, depending on where the AI amplifier circuit is located. They ar e
required because there is no camera/lens combination yet devised which
can automatically cope with every possible lighting scenario, and sometimes compromises need to be reached.
74
Closed Circuit Television
Drive
control
Level
d.c. control
Auto iris
amplifier and
control
Galvanometer
coil
Peak/Average
To
camera
circuits
Light input
Video signal
CCD
Aperture
Lens
Average video
signal level
Too bright
Acceptable
Too dark
Time
0V
Aperture
closed
Time
Aperture
open
a
b
c
d
d.c. drive to galvanometer
Figure 4.11 Unlike the motor, a galvanometer requires a constant current to
maintain the aperture setting
The level control performs a similar function to the manual iris ring in
that it allows the engineer to set the sensitivity of the lens. The contr ol is
usually labelled ‘H–L’ or ‘Hi–Lo’. Setting the control towards the high end
tends to pr oduce satisfactory r esults in the day , but the pictur e may be
dark at night. A low setting will allow the iris to open in the dark, producing a clearer picture. However, this may lead to excessive brightness in the
day. Unless the system incorporates a facility wher eby an operator can
control the iris via telemetry, the level has to be set taking account of the
extremes under which the lens will be expected to operate, and the particular requirements of the customer.
The peak/average contr ol, often labelled ‘P–A ’ or ‘Pk–Av’, enables the
installer to introduce a degree of compromise with regard to large differences in lighting levels within the picture area. Consider the two situations
Lenses
75
Peak area
resulting
from a wall light
Figure 4.12a Peak levels allowed to burn out; average levels are viewable
Peak area resulting
from person
Tarmac area
Figure 4.12b Average area appears dark; peak areas are viewable
shown in Figure 4.12. The wall light in Figure 4.12a would cause the iris to
close down, resulting in a very dark image acr oss the rest of the viewing
area. In this situation the peak/ average control may be used to open the
iris up (by adjusting towards the ‘average’ position) in order to give a reasonable contrast across the picture. The compromise is that there will be a
‘hot spot’ in the area surrounding the wall light.
For the situation in Figure 4.12b, if the control were set for ‘average’, the
iris would open fully to bring out the tarmac ar ea, causing the person to
76
Closed Circuit Television
bleach out. The pr oblem would be particularly acute in low light conditions. Adjusting the control towards the ‘Peak’ setting causes the iris to
close down to a point where the tarmac is not really visible; however, the
person becomes more clearly visible on the screen.
You will frequently see another setting on many cameras labelled ‘EI’,
meaning electronic iris. This is a cir cuit within the camera that maintains
the correct video signal level using electr onic means, and has nothing to
do with the iris in the lens. This principle will be discussed in Chapter 6.
The final parameter we need to examine is the depth of field. This is the
range (in distance) in front of the lens where objects remain in sharp focus.
As much as we would like a lens to display everything, both far and
near, within the field of view in perfect focus, this is not possible. Look at
Figure 4.13, where a lens is focused at a particular distance. Items closer
Line of
focus
Distance to object
Depth of field
Figure 4.13 The depth of field is the area on either side of the object in focus
where correct focus is still maintained
and further away than the true focusing distance may also appear to be in
focus. This range of field is called the depth of field.
Depth of field is dependent upon two things: the angle of view of the
lens, and the apertur e setting. When considering the depth of field it is
important to remember that either a wider angle of view, or an increase in the
F-number (reduced size of aperture), increases the depth of field.
In general, wide-angle, short focal length lenses have a very large depth
of field, even when the F-number is low . However, the advantage of the
large depth of field is somewhat of fset by the fact that distant objects
appear very small on the monitor scr een. As focal lengths incr ease, the
lens views narrow and distant objects become larger. However, the depth
of field reduces.
The problem now is that once the lens has been trained onto a tar get,
and the focus has been set for that tar get, as soon as the lens is made to
zoom onto a target at a different distance, the focus has to be re-adjusted.
This is illustrated in Figure 4.14 where a zoom lens has been set to a target
(object A) at distance D1, and the focus has been adjusted accordingly. The
depth of field around that object would be that shown by the dotted lines.
Lenses
77
Object A
D1
Object B
D2
Figure 4.14 When the zoom is altered, the focus will be incorrect
However, when the focal length is re-adjusted to zoom in on object B at a
distance D2, the focus is incorr ect. This is why the mor e powerful zoom
lenses need to be re-focused when moved from one target to another.
Another related problem is when an auto iris, long focal length (telephoto) lens is used where there is insufficient lighting after dark. During the
day the aperture will set to a high F-number , and there will be an acceptable depth of field. However, after dark the auto iris will reduce the F-number
in an attempt to maintain the light thr oughput. When this happens the
depth of field reduces. Of course, if the engineer only visits the site during
the day he may wonder what the customer is complaining about!
Zoom lenses
Fixed focal length lenses are all right for locations where the area we wish to
view is only a few metres square. However, as soon as we attempt to view a
larger expanse such as a car park or shopping mall, then the limitations of a
fixed lens become patently obvious. To attain the field of view a wide-angle
lens must be used, but then everything and everyone appears so small on
the screen that the image is useless for evidential purposes. On the other
hand, if a zoom lens is fitted the field of view is lost meaning that, unless the
lens is motorized and can be operated from a control room, it is of little use.
This in turn means having to employ staff to operate the CCTV system. For
situations where the level of risk and the size of the business where the system is to be installed does not warrant employing full-time CCTV operators, let alone invest in expensive zoom lenses and telemetry systems, it is
possible to obtain a reasonable coverage by installing two cameras at the
same location, one fitted with a wide-angle lens and the other with a zoom
lens. However, the installer must be certain that such a system will pr ovide the level of cover required before selling the idea to the client.
78
Closed Circuit Television
It is common practice to state the performance of a zoom lens in terms of
the ratio of the change in focal length. For example, a lens having a focal
length from 16 mm to 160 mm would be quoted as being a 10:1 zoom lens.
However, just describing the ratio gives no clue as to the range of views provided; i.e., 10:1 could mean 10mm to 100 mm or 15 mm to 150 mm, therefore
manufacturers usually quote the actual focal length range as well as the ratio.
The principle of a zoom lens is shown in Figure 4.15. Within the assembly,
the zoom lens group is the most complex part of the mechanism because the
optical devices within this group have to move in such a way that the image
remains undistorted and in correct focus. It would be very difficult to maintain correct focusing over a large change in focal length, and therefore a manual focus control is provided by making the front lens group adjustable.
Iris Control
Focus Motor
Zoom Motor
Zoom Lens Group
Fixed Lens Group
Front Lens Group
Figure 4.15 Principle of the zoom lens. In a manually operated lens, the two motors
are omitted and two small levers are available to turn the focus and zoom controls
Specifying the optical speed of a zoom lens is not as straightforward as for
a fixed lens because, as shown earlier , the F-number is dependent on the
focal length and aperture, and in this case the focal length is variable. To keep
the specifications simple, the majority of manufactur ers simply state the
F-number for a lens at the two extremes of its focal length. So, for example, a
2
⁄3 6:1 zoom lens having a focal length range between 12.5 mm and 75 mm
may be quoted as having an F-number between F1.2 and F560.
Lenses
79
When a zoom lens is set to its maximum focal length (telephoto), it must
be fixed firmly in position because the slightest movement r esults in a very
large picture shake on the monitor . The longer the focal length, the mor e
acute the problem. Where the camera is fixed to a solid structure there is not
usually a problem; however, when it is mounted on top of a tower, the slightest wind can throw the picture all over the monitor screen when the lens is at
a high telephoto setting. Gyroscopically corrected lenses are available (at
considerable cost). The optics are mounted in a mechanism that moves in
the opposite direction to the motion of the main lens casing. Needless to
say, because of their high cost these devices are not commonplace.
Electrical connections
Correct electrical connection between the various lens drive cir cuits and
the lens is very important. Connection methods, terminology , and drive
potentials can differ between lens types and manufacturers, and an engineer must be able to decipher the wiring diagrams pr ovided in the manufacturers’ technical information.
The most popular lenses for use in smaller installations are either fixed
focal length or varifocal types, employing a DD or anAI iris control. In this
case the only electrical connections between camera and lens ar e for the
iris. Terminal connectors may be used on some cameras; however , many
lenses come with a prefitted connector plug. When using one of these with
a camera from the same manufacturer, it is usually very difficult to get the
connections wrong. However, where a mix of equipment is being installed
it is important to check for compatibility because, although some standardization has taken place with the acceptance of the ‘P Plug’ (illustrated
in Figure 4.16), there are still a number of wiring configurations. It is also
Pin 1
Pin 2
Pin 3
Pin 1
Damping coil
ve
Pin 2
Damping coil
ve
Pin 3
Drive coil
ve
Pin 4
Drive coil
ve
Pin 4
Figure 4.16 Standard connection configuration for the P-Plug. (So called
because it was developed by Panasonic)
80
Closed Circuit Television
essential that, if fitted, the Video/DD selector switch on the camera for iris
control output is correctly set.
A galvanometric iris usually has a four-pin connector. Two terminals connect the d.c. drive voltage from the camera to the actuator coil which controls
the aperture. The other two carry a control voltage to a second coil in the iris
assembly, which is used for damping. The damping coil is necessary because
the galvanometer is a very sensitive device and is pr one to over-reaction to
the drive voltage. Because the iris contr ol is effectively a closed loop servo
(the light input forming the feedback), this would r esult in a cond ition
known as hunting. Hunting is wher e a servo continually over compensates
and the unit that it is controlling, in this case the iris, oscillates. Thus the iri s
control circuit in the camera outputs not only the drive voltage but also a
damping voltage which is fed to the damping coil. The arrangement is
shown in Figure 4.17.
Damping
coil
Drive
coil
Camera
Figure 4.17 Connections between camera and a DD auto iris lens
In cases wher e the iris contr ol circuit is located within the lens,
three connections are required: one for the video signal, one to pr ovide a
12 Vd.c. supply from the camera to power the contr ol circuit, and one for
a common negative (ground) return. Although not a guaranteed standard,
the wiring colours for these functions ar e normally white for video, r ed
for 12 V and black for 0V. A green wire is often found in the lens connector
cable. This is only used wher e the system incorporates fully functional
telemetry-controlled cameras where the operator is able to manually adjust
the iris from the control room. In this case the green wire connects a varying
d.c. control from the telemetry receiver to the lens.
The lens connector cable may have a braided screen, which is intended
to reduce the possibility of RFI entering at this point. In theory this would
be connected to the earthed body of the camera, or an alternative earth
point in cases wher e the camera body is non-metallic. In practice this
screen is often left disconnected because it can produce an earth loop, and
as the connecting lead is so short it is rar e for RFI to be induced. Wher e
interference is proving to be a problem then one test which should be performed would be to connect the screen to earth.
When we come to look at motorized zoom lenses, the electrical connections become a little mor e involved. A zoom lens usually employs
Lenses
81
low-voltage d.c. motors to move the zoom lens gr oup (focal length) and
the front lens group (focus). Typical operating voltages for these motors
can be between 5 and 12 Vd.c., the control voltage being provided by the
telemetry controller, which may vary depending on the type of system
employed. It is important to check that the lenses selected for use with a
control system ar e compatible. For example, if the lens employs 5
V
motors, and the contr ol unit is applying 12 V when zoom and focus is
operating, you will find that the lenses might r eact very quickly(!), but
also fail soon after installation because the motors have burnt out. On the
other hand, operating 12V motors at 5V will mean that the lens reaction is
very slow, if there is any reaction at all.
The motors drive the lenses via gearboxes and must not be allowed to
over-drive the mechanism, otherwise permanent damage may occur . Take
for example the zoom mechanism. When the motor has moved the zoom
lens group to the maximum focal length position, the power to the motor
must be cut immediately. This can be achieved using limit switches. However,
there is no point in simply placing a switch in series with the motor , otherwise once it has opened ther e is no way of applying a r everse voltage to
move the lens in the other direction. The solution is to place diodes across the
switches. A typical circuit arrangement in shown in Figure 4.18.
D1
S1 open in ‘Telephoto’
S2 open in ‘Wide’
S1
A
M
B
S2
Zoom Out
Zoom In
A = 12 V
A = 0V
B = 0V
B = 12 V
D2
Figure 4.18 Limit switch circuit for a zoom motor. The same arrangement can
be used for focus and iris motors
In this circuit, switch S 1 opens when the lens r eaches maximum zoom
(telephoto). When the operator wishes to zoom out again it is necessary to
reverse the motor direction, which means reversing the polarity of the drive
82
Closed Circuit Television
voltage at terminals A and B. The current path will now be from terminal
A, through diode D1, through the motor, returning through switch S2, which
will still be closed. Moving fr om the ‘wide’ position is simply the r everse,
because S2 will now be open, and diode D2 provides the current path.
It is not uncommon to find potentiometers in a zoom lens for
preset
zoom and preset focus. These potentiometers rotate as the zoom and focus
motors are operated, thus deriving a range of voltages corr esponding to
the zoom/focus ranges. The output voltage fr om each of the potentiometer sliders is connected back to the control unit, making possible a facility
whereby the control unit can have programmable preset zoom positions.
A typical potentiometer wiring arrangement is shown in Figure 4.19.
V Supply
Focus Error
Output
5 kΩ
Focus Preset
V Supply
Zoom Error
Output
Figure 4.19 Preset control circuit arrangement
5 kΩ
Zoom Preset
Lenses
83
When setting up the pr eset positions, the engineer or operator moves
the zoom to a desired position (it may be that the pan/tilt is also moved at
this point) and adjusts the focus. This position can then be stored into the
memory of the control unit, which does this by converting the d.c. value
from each potentiometer into a digital value. During normal operation,
when the operator instructs the control unit to move the camera to the preset position, the zoom and focus motors (plus pan/tilt motors) ar e operated until the d.c. fr om the potentiometers r eaches the pr edetermined
value.
Modern control units can memorize mor e than one pr eset position;
indeed, some can store up to one hundred presets. Yet in most cases five is
about the most that is required because, when you think of it, when would
you need one camera to have a hundred zoom positions, and how can an
operator remember where they are anyway?
Lens mounts
A type of lens mount referred to as C mount has been used for many years
in the cine industry, and because these lenses were already in manufacture
they were adopted by the early CCTV industry. The term ‘C mount’ refers
to the type of screw thread on the lens, and the distance fr om the back of
the lens to the CCD pick-up device in the camera.
C mount lenses are comparatively large, and as the size of CCTV cameras continued to r educe, it was necessary to have a mor e compact lens.
Furthermore, the lenses can be one of the most expensive components in a
CCTV system, and as the size of systems increased in terms of the number
of cameras, a more cost-effective lens was called for. It was with these factors in mind that the CS mount lens was developed.
The CS mount lens is much smaller; however, this is not without some
sacrifice in performance. Compar ed to C mount lenses, CS mount types
suffer greater optical distortion because nearly all of the glass ar
ea is
employed, whereas a C mount only makes use of the inner ar ea where
there is minimal distortion.
From an engineer ’s point of view , the main dif ference between these
lenses is the distance fr om the back of the lens to the image device in
the camera. For a C mount this distance is 17.5 mm, whereas for a CS mount it
is only 12.5 mm. In practice this means that these lens types cannot simply
be interchanged because the focal point will not fall onto the pick-up
device.
CCTV cameras are specified as being either C or CS mount, although
some have a mechanical adjustment which moves the pick-up device
between 12.5 mm and 17.5 mm. Where the mount cannot be adjusted, the
following rules apply.
A C mount lens can work on a CS mount camera as long as a special
5 mm adapter ring is fitted to extend the distance to 17.5
mm. This is
84
Closed Circuit Television
Image
device
Lens
Camera
d
d = 17.5 mm for a C mount lens
d = 12.5 mm for a CS mount lens
Adaptor
ring
C mount lens
Image
device
CS mount camera
5 mm
12.5 mm
Use of a 5 mm adaptor to marry
a C mount lens to a CS mount camera
Figure 4.20
illustrated in Figure 4.20. However, a CS mount lens will never work on a C
mount camera because there is no way of reducing the 17.5 mm distance.
Filters
There are occasions where it is useful to fit a light filter to the lens. For
example, where a camera is required to look through a glass windowpane,
a polarizing filter can be fitted to remove the effects of glare or reflections.
This type of filter is fitted to the lens and then r otated until the desir ed
result is achieved.
Lenses
85
Another type of external filter is the ND type, which we discussed earlier
in this chapter when considering the action of the spot type filter , which is
used to increase the possible number of F-stops. Full-sized versions af fecting the entire lens area are available, and in a range of attenuation factors.
However, it is not often in CCTV applications that these would be a permanent fixture to a camera. One practical application of the ND filter is in the
adjustment of back-focus, which will be discussed later in this chapter.
Infrared cut filters are used where it is desirable to remove IR light from
the camera. All colour cameras have an in-built IR cut filter; however ,
some monochrome models are designed to respond to IR light. Sometimes
where there is a lot of infrar ed light in the viewing ar ea it is necessary to
negate this feature because parts of the picture become slightly burnt out.
Fitting an external IR cut filter often results in a picture with an improved
grey scale contrast range. Another application for IR cut filters is when
adjusting the focus. Removal of infrar ed light from the lens enables the
focus to be adjusted using only visible light. This pr oduces a slightly different focusing point, but one which is per haps more accurate for a camera which is expected to rely largely on visible light.
Opposite to the IR cut filter is the IR pass filter , which r emoves the
majority of visible light so that the camera can only ‘see’ infrar ed. In the
case of cameras that are intended primarily for night-time operation using
infrared lighting, adjustment can be performed in the daytime if an IR
pass filter is used to remove the visible light. The reliability of the adjustment can be further improved by adding an ND filter, in addition to the IR
pass filter, to further block visible light.
Lens adjustment
One of the biggest problems in CCTV installations is incorrect adjustment
of the back-focus. This refers to the 17.5 mm/12.5 mm distance between the
back of the lens and the camera pick-up device. Wher e fixed lenses ar e
involved this is not as critical as long as the engineer is able to obtain a correctly focused image. However, when a zoom lens is employed, if the backfocus is not correctly adjusted, the operator will find that the focus moves
out every time the zoom function is used.Another problem can be that the
zoom lens functions satisfactorily during the day when the auto iris is
closed down, but when the light begins to fade and the iris opens, the
focus becomes poor due to the reduced depth of field.
We already touched on this pr oblem earlier when we discussed depth
of field, and the ideal solution is to perform back-focus adjustment at dusk
when the adjustment is far mor e sensitive. However, in reality this is not
always practical, and therefore another method is to fit an external ND filter to the lens whilst performing the adjustment – the filter simulating twilight conditions by reducing the amount of light input.
86
Closed Circuit Television
There is more than one acceptable method for obtaining corr ect backfocus adjustment, and engineers will adopt their own pr eferred routines
as they become proficient at this. However, when starting out, the generally accepted method for adjusting the back-focus on a camera which is fitted with a zoom lens is as follows.
1. Manually open the iris, or fit an ND filter
, or work in low light
conditions.
2. Select a target at the maximum operational range for the particular
camera/lens (a target with a lot of detail such as a wall is ideal).
3. Adjust the lens focus to ‘far’.
4. Set the zoom to maximum wide angle.
5. Move the back-focus adjustment on the camera forwar ds and back
until optimum focus is obtained.
6. Set the zoom to full telephoto.
7. Adjust the lens focus for optimum focus.
8. Set the zoom to maximum wide angle.
9. Re-adjust the back-focus for optimum focus.
10. Repeat steps 4 to 10 until optimum focus is obtained at all points
between wide and telephoto.
In some cases the back-focus is fixed with a locking scr ew or nut. Be sure
to slacken this off before commencing adjustment, and be sure to secure it
again afterwards, otherwise vibrations from the zoom or PTZ action may
move the back-focus adjustment.
From a practical point of view , it has been assumed all along that the
engineer is looking at the pictur e on a monitor whilst performing focus
adjustments. However, this is not always as straightforwar d as it might
seem, especially when working on a camera assembly mounted atop a 7m
tower. There are a number of ways of overcoming this issue. One of them
is to adjust the back-focus befor e mounting the camera, per haps even in
the workshop. When performing the adjustment in this manner , remember to reduce the lighting somehow , otherwise you might find yourself
performing the adjustment again once the camera has been installed.
Another approach is to use a portable, battery-powered monitor. There
are test monitors made specifically for this purpose, and they can pr ove
very useful in the field for performing all manner of tests and adjustments. Difficulties are sometimes experienced when using these in dir ect
sunlight, as viewing can be somewhat difficult.
Another device that is useful for performing focus adjustment is the
focus meter (Figure 4.21). This is designed to connect to the video output
from the camera, where it analyses the high-frequency video components
in the signal. The sharper the focus the greater the amount of h.f., and thus
if the engineer moves the focus adjustment to one end of its travel and
then progressively moves it back once again, the meter will indicate an
increasing h.f. content up until the point wher e the optimum focus point
has been passed, whereupon the indication will begin to fall off. The engineer is therefore able to set the focus adjustment for the peak indication.
Lenses
87
Figure 4.21 A focus meter. (Courtesy of NG Systems Ltd)
Lens finding
Earlier in this chapter we saw how we can calculate the required lens size
for a given application; however , there are alternatives to using mathematics and an electronic calculator. Perhaps the simplest of these is to use
one of the lens calculators which ar e often available fr ee of charge from
lens manufacturers or suppliers. These differ slightly in both the way that
they function, and the data they pr oduce, but they are simple to use, ar e
quite accurate and are adequate for the majority of applications. A typical
lens calculator is illustrated in Figur e 4.22. This particular type r equires
you to know the primary target height or width, and the distance from the
camera to the target. By adjusting the cursors on the calculator , the focal
length and the horizontal or vertical angle of view can be found for any
given camera format size.
Another method is to use one of the look-up tables pr oduced by various lens manufacturers; however, these often quote the horizontal angle of
view produced by a range of lenses when fitted to dif
ferent camera
88
Closed Circuit Television
Figure 4.22 Typical lens calculator
formats. In this case the engineer still has no idea of the object width or
height for any given distance fr om the lens, but this can be calculated
using trigonometry:
W d tan where W is the width of ar ea across the image, d is the distance of object
from lens, and is the horizontal angle of the lens.
Example
A look-up table states that a 30 mm, 1⁄2 format lens has a horizontal angle
of approximately 11°. Find the image width at a distance of 40 m.
Lenses
89
Solution
W 40 tan 11°
40 0.19
7.6 m
Of all the tools available for lens finding, per haps the most ef fective are
the optical viewfinder types, an example of which is shown in Figure 4.23.
The engineer positions himself wher e the camera is to be located and
Figure 4.23 Optical device used for determining the required lens focal length
for a given field of view, target distance and camera format size. (Photo courtesy
of CBC (Europe) Ltd, manufacturers of Computar lenses)
looks through the device, adjusting it until the desired image is obtained.
Calibrated markings on the side of the device enable the focal length to be
read off. The scales include the various format sizes.Apart from removing
the need for any calculations, another bonus of using an optical viewfinder
is that the customer can be shown what he is getting before agreeing to the
installation. Indeed, the customer can set the viewfinder to show the engineer/surveyor what he requires, and if the engineer is seen to log the agreed
lens size at the time, he need have no worries about comeback from the customer at a later stage in the installation.
5 Fundamentals of television
Before we can look at the operation of cameras and monitors, it is important to have an understanding of the makeup of the television picture, and
the signals required to produce the picture. Later, in Chapter 7, we will
examine the various picture display device technologies.
Development of television really began during the 1930s with a number
of ideas being tested, with differing degrees of success. However, all of the
ideas had one thing in common: they were designed to produce a picture
using a cathode ray tube (CRT), and consequently modern analogue television systems are still tied to a video signal waveform that is designed to
drive a CRT, even when a flat panel display device of some type is being
used. Of course, digital signal transmission systems ar e replacing analogue in both the broadcast and CCTV arenas but, nevertheless, analogue
signals will still be employed for a number of years to come because of the
large amount of analogue equipment still in use and the prohibitive cost of
replacing all of this overnight (not to mention the fact that manufacturers
could not pr oduce the amount of equipment necessary for a sudden
upgrade in such a short period of time).
Producing a raster
A raster is the term given to the blank white display produced when a CRT
is made to scan its screen without any video signal input. It is constructed by
making the spot produced by the electron beam in a monochrome CRT (or
three spots in the case of a colour CRT) deflect vertically and horizontally at
high speed across the screen. The spot is moving so fast that it is indiscernible to the eye, and hence the brain is tricked into thinking that it is see
ing a solid white display. The principle of sequentially scanning the screen to
produce a raster is illustrated in Figure 5.1. For simplicity a raster with only
nine lines has been drawn, although, as we shall see later,both broadcast TV
and CCTV in the UK use a 625-line raster (525 lines for NTSC systems).
The horizontal (line scan) speed is made to be many times faster than the
vertical (field scan) speed, and so the spot zigzags its way down the CR T
screen. However, the line scan period from left to right is much longer than
the line flyback period, where the spot moves rapidly fr om right to left.
Similarly, the field scan period is many times longer than the field flyback
period, where the spot is made to return very quickly to the top of the screen.
During both the line and field flyback periods the electr on beam is cut
off, and so only the solid lines number ed 1 to 9 in Figur e 5.1 are actually
Fundamentals of television
91
Line scan period
Line flyback period
Field flyback period
1
2
3
4
5
6
7
8
9
Figure 5.1 Nine-line sequentially scanned raster
displayed. This is why the flyback periods have been shown dotted in the
illustration. The line scan period is often referred to as the active line period.
The electron beam in a CR T is deflected by passing curr ents through
scanning coils that are wrapped around the CRT neck. During the flyback
periods the currents are reversed to produce an equal but opposite magnetic field. This cannot be done instantly because of the inductive ef fects
of the scan coils – hence the r eason for having to turn of f the beam and
wait for flyback to complete.
So we see how a nine-line raster is produced using sequential scanning.
But what scanning speeds would we r equire to produce a practical 625line or, for NTSC, a 525-line raster? The answer to this question is not as
straightforward as it may appear . During the development of the television system there were a number of factors the designers had to take
into consideration when determining the scanning speeds. Primarily these
were the picture rate, the picture resolution, the system bandwidth, and the
problem of picture flicker.
It must be r emembered that the television system had to be designed
around the existing cinema system, because at the time it was accepted
that much of the transmitted material would be taken from movie sources.
Even television newsreel was sourced on celluloid film (there were no camcorders in the 1940s!). So, before we can begin to look at the rationale for constructing television pictures using a CR T, we need to look briefly at the
structure of the source material that was to come from the cinema industry.
The ‘moving pictures’ we see at the cinema ar e actually a series of still
photographs being flashed onto a screen at high speed. Early systems used
92
Closed Circuit Television
a frame rate of 16 pictures per second, but this produced poor quality and
jerky pictures, and a 16 Hz brightness flicker was very discernible. It was
generally accepted that a pictur e rate of between 20 Hz and 25 Hz would
cure the pro blem of jerky pictures, but the eye could still r esolve some
flicker at these rates. This problem could only be over come by increasing
the picture rate to something in the order of 45 Hz to 50 Hz, but this meant
a doubling of film consumption. The solution was (and still is to this day)
to block the light output from the projector lamp not only whilst the frame
is being pulled into the gate, but also once again whilst it is being held stationary in the gate. Thus each frame is flashed onto the scr een twice per
second. A picture rate of 24 per second was decided on as the industry standard, and so the cine flicker rate is 48 Hz, which is indiscernible to the eye.
The flicker is a product of how the eye functions. The eye has a characteristic referred to as the persistence of vision. Ther e is a time delay between
a part of the retina being excited by a light input, and the retina output fading
once the light input has been removed. In the case of the 24Hz cine picture
rate, the eye was able to distinguish the projector light output being strobed
as the frames were advanced. By strobing the lamp at twice the rate, the
retina is unable to respond quickly enough for the flicker to be resolved.
Because the UK mains fr equency is 50 Hz, there were reasons relating
to receiver design why a picture rate of 25 per second rather than 24 was
chosen for the television system. (In the USA, where the mains frequency
is 60 Hz, the picture rate is 30 per second.) However , displaying the pictures at a rate of 25 /30 Hz using sequential scanning would pr oduce the
same flicker pr oblems as a cine film flashed at 24 Hz. To overcome the
problem of picture flicker in television, a technique known as interlaced
scanning was developed. The principle of interlaced scanning is illustrated
in Figure 5.2 where, again for simplicity, a nine-line raster has been shown.
Start of second field scan
Start of first field scan
1
2
3
4
5
6
7
8
9
Figure 5.2 Nine-line interlaced raster
Fundamentals of television
93
In this case the spot begins at the top centre of the screen and makes its
way down, finishing at the bottom right-hand corner . The field flyback
now returns the spot to the top left-hand corner, where it once again scans
downwards to end in the bottom centr e. During the first vertical scan
period, known as the odd field, the odd TV lines are scanned, and 50% of a
picture frame is displayed. During the even field period when the even line
numbers are scanned, the other 50% of the frame is reproduced. So we see
that two TV fields equals one TV frame. The principle is illustrated in Figure 5.3
where for simplicity an eleven-line raster has been used.
Odd field (312.5 lines PAL)
(262.5 lines NTSC)
Even field (312.5 lines PAL)
(262.5 lines NTSC)
1 TV frame (625 lines PAL)
(525 lines NTSC)
Figure 5.3 Two interlaced fields build up a complete TV picture frame
The flicker is eliminated because all ar eas of the scr een are being illuminated at a rate of 50Hz (60 Hz NTSC). At a normal viewing distance the
eye cannot discern the individual TV lines, and so just as one line would
be appearing to fade, two more are scanned, one on either side.
In the PAL UK TV system (see later in this chapter), the vertical scanning rate is made to be 50Hz, twice the 25Hz picture rate. Thus during the
first field scan the spot r eproduces the odd 312.5 lines, and during the
second field scan the even 312.5 lines ar e reproduced. So in 1/25th of a
second (i.e., 40 ms or 25 Hz), a complete picture is built up on the screen.
For the NTSC system, the vertical scanning rate is made to be 60 Hz,
twice the 30 Hz picture rate. Thus during the first field scan the spot reproduces the odd 262.5 lines, and during the second field scan the even 262.5
lines are reproduced. So in 1/30th of a second (i.e., 33 ms or 30 Hz) a complete picture is built up on the screen.
Returning to the question, ‘what scanning speeds would we r equire to
produce a practical 625-line raster?’, having dealt with the issues of picture rate and pictur e flicker, let’s now see wher e picture resolution and
system bandwidth come in.
Picture resolution
This relates to the definition of a TV picture, and can be expressed in terms of
either vertical or horizontal resolution. The vertical resolution is determined
94
Closed Circuit Television
largely by the number of lines that make up the picture. The horizontal resolution is determined by the number of black and white elements that the
system is capable of displaying along any line. Ideally, the maximum horizontal resolution for a television system should be at least equal to the vertical resolution – that is, wher e the width of each black or white squar e is
equal to the thickness of one TV line. This display situation is illustrated in
Figure 5.4.
Picture
elements
TV
lines
585
elements
780 elements
Figure 5.4 Deriving horizontal resolution for a 625-line picture. (Pixels are not
drawn to scale as they would appear too small)
Although the vertical resolution is determined by the number of TV lines
in the picture, it is also dependent on the position of the camera in relation to
the image in view. The more horizontal lines a system has, the more squares
we can reproduce. However, it does not necessarily hold that a 600-line
raster would always give a 600-line resolution. Looking at Figure 5.4, consider what happens when the camera is positioned such that each line is
scanning equally between two sets of squar es. The line cannot be black
and white at the same time, and so the entire checkerboard would appear
mid-grey. This is an extr eme situation, but it does make the point that in
any picture there will almost certainly be some loss in vertical resolution.
Ignoring widescreen television, which has yet to make an inr oad into
CCTV, all monitor screens have an aspect ratio of 4:3. That is, although the
screen size of any monitor, e.g. 30cm, 51 cm, is measured diagonally, the ratio
of the sides is always 4 wide by 3 high. Because the horizontal r esolution
must be at least equal to the vertical r esolution, for any given number of
Fundamentals of television
95
TV lines (vertical resolution) the horizontal resolution will be the number
of TV lines multiplied by 4/3. Let’s see how this works for the 625-line
system employed in UK broadcast and CCTV.
As we have seen, a 625-line raster is constr ucted from two fields, each
having 312.5 lines. The period for one field scan is 20 ms; however, this
includes the flyback period which is (appr oximately) 1 ms, during which
time the electron beam is cut of f to prevent it being seen moving up the
screen. In 1 ms approximately 15 horizontal lines will have been scanned,
but the beam is switched off for 20 lines to ensure a clean flyback. In other
words, out of the 312.5 lines per field, only 292.5 are active.
This means that over two field periods 40 lines ar e unused and so, out
of 625 lines, only 585 actually contain picture information. Looking again
at Figure 5.4 we see that if the vertical resolution is 585 lines, then the horizontal resolution should be at least 585 4/3 780 pixels.
We can now turn things around and look at them from a different point of
view and ask the question: why was 625 chosen for the number of lines? It
was derived from the decision to begin the calculations with horizontal resolution 780 pixels. So how was the figur e of 780 arrived at? This is the ma ximum frequency of the signal drive voltage required at a CRT to reproduce
the checkerboard display, a portion of which is shown in Figure 5.5.
TV line
CRT off
CRT cathode
drive voltage
CRT on
t = 133 ηs
Figure 5.5
In the UK TV system, the horizontal scanning speed is set at 64s; 52 s
for the scan period (also referred to as the active line period); and 12s for
the flyback period. From the periodic time of 64 s, we can derive the
scanning frequency:
frequency 1
1
15 625 Hz
periodic time
64 s
For a display device to switch on and of f quickly enough to reproduce
780 pixels during the 52 s active line period, the periodic time for one
cycle of the signal would be 133 ns. In terms of fr equency this equates to
7.5 MHz. In other words, in order to reproduce a horizontal resolution of
780 pixels, the system including everything from the camera to the display
device would require a bandwidth of 0 Hz to 7.5 MHz. This exceeds the
96
Closed Circuit Television
5.5 MHz bandwidth originally allocated to broadcast TV in the UK, but it
was found that the bandwidth could be r educed by applying something
called the Kell factor. This is a figure of 0.7, which was derived as a result
of extensive work in 1933 by Ray Kell who, after performing many viewing tests, concluded that a reduction in horizontal resolution of 0.7 would
not produce an appreciable deterioration in picture quality.
Applying the Kell factor r educes the bandwidth to a mor e practical
figure of 5.5 MHz, and so it is that the UK TV system has a video bandwidth
of 0–5.5 MHz, which equates to a resolution in the order of 546 pixels.
For NTSC the horizontal line fr equency is 15 734 Hz, which produces a
line period of 63.6s. The horizontal flyback period is approximately 10.3 s,
which gives an active line period of ar ound 53.3 s. The vertical flyback
period is typically 15 lines per field (30 lines per frame), r esulting in 495
active lines per TV frame. Ther efore, applying the r ule that the horizontal
resolution should be at least equal to the vertical resolution, an ideal system
should be capable of reproducing 495 4/3 0.7 (Kell factor) 462 pixels
along one line. This would equate to a video signal bandwidth of 0– 4.3 MHz.
In practice the video bandwidth for NTSC is specified as being 0– 4.2 MHz,
which would produce a horizontal resolution of around 450 pixels.
It is more common to expr ess horizontal resolution in terms of ‘television lines’ or TVL. This measurement is related to the number of horizontal pixels that we have just been looking at, and will be discussed in
Chapter 6.
It must be remembered that the figures quoted above are for broadcast
television and, whilst in times gone by , CCTV has in general fallen short
of these specifications, with modern technology cameras and monitors,
coupled with fibre-optic or other broadband transmission techniques, CCTV
is actually capable of producing a superior image resolution to broadcast
television, which is dogged with bandwidth restrictions. However, it must
be stressed that to achieve a high r esolution in CCTV, everything fr om
the lens to the monitor must be of a high specification and the installation
must be sound; one weak link in the system, for example poor cabling with
a multiple of connectors, or a poor -quality lens, will r esult in an overall
degradation of resolution.
Synchronization
The image which has been focused onto the pick-up device by the lens is
scanned at a rate of 50Hz, and 15 625 Hz (PAL). It is essential that the monitor scans at the same rate as the camera, not just in terms of frequency, but
also in terms of the precise position of the spot on the screen. (Remember
that, even though modern solid-state devices such as camera CCD chips
and flat panel monitors do not actually scan, the analogue video signal
must still conform to the CCIR standards that relate to the old camera tubes
and CRT monitors. Hence the reason why we still talk in terms of scanning.)
Fundamentals of television
97
Consider the two conditions illustrated in Figure 5.6. In condition A, the
electron beam in the monitor begins to scan a field at pr ecisely the same
time that the camera begins a field scan, so the displayed pictur e appears
on the screen in exactly the same position as it would if you were looking
directly through the camera lens.
Relative scanning positions
Condition A. Monitor and
camera synchronized
Camera pick-up
Monitor display
Relative scanning position
Condition B. Monitor and
camera out of sync.
Figure 5.6 Effect on the display when camera and monitor are unsynchronized
In condition B, the beam in the monitor CR T is in the centr e of the
screen when the camera begins a field scan. In this case the top left corner
of the image is displayed in the centre of the screen, with a corresponding
displacement of the r est of the image. The vertical and horizontal dark
stripes appear as a r esult of the camera flyback periods being displayed;
remember that during the flyback period the beam is cut of f, and so the
camera outputs a black level signal.
The monitor must be synchr onized to the camera , and to achieve this the
camera generates a series of pulses which are added to the video signal at
the output. At the end of each horizontal scan, at the instant that the camera is initiating line flyback, a line synchronization (sync) pulse is generated
by the camera and added to the video signal. Likewise, at the end of each
field scan a field sync pulse is added to the video signal. Thus, during one
20 ms field period the camera will output 312 line sync pulses and 1 field
sync pulse. During one complete TV frame (40 ms) a total of 625-line, and
two field sync pulses will be output from the camera.
The shape and timing of the line and field sync pulses is complex, and
it is beyond the scope of this book to look into the r easons behind this
complex make-up. However, it is important that a CCTV engineer can recognize these signals when they are viewed on an oscilloscope.
98
Closed Circuit Television
The line sync pulse is shown in Figure 5.7. The monitor will initiate line
flyback at the instant the first falling edge, immediately following the
front porch, of the pulse appears. Note that the total duration of the pulse
is 12 s, which is equal to the line flyback period. The porches are set to be
at a level in the video signal waveform equal to black to ensur e that the
beam is cut off during the flyback period. This is called the blanking period.
12 s
1.4 s
4.7 s
5.9 s
Back porch
Front porch
Sync pulse
Figure 5.7 A line sync pulse. The timings relate to a 625-line system. For a 525line system, the blanking period is 10-s, with similar proportions for the porch
and sync periods
When viewed on an oscilloscope at the output from a camera, the field
sync pulse actually appears nothing like a pulse at all. This is because it is
made up from a series of pulses beginning with five equalizing pulses, followed by five pulses which we refer to as the field sync period, and finally
five or four (depending whether it is the odd or even field) more equalizing
pulses. At the monitor, after separation fr om the video signal, this series
of pulses is passed through a low-pass filter which integrates them into a single pulse, which is the field sync pulse. The complete field sync period is
shown in Figure 5.8.
Following the field sync period, a series of black lines is sent out. These are
essential to ensure that the beam is cut off during the field flyback period.
As discussed previously, there are twenty lines in the field-blanking period (15
for NTSC). In a UK br oadcast transmission, it is during this period that
Teletext information is transmitted and, although this might not appear
significant for CCTV engineers, the idea of transmitting data during the
Fundamentals of television
99
White
level
Video signal for the last
line of a field
Video signal for first line
of the following field
Black level
Line sync pulses
Equalizing Field sync Equalizing
pulse
pulses
pulses
Field flyback period
(20 lines approx)
Figure 5.8 Field synchronizing and flyback period
field flyback period has been taken up by the majority of manufacturers of
CCTV telemetry control equipment, and camera control data is sent out during this period in a very similar way to broadcast Teletext data. This will be
considered further in Chapter 10.
Synchronization is very important and any loss, distortion or attenuation
of the sync pulses will result in vertical jittering or rolling, and/or horizontal pulling or rolling. Such problems are all too common in CCTV installations, and can be the result of many things. To mention but a few, there can
be different earth potentials between various points in the system, faulty or
incorrectly installed cables and/or connectors, incorrect cables or connectors, poor-quality camera switching units which interrupt the sync signals,
and incorrect positioning of terminator switches on monitors or other equipment. All of these are covered in the relevant chapters in this book.
The luminance signal
The luminance signal, usually abbr eviated luma and represented by the
symbol Y, is the black and white information required to satisfy the rods in
the eye. It contains information relating to brightness and contrast changes.
A monochrome camera outputs a luma signal, plus sync pulses. A colour
camera produces the luma component by adding red, green and blue in the
correct proportions.
100
Closed Circuit Television
It can be shown that one unit of white light is made up fr om proportions of red, green and blue light following the expression:
1Y 0.3R 0.59G 0.11B
The signal pr ocessing stage in the colour camera employs a matrix
which adds the R, G and B signals in these proportions to produce a luma
component.
A common test pattern used in television is the eight-bar colour display
.
The order of the colours is such that when the colour is removed, the bars
appear with white on the left and black on the right, and descending order
of grey in between. This is termed the grey scale. The luminance signal
required at a CRT cathode to pr oduce this display is termed the staircase
waveshape, and a quick look at Figur e 5.9 r eveals how the name was
derived. Bear in mind that a single stair case produces only one TV line,
V
White
level
70%
Black
level
30%
0
time
52 s (53.3 s NTSC)
12 s
(10.3 s NTSC)
Figure 5.9 Staircase luma signal which, when applied to a display device,
produces eight grey scale bars
Fundamentals of television
101
and to produce a complete frame a total of 585 of these, plus 40 black lines
for field flyback blanking, are required.
The CCIR standard for transmission of luminance signals along a cable
requires that the voltage level should be 1 Vpp when measured from the
peak white to sync tip levels, and when the input/output impedances of
the equipment are 75 . In this 1 V signal, the pictur e content will be no
more than 0.7 V (70% of the total signal level), and the sync pulse will be a
constant 0.3 V (30% of the total signal level).
The chrominance signal
The chrominance, or chroma (C), signal containing the colour information
is far more complex than the luma component, and although it is not necessary for the CCTV engineer to be fully conversant with the theory , there
are some essential features which he must be aware of.
The colour system employed in the UK is known as the PAL system, PAL
being the abbreviation for phase alternating line. PAL evolved out of the earlier American NTSC (National T elevision System Committee) colour television system, which is used throughout the USA, Canada, Japan and parts
of South America. A forerunner to PAL is the French SECAM (sequential
couleur avec memoire) system; however, this is not widely used throughout
the world. None of these three television systems is directly compatible and
engineers must be aware of this when they ar e confronted with equipment
that has switches (either mechanical or menu-driven) which allow different
systems to be selected. For example, a DVR switched to operate in NTSC
mode will record in black and white with synchronization errors if it is connected to a system comprising PAL cameras and control equipment.
In order to produce a colour pictur e at the monitor , it is not necessary
to transmit all three primary colours. This is because the Y signal is already
being transmitted, and r emembering that 1Y 0.3R 0.59G 0.11B, we
can see that if we transmit any two of the primary colours, the thir d can be
recovered by matrixing the other two and theY signal. For example, if we do
not transmit the gr een, a matrix cir cuit in the monitor can be made to perform the function G 1Y 0.3R 0.11B. Any of the three colour signals can
be omitted, but green was chosen because, after it has been processed in the
PAL colour encoder, it is the smallest of the thr ee colour signals and is thus
more prone to being swamped by noise during the transmission process.
To aid transmission of the r ed and blue colour signals it is necessary
to reduce their amplitude. Because the luma is sent independently of the
colour, it is possible to achieve this reduction by removing the luma content
from the red and blue signals, pr oducing colour difference signals, R-Y and
B-Y. However, further attenuation of the colour difference signals is necessary, and thus the camera applies a weighting factor to each of them. The
weighted signals are referred to as u and v. The u signal contains the B-Y
information, and the v signal contains the R-Y information.
102
Closed Circuit Television
In order to transmit the u and v signals they must be modulated onto a
carrier. Modulation is a process where two signals are added together in a
certain way to enable them to be transmitted either thr ough space, or
along a cable. One signal is the desir ed information, and the other is a
high-frequency carrier which is used to ‘transport’ the information to the
receiving equipment, after which it is dispensed with. The principle is
illustrated in Figure 5.10.
Information
signal
High-frequency
carrier signal
Amplitude-modulated
carrier signal
Figure 5.10 Principle of amplitude modulation
For the PAL colour system the carrier, known as the colour subcarrier, has
a very precise frequency of 4.433618 75 MHz, although it is usually referred
to as the 4.43 MHz subcarrier. Similarly, for NTSC, the 3.579 545 MHz
colour subcarrier may be r eferred to as the 3.58 MHz subcarrier. These
exacting figures were chosen to ensure that the colour signal would of fer
minimum interference with the high-frequency luma components because,
as we can see in Figure 5.11, the colour subcarrier is positioned within the
luma passband.
The name PAL was given to the system because of the fact that thev signal
changes its phase by 180° at the end of every TV line. This is unique to the
PAL system and is employed to pr ovide a built-in corr ection for phase
errors which occur in the chroma signal as it passes through the transmission medium. Such phase errors would result in noticeable colour errors.
A colour monitor or VCR/DVR, for reasons which are beyond the scope
of this textbook, uses a crystal oscillator to generate a pr ecise 4.43 MHz
(3.58 MHz NTSC) subcarrier. However, this oscillator must not only produce
an accurate frequency; its output must be in exactly the same phase as the
subcarrier coming from the camera. To achieve this a chroma burst signal is
generated in the camera. This burst comprises ten cycles of the subcarrier
and is placed onto the back porch of each line sync pulse. The chroma burst
can be considered to be a sync pulse for the colour processing circuits in the
monitor or other processing equipment and, if for any reason it is lost, the
decoders in all equipment will default to black and white mode of operation.
If a camera were producing a picture of the standard eight-bar colour
display, then the chroma signal, when viewed on an oscilloscope, would
appear something like that shown in Figure 5.12. During the periods of the
white and black bars ther e is no colour signal. Between the yellow and
Fundamentals of television
103
Luma signal spectrum
Chroma signal spectrum
Y and C components
overlap causing cross-modulation
Chroma subcarrier component
PAL transmission
f MHz
3.33
4.43
5.53
Chroma signal bandwidth 2.2 MHz
Chroma subcarrier component
NTSC transmission
f MHz
2.08
3.58
5.08
Chroma signal bandwidth 3 MHz
Figure 5.11 Frequency relationship between the luma and chroma signals (PAL
and NTSC transmissions)
blue bars the amplitude-modulated subcarrier is pr esent. Note also the
chroma burst signal on the back porch of the line sync period.
Television signals
When it comes to sending the luma and chr oma signals from the camera
to the monitor there are a number of options available. In the CCTV industry, perhaps still the most common method is to use composite video, where
the luma and chr oma are sent simultaneously along the same co-axial
cable. The prime advantage of this is the low cost compared with the other
104
Closed Circuit Television
Standard colour bar display
White
Yellow
Cyan
Green Magenta
Red
Blue
Black
Chroma burst
signal
4.43 MHz carrier modulated
with chroma signals
52 s
12 s
Figure 5.12 Chrominance signal in relation to the standard colour bar display
(PAL system)
options available, because only one co-axial cable is r equired from each
camera, and between each item of equipment in the control room.
When the two signals ar e mixed, the chroma signal tends to sit on the
luma voltage waveshape. If we add the chroma signal shown in Figure 5.12
to the luma staircase shown in Figure 5.9, then we have a composite video
signal. This addition is illustrated in Figure 5.13, where the chroma signals
for the six colour ed bars sit on the corr esponding luma steps. Note the
chroma burst signal positioned on the back por ch of the horizontal sync
signal. This figure also raises another point: it was stated earlier that the
CCIR standard for video signals is 1 Vpp into 75 , yet when a colour signal is viewed on an oscilloscope, the signal often appears higher than this,
even though it is corr ect. The r eason for this is that the CCIR standar d
relates to the level of the monochr ome (luminance) signal which, as we
saw from Figure 5.9, is measur ed between the sync tip and peak white
Fundamentals of television
White
Yellow
Cyan
Green Magenta
Red
Blue
105
Black
Chroma burst
signal
52 s
12 s
Figure 5.13a Composite video signal where the modulated chroma carrier sits
on top of the luma signal
levels. When the chrominance signal is added, the blue and red components
frequently exceed the sync tip level. In conclusion, when using an oscilloscope to take an exacting measur ement of a video signal waveform,
a monochrome signal should be used. Alternatively, if a colour bar display
is to be used, be prepared to accept a signal level in the order of 1.8 Vpp.
106
Closed Circuit Television
Figure 5.13b Composite video signal as it would appear when viewed on an
oscilloscope
The disadvantage of composite video is the cross-modulation effect. This
is where luma signal components at fr equencies around 4.43 MHz mix
with the 4.43 MHz chroma signal components (3.58 MHz NTSC). Once
they are mixed (composite), analogue VCRs and monitors ar e unable to
separate them, and the displayed picture contains interference. This interference manifests itself as a colour ed patterning in ar eas of the pictur e
where there is high-resolution luma. A classic example is frequently seen
on broadcast TV when a person appears wearing a suit or jacket with a fine
pattern, and at certain distances from the camera a rainbow effect appears
over the person’s clothing.
Cross-modulation can be avoided by keeping the luma and chroma signals separate at every point between the camera and the monitor . This
method of signal transmission is known as Y/C.
Many CCTV cameras, VCRs and monitors, and other contr ol equipment
have optional Y/C provision through the four-pin S-VHS socket (Figur e
5.14) which carries the luma on one pair of conductors (pins 1 and 3), and
the chroma on another (pins 2 and 4). Each pair of conductors is an individual co-axial cable, ensuring that the luma/chr oma signals do not mix.
This is the good news. The bad news is that to make the system ef fective
every camera requires two co-axial cables, and to further complicate things,
adaptors are required to marry the large BNC connectors to the very small
Y/C connectors, which in reality were never intended for use in CCTV system building. In practice it is generally consider ed to be too expensive and
impractical to use S-VHS in a complete CCTV system installation, and use
of these connectors is confined to such things as VCR input/output, etc.
In CCTV wher e picture resolution is of paramount importance, the
losses incurred by cross-modulation are unwelcome; however, there is no
Fundamentals of television
3
107
Pin 1 Luma screen
4
Pin 2 Chroma screen
Pin 3 Luma signal
Pin 4 Chroma signal
1
2
Figure 5.14 S-VHS connector pin configuration
cost-effective way of avoiding them in analogue transmission systems.
Where analogue signals are later digitized, there are methods of detecting
and largely cancelling cross-modulation effects at the contr ol room, but
this equipment is still expensive and is only feasible for lar ger installations. On the other hand, wher e the video signal r emains in the digital
domain from camera to control room, and S-VHS connectors are employed
for monitor connection, the story is somewhat different.
Digital video signals
The concept of converting analogue signals into digital form is not new .
In fact, like many of the principles employed in modern computing, the
complex mathematics relating to the pr oblem were resolved many years
before the technology was available to build equipment capable of performing the operations.
We are all now familiar with digital audio, which has been around for a
few decades in the form of various tape formats, and for many years in the
form of compact disk. Yet digital video took much longer to make an appearance. There is a good reason for this. A digitized video signal amounts to
many times more data than an audio signal of any equivalent time period,
and even with the high-capacity har d drives of today’s computers, without the application of the compression techniques used with digital video,
an 80 Gbyte hard drive would store no more than a few minutes of video
information. It was the development of video compr ession techniques
that made digital video possible, but it has taken a number of years for this
compression to be perfected, and for chip sets to be developed and manufactured in sufficient quantities to make them commercially viable.
The process of analogue-to-digital conversion is very complex, but a
simple overview is shown in Figure 5.15.
108
Closed Circuit Television
V
1.0
256
levels
a
0.3
t
0
Sample period
Clock
signal
t
Data out
Video signal
input
Store
ADC
Clock
Figure 5.15 Principle of analogue-to-digital conversion
The video signal is fed into the cir cuit and the switch closes for a brief
moment every time a clock pulse is present. For this illustration, we must
assume that the capacitor is capable of instantly char ging to the instantaneous voltage level of the video signal. When the first clock pulse arrives,
the switch closes and the capacitor acquir es a charge. In the brief period
between the first and second clock pulses, theanalogue-to-digital (A/D) converter measures the potential on the capacitor, and assigns an eight-bit binary
word which corresponds to this level. When the second clock pulse arrives,
a new voltage level corresponding to the video signal level at that instant is
stored in the capacitor. The A/D converter now assigns an eight-bit word
for this level, and the cycle repeats.
An eight-bit binary word has 256 combinations between 00 000 000 and
11111111, meaning that it is possible for the video signal to have 256 voltage
levels. The A/D converter functions by measuring the voltage in the store
capacitor, and assigning the binary wor d that corresponds to the closest
of these 256 voltage values. Note that these levels do not include the sync
components because the A/D converter only samples the 0.7 Vpp video
Fundamentals of television
109
White
Portion ‘a’ of the video
waveform shown in figure 5.15
Black
Clock signal
8
8
8
8
8
Each level is converted into
an 8-bit binary word
D/A conversion produces pixels
rather than a true analogue
waveshape (the dashed line
represents the original
signal shape)
Figure 5.16 Process of analogue to digital/digital to analogue conversion
signal components. The sync pulses have a constant level, fr equency and
duration and therefore may be represented by a simple data string.
Of course, for a television monitor or VCR to process the signal, it must
first be converted from its digital format back into an analogue waveform.
This is known as digital-to-analogue (D/A) conversion. It is at this point where
the problems associated with digitizing video signals become appar ent.
Consider the portion of video signal shown in Figur e 5.16. If connected
110
Closed Circuit Television
into a monitor, the analogue signal would pr oduce a steadily lightening
grey image along part of a TV line. However, after A/D and D/A processing, because the signal was sampled at the clock rate, the linear video
waveform is now a series of steps. When viewed on a monitor, this would
appear as a series of tiny rectangular blocks of increasing brightness.
The steps, or brightness increments, can be made smaller by increasing
the clock rate, but this means that we produce more bytes per TV line. For
argument’s sake, let’s suppose that we decide to produce the 780 pixels we
looked at in Figur e 5.4. This means that for each active TV line we will
have 780 bytes, which amounts to 6240 data bits. Thus, in one TV frame
we will produce 6240 585 active lines 3.7 Mbits. In one second we will
produce 3.7 Mbits 25 92.5 Mbits. Dividing by 8 to convert this figur e
to bytes, we have 1 1.6 Mbytes per second! So we can now see that if we
have an 80 Gbyte hard drive, we would be able to store 80 G 11.6 M 115
minutes of black and white video information.
A colour signal could
reduce this time by up to 50%.
The figures above are just an example, but they clearly illustrate how
much data is pr oduced once a video signal is digitized. The ‘r ecording
time’ of our 80-Gbyte har d disk can be incr eased in a number of ways.
For example, we do not actually require the signal to be broken down into
256 levels. We could use a six-bit binary wor d to r epresent each video
signal level. This would still give us 64 levels of grey scale, which produces
a reasonably acceptable image, with an incr ease in the storage time to
155 minutes.
Another method of increasing the recording time is to reduce the actual
area of the TV picture frame that is recorded. In other words, do not record
the full TV frame area. At first glance this may appear somewhat odd, but
when we consider that the field of view of many fixed CCTV cameras
includes irrelevant information such as brick walls or sky , why waste
valuable disk space recording such information?
Video compression
Although this process involves some of the most complex mathematics, it
is possible for it to be understood in simple terms. Ideally, compression is
all about removing any data from a digitized video signal which we know
can somehow be restored following ‘replay’ from the storage device. I say
‘ideally’ because in practice this is often not the case. In or der to achieve
longer storage times on digital recording equipment, high levels of compression are frequently applied – at the expense of replay picture quality because
it becomes impossible for the decoder to recover all of the discarded data.
There are a number of ways of recovering data (and therefore not having to transmit/store it in the first place). One of these is to employ different forms of redundancy. Spatial redundancy removes repeated information
within a frame, temporal redundancy removes repeated information over a
Fundamentals of television
111
number of frames, chromatic redundancy removes colour information that
is repeated, and perceptual redundancy removes information that may be
considered not relevant to the image (i.e., the viewer won’t miss it).
To illustrate temporal r edundancy, consider the television pictur e in
Figure 5.17a. This has one thing in common with almost all other TV pictures and that is, between one TV frame and the next (1/25th of a second),
most of the picture information does not alter. Thus, having digitized all of
the picture once and stored it, what is the point in storing the same data
again and again? Obviously we ar e referring to the backgr ound areas
which for most of the time ar e stationary; that is, until either the camera
pans, or the picture cuts to another shot.
Figure 5.17a A typical television image contains a lot of redundant information.
In this example of two consecutive TV frames there are subtle changes in the
face, and the body has moved very slightly
Figure 5.17b clearly illustrates the ef fectiveness of temporal r edundancy by showing only the difference information between one frame and
the following frame. From this illustration we begin to appreciate just how
much data can be excluded in transmission/storage of subsequent frames
without any loss of information in the final reconstituted picture.
So, one way of compr essing a digital video signal is to r emove duplicated data, thus making it redundant. To rebuild the picture, the compression decoder simply uses the same data over and over again to pr oduce
the walls and floor in our example. Another form of compression is prediction. Looking again at the example in Figur e 5.17, when the camera pans,
all of the repeated data will move in the same direction. Upon seeing this,
the compression processing chip, instead of passing all of the wall and
floor area for storage, produces a relatively small amount of mathematical
data which, during the recovery process, will be used to predict where the
wall and floor will have moved to on a frame-by-frame basis. In other
words, the same data is used over and over just like before, only now it is
being moved around the screen.
112
Closed Circuit Television
Figure 5.17b Illustrating temporal redundancy. Here only the changes between
two consecutive frames are shown. (Illustrations courtesy of Tim Morris)
As outlined at the start of this section, we cannot expect to heavily compress a signal and not lose something in terms of quality. The compression
chip sets are so designed that equipment manufactur ers using them can
set the level of compression, and the rule is very simple. A lot of compression results in a lot of redundant data, which means long recording times,
but at the expense of pictur e quality. Minimal compression means that a
lot of data need to be stor ed, reducing the recording time, but producing
high-quality (resolution) pictures.
Compression is measured as a ratio of the amount of data entering a compression system to the amount that comes out. Thus, a compression ratio of
1:1 indicates no compression, 5:1 would indicate that there is five times less
data at the output, 10:1 indicates a reduction of ten times, and so on.
The process of decompressing the data involves taking the data that has
been stored and applying it to complex mathematical algorithms. The success with which the original signal is r estored depends upon the amount
of original data that was r etained (that is, the amount of compr ession
applied) and the ef fectiveness of the algorithms (that is, how well they
have been derived). The term ‘lossless compr ession’ is becoming used
more and more, but only where little compression has been applied in the
first place. At the time of writing, once even moderate amounts of c ompression have been applied to an image, some information loss and/or
artefacts will be evident, even if only on certain types of pictur e information. This is because a point is r eached where the decoder has to make a
Fundamentals of television
113
‘best guess’ at what information should be put back into a certain pictur e
location and each time it gets this wr ong, an unwanted pixel is produced
on the picture (i.e., an artefact).
It is difficult to relate a specific compression ratio to any given number of
artefacts because there are too many variables. For any given compression
ratio, the point at which compression artefacts become noticeable depends
on the picture information, the type of compr ession used, the algorithms
employed by the manufacturer and the processing circuitry. Nevertheless,
as an aid to installers and end users, manufacturers often compare a given
compression ratio with known analogue formats such as VHS and S-VHS.
In practice it is difficult to make direct comparisons because the artefacts
produced by digital compr ession errors often manifest themselves quite
differently to those produced by analogue processes, but making such comparisons does serve the purpose of giving some idea as to how the r eproduced image will appear.
MPEG-2 compression
Digital CCTV equipment employs one of two forms of compr
ession:
MPEG or Wavelet.
MPEG (Moving Pictures Experts Group – a body set up by the ISO in 1988
to devise standards for audio and video compression) video signal compression is a very effective, robust and reliable compression format. MPEG-2 is
a natural evolution from MPEG (or MPEG-1) and, although it is the compression format used in a lot of CCTV equipment, it was primarily developed
with broadcast television and home entertainment in mind.
A TV signal can be said to be four-dimensional, having attributes of sample, horizontal axis, vertical axis and time axis.All of these can be explored
for redundant information by a compression encoder, and MPEG-2 employs
three types of compression to explore these four attributes.
The first type of compr ession is temporal r edundancy, also known as
inter-frame (P frame) compression because it is applied acr oss the frames.
This type of compr ession takes account of the fact that much of the video
information is the same fr om one frame to the next and ther efore it is only
necessary to transmit the dif ferences between the frames. Consider the
image in Figure 5.18. In this case, assuming that the camera angle is fixed, the
only significant changes between frames will be the movements of the vehicle in the centre of the picture. Thus we see that throughout the entire scene
when the vehicle is moving r ound the corner, which for the purpose of this
example we shall assume lasts for three seconds, large amounts of data in the
75 frames will be identical. If only data r elating to changes in pictur e information are transmitted, it may be possible to reduce the data content by 90%.
The second type of compr ession is spatial r edundancy, also known as
intra-frame (I frame) compression because it is only applied within a single
frame. Looking at the image shown in Figure 5.19, assuming a horizontal
114
Closed Circuit Television
Figure 5.18 The only significant movement in this clip is that of the vehicle, and
therefore the majority of the video information only needs to be transmitted for
the first frame
Line 10
Figure 5.19
resolution of 720 TVL, during each active line period many of the 720 luma
samples, or pixels, will be identical. Therefore, it can be reasoned that it is
only necessary to transmit the first pixel in a sequence, plus information to
inform the decoder how many times the pixel must be r
epeated. For
example, if we consider line 10 of the image in Figure 5.18, we can see that
instead of sending 720 pixels of data at 8 bits per pixel, we need only send
the 8-bit value of pixel number 1, along with a small amount of data relating to the position (line 10) and the number of repetitions (719).
Fundamentals of television
115
Intra-frame compression also makes use of the fact that, in general, the
amplitude of video components r educes as their fr equency increases.
Therefore, if we assign fewer bits to the high-fr equency end of the spectrum, a further saving can be made. However , this will lead to increased
noise in areas of picture that contain a lot of fine detail but, as long as the
process is not pressed too far, the eye will not easily discern this.
The third type of compr ession is statistical redundancy, also known as
prediction or motion compensation . Looking again at the image in Figur e
5.18, once the vehicle is fully in shot, it is a relatively simple operation for
the encoder to produce a plot of the pixel movements across the 75 frames
of the video clip. Ther efore it is not necessary to transmit all of the data
relating to the changes between frames, but simply the pr ediction data
required by the decoder to move the pixels between each frame.
Further data reduction may be introduced by replacing the regular and
predictable line and field sync signals with short codes that will indicate
to the decoder the start of these periods. Similarly , short codes may be
used to indicate black levels, which occur fr equently in most video signals. And finally, replacing long repetitions of ‘0’ or ‘1’ with shorter words
that simply indicate the number of r epetitions can reduce the amount of
data considerably.
Figure 5.20 shows how MPEG arranges the TV frames for storage and
subsequent transmission to the decoder. This illustration shows how one
I frame is transmitted on every 12th frame to serve as a ‘detailed’ eference
r
block, with P frames (predicted frames) interleaved between I and B frames.
I
B
B
P
R
B
P
R
R
P
B
B
I
Figure 5.20 The MPEG frame structure comprising I, P and B frames
The B frames (bidirectional frames) are made up purely from interpolated
information from adjacent P and I frames. One weakness with this system
is that an error in an I frame will pr opagate through the following frame
sequence until the next I frame. Such an error would persist for about 0.5s.
Errors occurring later in the sequence will endure for a lesser time period.
The twelve-frame cycle illustrated in Figure 5.20 relates to a typical broadcast MPEG-2 transmission although, in practice, the encoder designers
may employ more or fewer I frames, depending of the application and/or
amount of compression desired.
At this point we may be forgiven for believing that we have reduced the
data by as much as possible; however, it is here that the mathematicians take
over. The remaining 8-bit samples are arrayed into 8 8 matrices where
a process called discrete cosine transformation (DCT) is performed. This is
followed by a quantization process, resulting in a large reduction in data per
frame compared with the original 8-bit samples. DCT is a key part of MPEG
116
Closed Circuit Television
because this part of the compression process is truly lossless as every bit can
be recovered at the decoder by reversing the mathematical process. On the
other hand, quantization is distinctly lossy, and these losses are irreversible.
Within the encoding and decoding process there are a number of errors
that may occur . Data err ors can occur even though err or correction is
employed in MPEG because, if an error is too great, entire ‘slices’ of picture
will be r ejected because of the way in which data is arrayed within a
frame. Decompression errors, although not an err or as such, may occur
because of excessive removal in the redundancy process (i.e., a high compression ratio). Where large amounts of information have been identified
as redundant there is no way that the original data (and thus image) can be
restored. This results in a ‘blocky’ ef fect (pixilation) over ar eas of the picture. Finally, data may be insuf ficient because of ‘overloading’ at the
encoder. Overloading describes a situation where the video clip contains a
large amount of high-definition, fast-changing information. In such cases,
each frame bears little r esemblance to the adjacent frames in the footage
and therefore there is little, if any, redundancy. Consequently the encoder
is unable to output the large amount of data necessary to communicate the
images, resulting in large areas of pixilation and/or numerous freeze frames.
In a CCTV system, this situation is frequently encountered whilst a camera
is panning rapidly.
MPEG-4 compression
Available in October 1998, this compression format may be seen as MPEG-2
with a lot of additional functionality and features. The basic compression/
decompression remains the same as for MPEG-2; that is, it still employs spatial and statistical redundancy, 8 8 pixel blocks and DCT. Where MPEG-4
differs is in the functionality options available to system and equipment
designers.
It was stated earlier that MPEG-2 was developed primarily with the
broadcast and home entertainment industries in mind, i.e. digital television,
surround sound, DVD, etc. However , it must r emembered that these ar e
all two-dimensional video formats wher e all that is r equired is a highresolution picture with good-quality sound. By the early 1990s it was becoming clear that the thr ee emerging technologies of commer cial film/TV/
entertainment, computing and communications (in particular, the Internet
and network communications) were rapidly converging, and that a common standard for compression/decompression of this multimedia video
and audio content would be required.
To meet this requirement, MPEG set up the MPEG-4 committee, which
was given the task of pr oducing a suitable standar d. They first met in
September 1993. As this committee analysed the pr oblems, they realized
that a single set of compression rules set in stone (like MPEG-2) would never
be able to satisfy the complex and varied r equirements of the converging
Fundamentals of television
117
technologies. What they eventually devised was a whole range of digital
video/audio manipulation tools that would all be included in the MPEG4 standard, and from which system and equipment designers could draw.
Consequently, different items of equipment which incorporate MPEG-4
and perform the same function may in practice utilize a dif ferent tool set
to achieve their objective. From the point of view of CCTV equipment, many
of the available tools and functions ar e not really relevant as they belong
in the world of multimedia pr oduction and delivery. Nevertheless, some
CCTV equipment does utilize MPEG-4 and ther efore it is helpful for the
CCTV engineer to have an overall idea of MPEG-4.
So what does MPEG-4 offer in addition to MPEG-2 capabilities? Well, to
begin with, MPEG-4 has the ability to texture map 3D as well as 2D objects
using a Wavelet algorithm (Wavelet will be discussed later in this chapter)
although, at the time of writing, no one has actually managed to develop
a method of identifying these objects in real time. Another significant feature is its robustness, even when the bit error rate is high, making MPEG4 an ideal choice for digital video signal transmission over Ethernet and
wireless LAN.
However perhaps the most significant feature is the ability of MPEG-4
to separately identify objects within a scene, which enables these objects to
be individually manipulated and also interacted with by the user . Figure
5.21 illustrates the primary video analysis methods used by MPEG-4 to
separate video information in order to sample, compress and manipulate
the information in a video sequence with maximum ef ficiency. Because
objects are identified as separate entities (2D backgr
ound, arbitrary
shapes, 3D, etc.), the encoder is able deal with them as such from the point
of view of both compr ession and manipulation. In Figur e 5.21, the main
point of inter est is the goalkeeper , although the two girls in the background also represent areas of lesser significance. In MPEG-4 these thr ee
objects can be identified as arbitrary shapes and processed separately from
the rest of the video image. Although MPEG-4 utilizes 8 8 pixel blocks
and applies DCT to each block as in MPEG-2, it also has a featur e known
as shape-adaptive DCT which is applied to arbitrary shapes such as the
three girls in our illustration, with the effect of improving shape accuracy
and coding efficiency.
It is also possible to apply dif fering levels of compr ession to different
arbitrary shapes. For example, objects of little inter est could be subjected
to higher levels of compression than those of greater interest, thus improving the S/N ratio and image quality in the areas of greater interest. Referring
again to our image in Figure 5.21, the girl in the foreground is clearly more
significant than the two girls in the backgr ound and therefore these two
objects could be subjected to a higher level of compr ession than the data
relating to the goalkeeper.
Arbitrary shapes do not have to come fr om the original video sour ce.
They could in practice be computer -generated objects (animations, etc.),
or video objects that have been extracted from another video source. Such
118
Closed Circuit Television
Arbitrary shapes
of lesser interest
Arbitrary shape deemed to
be of high interest; potential 3D object
Figure 5.21 A TV image may be broken down into component parts by the
MPEG-4 encoder in order that each part may be processed separately. This
technique greatly improves the efficiency of the compression process as well
as offering features such as user interaction and the ability to add images
from other sources. (Photos courtesy of David Close).
objects may be inserted into each frame in a video sequence, even though
they are a sequence in their own right.
Figure 5.22 illustrates how MPEG-4 breaks an image down into separate
component parts. The main background forms the largest (2D) object and is
known as a sprite. Assuming that the camera is stationary, the sprite contains
all of the data that need only be transmitted occasionally (I frames) and for
all other frames can be replaced by a single 8-bit word which informs the
decoder that it is to use the same sprite again. When we consider that one
8-bit word has replaced thousands of pixels, we can immediately appreciate how much compr ession has been applied. The sprite may be much
wider than the display area; for example, the camera may be fitted with a
fisheye lens that produces a 360° image, but we only wish to view a part of
that image (with all distortion removed) on a monitor.
For most CCTV applications the goal netting in the image would probably be included in the sprite. However , for multimedia applications it
could be separately identified and encoded as a 3D object, thus bringing it
out to the forefront of the image.
MPEG-4 is capable of defining the movingparts in the image as separate,
arbitrary shapes which can be both compressed and transmitted separately.
As discussed earlier, this means that the encoder can apply different levels
of compression to each object, depending on the perceived importance of the
object, and can also apply different animation tools depending on the type
of object (video, synthetic animation, etc.). This makes for maximum compression efficiency, optimum S/N ratio for objects of gr eater interest,
Fundamentals of television
119
Original image
2D background
(sprite)
Arbitrary shape deemed
to be of high interest;
potential 3D object
Arbitrary shapes
of lesser interest
Mouse
(user interaction)
Theoretical view point
Figure 5.22 Illustration showing how MPEG-4 separates images into components.
These may then be transmitted separately and re-constructed by the decoder.
Where the sprite is much wider than the viewing area, the encoder must be able
to determine the desired viewpoint in order to crop the image accordingly.
(Photos courtesy of David Close)
improved ability to round the edges of objects (by use of shape-adaptive
DCT), and simplified motion pr ediction. The decoder r eceives the separate images (plus audio, if present) and, making use of the sophisticated
and robust synchronization tools, re-constructs the original image.
120
Closed Circuit Television
Finally, MPEG-4 incorporates tools that enable user interaction with
objects, which is an essential feature in many multimedia applications.
From this discussion of MPEG-4 it can be seen that it is a truly remarkable toolkit for multimedia pr oduction and data transmission. However ,
at the time of writing, MPEG-2 r emains the more popular of the two formats for CCTV multiplexer and DVR applications where the requirement
remains, primarily, a two-dimensional video image, usually without any
sound. Nevertheless, IP cameras find MPEG-4 a more suitable format, as
do the emer ging range of 360° cameras that incorporate high levels of
digital processing in order to produce a number of distortion-fr ee 4:3 TV
images from the one 360° sprite.
Wavelet compression
Signal analysis using wavelets is based on Fourier analysis, which was
formulated by Joseph Fourier in the 1800s, although it was not until the
1980s that wavelet analysis was used as a method of signal analysis. This
type of analysis has many applications including astronomy, music, earthquake prediction, radar, neurophysiology and – of course – video signal
compression.
Wavelet analysis does not break the TV frames down into blocks as in
MPEG; rather, it analyses each frame individually and as a whole. In other
words, there are no ‘P’ or ‘B’ frames. Figur e 5.23 illustrates how wavelet
analysis separates the signal components in each frame into fr equency
bands (typically 42 although, for simplicity , this illustration only shows
10 discrete bands). Bearing in mind that each fr equency band represents
video components of a particular resolution, from this illustration it is possible to identify the picture content contained within each band.
Having identified the components within a TV frame, it is now possible
to individually analyse each band and decide, from a compression point of
view, which components must be retained to maintain image integrity and
which components may definitely be discar ded without detriment to the
final reconstructed image. What happens to the remaining components will
be determined by the level of compr ession set at the encoder . Generally
speaking, the signal components which ar e deemed to produce information that is not visible to the human eye ar e discarded (typically picture
content of fine detail). Spatial compression is then applied to the remaining
bands and, finally, applied algorithms provide further compression of the
remaining data. This compression is non-linear with the greater compression being applied to the higher components (Figure 5.24).
Each frame is processed individually in real time, and is scanned three
times at the encoder to determine the optimum compression ratio for that
frame. Compared with MPEG, wavelet encoding is very much simpler and
therefore less expensive, a factor that can make W avelet more attractive
for CCTV applications. In br oadcast television there is only one (MPEG)
Fundamentals of television
121
Figure 5.23 How Wavelet sees an image. The original image is broken down into
discrete layers in terms of picture resolution. Different levels of compression can
then be applied to each individual layer. (Photo courtesy of Tim Morris)
122
Closed Circuit Television
Original TV
Frame
LF
MF
HF
Wavelet Component
Analysis
LF
MF
HF
LF
MF
HF
Frequency
Rejection
Compressed TV
Frame
Spatial
Compression
Figure 5.24 Simplified illustration of the wavelet compression process
encoder – at the transmitter – whereas in CCTV we have to build anencoder
into every multiplexer, digital recorder, etc. This makes the simpler wavelet
encoder more attractive fr om a manufacturing point of view , although
large-scale integration in silicon chips is tending to close this gap.
Because each frame is individually processed, Wavelet is much simpler
to edit. In CCTV, fast-switching multiplexers can r esult in MPEG having
to create a lot of I frames, thus increasing the data capacity. This is not the
case with wavelet compression.
Wavelet technology allows much higher compression ratios than MPEG
for the same image quality (or resolution). A wavelet compression ratio of
20:1 tends to be comparable to that of an MPEG compr ession ratio of 10:1,
although the losses in both systems ar e quite dif ferent, making a dir ect
comparison difficult. This being the case, the larger data file size of wavelet
compression is countered by the fact that greater compression ratios can be
applied.
Because it does not break the picture down into the 8 8 blocks, wavelet
compression does not normally produce the ‘blocky’ effect associated with
MPEG when larger amounts of compression are applied. Rather, a heavily
compressed wavelet image takes on a blurr ed appearance, perhaps comparable to looking at a pictur e produced by a camera with a slightly defocused lens. Having said this, it is possible to experience a blocky ef fect
with wavelet compression if the encoder design utilizes particular (usually simpler) basis functions. In Wavelet algorithms ther e are numerous
basis functions, and the encoder designer will select certain ones, depending
on the image quality that they wish to achieve and the computational complexity they are willing to afford.
Fundamentals of television
123
In conclusion, wavelet compression sees the TV frame as a number of
layers of differing resolution. Each layer is individually analysed to determine whether it needs to be r etained, and those that are retained are further analysed to determine how much (ifany) compression may be applied.
Higher compression levels ar e applied to the higher -resolution layers –
however, this amount is variable. As with any compression system that is
based on mathematical algorithms, there are many ways that the compression may be applied, with varying results. The problem with wavelet compression throughout the 1990s was the lack of any industry standar d and
therefore, in the year 2000, the Joint Photographic Experts Gr oup (JPEG)
released the JPEG-2000 standard, which is the official standard for wavelet
video signal compression.
Common interchange format (CIF)
Sometimes referred to as the common intermediate format, these standa rds
specify picture formats in pixel sizes. CIF itself describes an image size of 352
horizontal 288 vertical pixels (352 240 NTSC) befor e compression,
which amounts to one quarter of the number of pixels in a br oadcast TV
picture frame. 2CIF describes an image that contains twice the number of
pixels as CIF, which is 704 288 (704 240 NTSC) and 4CIF describes an
image containing four times CIF: 704 576 pixels (704 480 NTSC). Note
that for 2CIF, only the horizontal r esolution is increased as this in ef fect
doubles the number of pixels, and 4CIF doubles both the horizontal and
vertical pixels, making it equal in resolution to a full broadcast frame.
Other CIF specifications include QCIF (quarter CIF) , which defines a
resolution of 176 144 pixels, SQCIF (sub-quarter CIF) , which defines a
resolution of 128 96 pixels, and 16CIF, which is 1408 1152.
ITU-T recommendations
Over the years the ITU (International Telecom Union) has produced a number of standards which deal largely with two-way audio/video communication. The ITU-T H.320 standard is a recommendation that specifies the
requirements for low-bandwidth audio/visual telephone and video conferencing equipment and services. As such, H.320 comprises a number of
other (H.26) substandards.
Published in 1990, the CCITT H.261 standard is the substandar d that
defines video compression for H.320. The specification was defined with
video transmission over ISDN (Integrated Services Digital Network) in
mind and, as such, specifies data rates in the order of 64 kb/s and 128 kb/s.
Images are non-interlaced and the DCT coded output comprises luminance
and colour difference signals which may be either CIF or QCIF resolution.
124
Closed Circuit Television
The compression methods applied contain 8 8 pixel blocks, intraframe
and prediction similar to those described earlier in this chapter.
Owing to its low resolution (which is important in order to maintain a
low bit rate when transmitting over a low bandwidth channel such as
ISDN), H.261 does not perform well in CCTV applications.
A more effective specification is the H.263, which is a natural evolution
of H.261. This standar d specifies resolutions that are higher than H.261
but at lower bit rates, which is made possible by more advanced compression techniques being applied. The output may be SQCIF, QCIF, CIF, 4CIF,
or 16CIF. The standard was introduced during the mid-1990s to accommodate video data transmission over the very narr ow bandwidth communication channels pr ovided via PSTN and GSM modems, which ar e
typically 9600 bps.
A more recent standard (May 2003) is the H.264, which is also specified
under a different but familiar standar d, MPEG-4 Part 10. The reason for
having two identical standards is because they were developed jointly by
the MPEG and the ITU-T gr oups, who set out to cr eate a standard that
would exceed MPEG-2, MPEG-4 Part 2 and ITU-T H.263 in terms of image
quality and bit rate.
H.264 offers bit rates at half that of H.263/MPEG-2 without a comparable loss of image quality . This is made possible by the application of a
range of compression tools, some of which are new, and many of which are
improvements on existing tools employed in earlier compr ession specifications. For example, motion detection spanning many frames offers large
reductions in data when video sequences contain scene cuts that jump rapidly back and forth (which is basically what happens when a CCTV DVR is
recording multiple cameras simultaneously). The pixel block sizes ar e also
no longer fixed at 8 8 but are dynamically variable between 16 16 and
4 4, which provides very precise definition of moving arbitrary objects and
reduces aliasing, resulting in sharper images.
6 The CCTV camera
The camera can be consider ed to have two parts: the pick-up device(s),
and the signal-processing circuits.
For many years cameras relied on thermionic (valve technology) tubes
to derive a signal voltage from the incoming light information. However,
cameras employed in modern CCTV systems use a solid-state pick-up device
known as a charge coupled device (CCD).
Similarly, advances in technology have meant that the traditional analogue
signal-processing techniques which have served us well for many years have
given way to digital signal processing, meaning that cameras are able to produce remarkably clear pictures under very hostile lighting conditions.
Charge coupled device
The CCD is a silicon device which can stor e an electrical char ge. A chip
containing a number of CCDs in an array can be used to stor e samples of
analogue video or audio signals where they can be manipulated. And so,
although the CCD chip is not in itself a digital device, when controlled by
a microprocessor it can be used to move analogue samples ar ound in the
fashion of a shift register.
The CCD design can be modified such that electrons are released when
photons (light) fall onto the device. Thus the CCD behaves somewhat like
a photodiode. If the light output fr om a lens is focused onto an array of
these photodiodes, each diode will derive an output voltage proportional to
the amount of light falling upon it. Thus the chip is converting light energy
into proportional electrical charges.
A typical imaging chip used in a CCTV camera contains many thousands
of CCDs arrayed in a rectangular pattern. As we shall see in a moment, the
voltages from the cells are integrated to create individual pixels (‘pixels’ is
derived from the term picture elements). For a CCD image device, the picture resolution is determined by the number of cells in the chip, and the
density of the cells. In theory a 1⁄2 CCD chip will have better r esolution
than a 1⁄3, and this is true if we compare like for like. However, a modern
1
⁄3 chip can have a gr eater cell density than an older 1⁄2 device, which
means that a new 1⁄3 camera could possibly offer a higher resolution than
the old 1⁄2 camera it is replacing. As a general rule, the greater the cell density, the higher the cost of the chip.
CCDs can have a tendency to overload under bright light conditions,
causing light areas of the picture to diffuse into a white mass. This effect is
126
Closed Circuit Television
termed ‘burn-out’; however, this is not inferring that the CCD itself is
damaged, only that the picture quality is degraded. The burnout effect can
be avoided by using suitable filters and a good-quality iris control.
A typical CCD chip is shown in Figure 6.1.
Figure 6.1 A 1/3 CCD imaging chip employed in CCTV cameras
CCD chip operation
A frame transfer chip is illustrated in Figure 6.2 and, although this chip is no
longer employed in TV cameras, it does serve as a good starting point when
describing the shift register action of a CCD chip.
Although called a frame transfer chip, this early device actually operated
at field rate. During one field period of 20 ms the imaging CCDs gather a
charge from the incident light. Then, during the field flyback period, the
charges are shifted downwards through the image CCD cells into the CCD
storage area. During the period of the following field the image devices
re-charge, whilst the information in the storage area is moved one TV line
at a time into the horizontal storage area, from where it is clocked out over
a 52 s (one active TV line) period. At the end of the field period, the cycle
repeats.
The problem with this method of char ge transfer is that the char ges
have to move down through the image area to get to the store area, meaning
that the image CCDs must r emain active not only thr oughout the field
period, but also during the charge transfer period. The charges are therefore ‘topped up’ as they travel down through the image area by the incident
light which is still falling onto the chip, resulting in a vertical smear in bright
areas of picture content.
The solution to the vertical smearing problem was to re-design the CCD
image chip, and employ a charge transfer technique known as interline transfer. The chip architecture for this is illustrated in Figure 6.3.
With this architecture, each CCD charge can be moved directly into its
allocated temporary store area without having to travel through the other
The CCTV camera
CCD photodiodes
Exposed area
(window)
Masked area
CCD storage area
(capacity 1 TV field)
Output
Horizontal CCD storage
devices
Figure 6.2 Frame transfer CCD chip principle
CCD photodiodes
Vertical CCD storage devices
Horizontal CCD storage devices
Output
Figure 6.3 Vertical and horizontal CCDs in an interline transfer image chip.
The time that it takes to clock the information out of the H-CCD is 52 s –
one active line period
127
128
Closed Circuit Television
image cells, thus eliminating the problem of vertical smearing. The image
cells are exposed for a brief moment during an active field period. A previously acquired TV field will have been moved from the image CCDs into
the vertical CCD storage devices (V-CCDs) during the field-blanking period.
Thus, as the image cells are re-charging, this information is clocked at line
rate, and pr ocessed to pr oduce the video information for the field currently being viewed. By the end of the field period, all of the information
in the V-CCD will have been clocked out and thus, during the field flyback
period when no video signal is r equired (other than a black level which
can be generated by the camera), the image CCD charges for the next field are
simultaneously downloaded into the V -CCD area. At the start of the following field period the image CCDs ar e once again exposed, whilst the
stored field is once again clocked out.
The horizontal CCD (H-CCD) has the capacity to stor e one TV line of
information. During each horizontal flyback period one charge from each
V-CCD is moved into the H-CCD.At the same time all of the other charges
in the V-CCDs are moved downwards. During the following active line
period, the charges in the H-CCD are clocked out in serial form. From here
they pass into the camera signal-processing circuits to be output as a video
signal. The action can be likened to that of a cigarette vending machine where
the bottom packet is removed and all those above fall down by one position.
The main drawback with the interline transfer chip constr uction lay in
the fact that the image surface is no longer occupied solely by image CCDs.
Because the vertical CCDs ar e adjacent to the image devices, a lot of the
light falling onto the chip is unused. This reduces the sensitivity of the CCD
chip, making it less able to cope under low light conditions. Other pr oblems encountered with this technology are the migration of charges within
the molecular structure, and the problem of random electron release in the
store areas caused by thermal action and photons penetrating the stor e
area. Such effects result in a r eduction in the signal-to-noise (S/N) ratio
and the possibility of vertical smearing under high lighting conditions.
The solution to these problems is found in the type of chip that has been
in use for a number of years, the frame interline transfer (FIT) chip. This still
operates on the interline transfer principle, but instead of holding the
charges in the vertical CCDs for the complete field period, they are moved
into a lower stor e area just as in the frame transfer chip. The principle is
illustrated in Figure 6.4. Compared with the interline transfer chip, the FIT
chip offers a very high S/N ratio, low smear, and much improved low light
performance.
A development that has done much to improve CCD chip sensitivity is
the micro lens, where a microscopic lens is fabricated over each individual
image CCD. The principle is illustrated in Figure 6.5. Without the micro lens,
light falling between cells is lost, resulting in a reduction in chip sensitivity. The micro lenses gather this light and focus it onto the cells, effectively
increasing the chip sensitivity. Sony took this principle a step further with
the introduction of the exwave chip, which employs a more effective micro
The CCTV camera
129
Vertical CCD storage devices
CCD photodiodes
Exposed area
(window)
CCD storage area
(capacity 1 TV field)
Output
Masked area
Horizontal CCD storage devices
Figure 6.4 The frame interline transfer (FIT) chip combines the advantages of
the frame transfer chip (having a masked storage area safe from corruption from
incident light) and the interline transfer chip (adjacent storage areas removing the
effect of vertical smear)
lens, as illustrated in Figure 6.5c. The majority of CCTV cameras used today
employ Exwave technology CCD chips.
An alternative to the CCD chip technologies discussed so far is the
digital pixel system (DPS) image sensor produced by Pixim. The main difference between this and a conventional CCD image chip is the move away
from the analogue shift register principle. DPS image sensor chips have an
analogue-to-digital (A/D) converter dir ectly attached to each individual
photodiode pick-up, or pixel. This means that the image chip is able to
perform complex pr ocessing functions even befor e the information is
clocked out, resulting in very high sensitivity, high resilience to burn-out,
wide dynamic range, high S/N ratio and excellent colour quality.
The principle of operation is shown in Figur e 6.6. The A/D converter
on each photodiode is connected to a data bus where its binary output (relating to the incident light level) is passed to a RAM. Because the information
is being shifted in digital form, the problems of smear and noise associated
with incident light are eradicated.
130
Closed Circuit Television
Incident light
CCD cell
(a)
Micro lens
Incident light
(b)
Incident light
(c)
Figure 6.5 (a) In a conventional arrangement, light falling between cells is
lost. (b) The micro lens increases the light input to each cell, thus improving the
sensitivity of the chip. (c) Sony Exwave technology results in even better lightgathering ability
DPS chips offer a very wide dynamic range (that is, the range between
the darkest and the brightest visible information in a scene) and have a
high resilience to burnout caused by overexposure because of the unique way
that each pixel samples the incident light. In a CCD chip, each cell char
ges up
to a voltage potential that is proportional to the level of light that is falling
upon it. However, this means that under low lighting conditions the charge
level will be very low , making the signal voltage liable to fluctuations
caused by the action of free electron charges in the chip or, in other words,
background noise. At the opposite extreme, under high levels of lighting, the
cells will charge to their peak and the image will appear burnt out. DPS
chips do not sample the light level in the same manner.Because the charge
created by the incident light is immediately converted into a binary value,
the chip is able to monitor the rate of rise of the charge value at each photodiode and, from this information, determine not only the peak level but
The CCTV camera
Light sensor
131
ADC
Data bus
RAM
CVBS
Digital image processor
Y/C
Figure 6.6 DPS image sensor principle. An ADC is attached to each individual
cell/pixel; therefore the light input can be continuously sampled, enabling a
very accurate measure of the light level at each point on the pick-up area
also the extreme light levels (i.e., very dark/very bright). With this information, the chip can decide at which point in time each pixel will take its
sample, thus avoiding the pr oblems of having some parts of the pictur e
under/overexposed.
The principle is illustrated in Figur e 6.7. The photodiodes r elating to
the dark areas in this photo will take a considerable time to accumulate a
charge, and so the chip waits until the point in time just befor
e these cells are
about to overexpose (time t3 in Figure 6.7) before taking the sample. The
digital image processor is then able to calculate the true light level at these
cells by equating the rate of rise of char ge over the time taken to r each a
maximum value. The photodiodes relating to the light areas of the image
will reach a maximum charge value much more rapidly, so the processor will
use time t1 to calculate this light level. Thus it can be seen that the DPS chip
avoids having to take samples at low cell char ge levels (so avoiding poor
S/N ratio in dark picture areas) and does not suffer from burn-out because
a sample will always be taken just befor e each photodiode cell reaches its
saturation level. In other words, DPS chips dynamically sample the image.
The dynamic range is improved using this dynamic sampling technique
because information in the extreme light and dark areas of the image are not
132
Closed Circuit Television
Light image area:
Photodiode charge rate a
Medium image area:
Photodiode charge rate b
Photodiode
saturation
level
100%
Nominal
photodiode
sample level
Dark image area:
Photodiode charge rate c
a
b
t1
c
t2
t3
time
One 50 Hz TV field period: 20 ms
Figure 6.7 DPS image chips sample the light level at a point where each
photodiode has a high charge potential. The actual light level can then
be computed from the time that it took for the charge to reach this level.
This ensures a good S/N ratio and no image burn-out
hidden by overexposure or background noise. There are, however, some
viewing conditions where a high dynamic range can prove counter-productive. For example, where a scene is generally dark but has a bright area within
it, the lens iris will by nature of its operation open up to accommodate the
darker area. This will cause the bright light content to reflect inside the lens
The CCTV camera
133
and scatter before reaching the image chip, resulting in an apparent misting
of the view. This does not mean to say that this ef fect is inevitable wh en
using cameras that employ DPS chips. The camera manufactur er can take
steps in the design to restrict the dynamic range under such conditions.
Electronic iris
Interline transfer made possible the introduction of electronic iris (EI). This
can be compared to the mechanical shutter in a photographic stills camera;
however, it is performed electronically by the application of a voltage to the
cells. The electronic iris circuits in the camera adjust the charge time of the
CCD cells to suit the average incoming light level.
The advantage of EI is that, where a fixed iris lens is used, the camera is
able to compensate for changes in lighting levels. However, where the lens
has an automatic iris (AI), the electronic iris should be switched off to prevent an effect known as ‘hunting’. This is wher e, following a rapid and
large change in light level, both iris cir cuits react. However, the EI normally reacts first, and thus when the mechanical iris closes a moment later
the light input becomes too low . This causes the EI to ‘open’ once again,
followed quickly by the mechanical iris, so the light level is once again too
high, and the cycle repeats a number of times until the iris circuits stabilize.
This oscillation can be sustained at a very rapid rate for quite long time,
resulting in a rapid changing in brightness level along each television
scanning line, pr oducing a patterning ef fect over the whole pictur e that
could be mistaken for an RF interference effect.
IR filters
CCD image chips ar e generally sensitive to IR radiation, which is what
makes them so sensitive under low light conditions.
At first glance thi s
might appear to be good news, and in some instances it may be. However,
unrestricted penetration of IR radiation into a CCD image chip can cause
problems.
Because of the longer wavelength, IR radiation penetrates deeper into the
silicon substrate of the CCD chip than visible light, and this penetration
can lead to the undesirable release of electrons in the charge storage areas,
changing the values of the wanted char ges, causing smearing and loss of
definition. To prevent this phenomenon an IR cut filter is placed over the
light input window of the CCD chip. This reduces the sensitivity of the chip
to some degree; however, the improvement in definition and S/N ratio
makes it a worthwhile trade-off.
Monochrome cameras employing chips without an IR filter are available,
and find applications where the area is to be illuminated by IR spotlights,
134
Closed Circuit Television
or where low light level operation is required. All colour CCD chips must
have an IR filter in order to produce accurate colours.
Colour imaging
Up to this point we have only considered the operation of the monochrome
image chip, which only produces a luminance signal. Producing a colour
signal is somewhat more involved because the chip must be able to generate
three signals: red, green and blue. There are two ways in which this can be
achieved: by using three chips, or by using a single chip with a colour filter.
The three-chip method, illustrated in Figure 6.8, is by far the best. The
incoming light is split into its three component parts using an array of mirrors, including special dichroic mirrors which reflect some frequencies of
light whilst allowing other frequencies to pass through. Each CCD operates
in very much the same fashion as it would in a monochrome camera, producing picture information relating to the colour of light falling onto it.An
optical filter is placed in front of each pick-up to correct for deficiencies in
the dichroic mirrors.
Because each CCD is operating in the same manner as a monochr ome
image device, the r esolution and light level performance of a thr ee-chip
colour camera are comparable to its monochr ome counterpart. However,
the combined cost of the optics and the three CCD chips makes this type of
Optical filter
CCD
Optical mirror
Red output
signal
Optical lens
assy
R
G
B
Green output
signal
Dichroic mirrors
Blue output
signal
Optical mirror
Figure 6.8 In a three-chip colour camera, the light is broken down into its
component parts and focused onto three CCD chips
The CCTV camera
135
camera very expensive and has for many years assigned these cameras
largely to br oadcast and semi-pr ofessional video pr oduction use. As
advances in technology bring the cost of thr ee-chip cameras down we
have, in more recent years, begun to see some three-chip cameras employed
in CCTV applications; however, they are still much more expensive than
their single-chip counterparts.
Colour cameras for the CCTV industry are generally single-chip units.
The chip is identical to that used in a monochrome camera; however, a filter
is placed in front of the CCD window to break the light up into red, green
and blue. There are two types of filter in use: thestriped filter and the mosaic
filter. Each of these has its str engths and weaknesses, and comparing the
two, the striped filter offers better colour reproduction, whereas the mosaic
filter offers superior resolution.
The striped filter, illustrated in Figure 6.9, forms a mask of alternate r ed,
green and blue strips of filter material in fr ont of the CCDs. Each CCD pr oduces an output for just one colour. These output signals are processed in the
same manner as we saw for the monochr ome chip in Figur e 6.3; however,
the output from the horizontal shift register has to be further processed to
derive the luminance and colour signals.
The reason for the poor resolution with this type of pick-up is apparent
when we look at Figure 6.9. In the horizontal direction, three CCD cells are
required to produce one pixel, whereas in a monochrome pick-up the same
cells would produce three pixels. It might therefore appear as though the
Red
Green Blue
Red
CCD photodiodes
Filters
Y
R–Y
B–Y
Processing
CCD storage area
Horizontal shift register
Outputs to luma/chroma
processing circuits
Figure 6.9 Frame interline transfer (FIT) chip with a striped colour filter
136
Closed Circuit Television
horizontal resolution is reduced by two thirds; however, when other factors
are taken into consideration it can be shown that the er solution of this type
of chip is about 50% of that of a monochrome chip.
Another problem which occurs with the striped filter is patterning on
areas of fine picture detail. This is caused by the generation of unwanted
frequency components when the picture detail falling onto the chip is about
the same size as the filter stripes. This interaction between the filter and the
picture information, known as beating, is overcome by placing a crystal filter
between the main optical lens and the CCD chip.
The mosaic filter is considerably more complex, both in terms of the filter
construction and the signal pr ocessing required. A filter is placed above
each CCD cell in a mosaic pattern as shown in Figure 6.10. This sixteen-part
block pattern is repeated across the entire chip.
Using the three secondary colours yellow, magenta and cyan, plus the
primary colour green, the processing block following the horizontal shift
Old (first)
field
Even (second)
field
G
M
G
M
Y
C
Y
C
M
G
M
G
Y
C
Y
C
Line 1
Line 2
Line 1a
Line 2a
R
G
B
Processing
CCD storage area
Outputs to luma/chroma
processing circuits
Horizontal shift register
Y yellow
G green
M magenta
C cyan
Figure 6.10 Frame interline transfer chip with a mosaic filter
register is able to derive red, green and blue signals which can then be further processed to produce the Y and C components. The algorithms used to
derive the colour signals can vary, depending on the number of CCD cells
there are in the chip; however, the greater the number of cells, the better the
CCD performance.
The CCTV camera
137
Both the striped and mosaic filters r educe the light input to the CCD
cells. This is one of the prime reasons why colour cameras are much less sensitive than monochrome cameras. While it is common to find monochrome
cameras with a specified minimum light input level of 0.1 lux, in general
the minimum level for colour cameras is in the order of 1 lux.
Camera operation
The principle of the colour camera is illustrated in Figure 6.11. The shift registers which pr ocess the CCD char ges are contained within the CCD chip,
and thus the input to the signal processing block is the output from the CCD
horizontal register. The char ge transfer pr ocess is contr olled by the CCD
‘Video iris’
control output
d.c. iris
control
‘DD iris’
output
CCD
Colour
filter
Electronic
iris
S1
Y
Lens
assy
CCD signal
processing
AGC
Y
Luma signal
process
Y output
Composite
output
R-Y
B-Y
CCD driver
Colour signal
process
3
Sync
generator
2
RV1
Crystal
oscillator
C output
1
V-phase
Low voltage
50 Hz from PSU
Genlock
input
S2 Sync
selector
Figure 6.11 Block diagram of a single chip colour camera (PAL)
driver, which in turn is clocked by an accurate crystal oscillator . Because of
the relationship between the charge transfer and the line and field sync signals, the CCD driver takes a reference from the sync pulse generator.
The signal processing is highly complex, but as far as we are concerned
it is sufficient to note that there are three signals present at the output: the
luma and the two colour difference signals R-Y and B-Y.
The average amplitude of the Y signal is determined by the average
light input level, and thus the Y is monitored by the EI cir cuit. If the light
138
Closed Circuit Television
input level is consistently high or low, the EI control changes the exposure
time of the CCD accor dingly. This function enables a camera to be fitted
with a manual iris lens and yet maintain a degree of control under changing lighting conditions. However , it should be r emembered that, unlike
the manual iris, EI does not af fect the depth of field and so, should the
engineer set the manual iris to a low F-number, the EI will ‘close’ to maintain a correct light level; however, the depth of field will be poor. The situation will become even worse at night when the depth of field further
reduces despite the fact that the EI has ‘opened’.
Another undesirable
effect of electronic iris is the increase in smear when the exposure time of the
CCD is reduced.
The EI circuit can be switched on or off by S1, which may be accessible
either on the side of the camera or through a software set-up menu. The EI
should be switched off when the camera is used with an auto iris lens otherwise the two may tend to fight each other when sudden changes in light
level occur, resulting in an iris oscillation effect.
The luma signal is also fed to both the d.c. iris contr ol circuit, and the
output socket on the camera provided for ‘video’ iris control. The action of
these was discussed in Chapter 4.
The signal voltage level derived by the CCDs varies considerably with
changes in light level, and in low light conditions the signal voltage is
so small that a lar ge amount of amplification is necessary to pr oduce an
acceptable output. However, in or der to pr oduce a high-quality pictur e
under all lighting conditions, the gain of the amplifier must be made variable so that it can reduce as the light input level increases. The automatic gain
control (AGC) circuit is an amplifier which includes a signal level monitoring
circuit. As the average level of the signal alters, the gain of the amplifier
alters accordingly so that, for example, when the signal level is high the
gain is reduced, and vice-versa.
As discussed earlier, the CCDs generate noise as well as a signal voltage.
Under reasonable lighting levels this noise is hidden by the high signal-tonoise (S/N) ratio. However, under low light conditions where there is less
signal voltage generated, the S/N ratio reduces. At the same time the gain of
the AGC amplifier rises to a very high level, and both the video signal and
the noise is amplified. This noise appears as a backgr ound grain on the
picture which detracts from the resolution.
As we have seen, the luma signal is derived in the CCD signal processing
circuits; however, further processing is required before the luma is r eady
to be sent to the monitor. Much of this processing is of little interest to the
CCTV engineer; however, one process which is worth mention is gamma
correction as it has a direct bearing on the shape of the output waveshape.
CCD pick-up devices generally have a linear output r esponse, and so
equal increments of brightness give rise to equal changes in output voltage.
Unfortunately the CRT characteristic is not linear but follows a curve like
that shown in Figure 6.12. If we were to apply an equal increment staircase
waveform (like that illustrated in Figure 5.9, Chapter 5) to the cathode of a
The CCTV camera
139
CRT, the light output would not rise in equal increments. Instead, the steps
in the dark and light r egions would give rise to lar ge increments in CRT
beam current, whereas the intermediate (mid-grey) steps would produce
compressed levels with little contrast difference between bars.
Light
output
Gamma correction
amplifier response
Resultant
Gamma response of CRT
CRT beam current
Figure 6.12 CRT response and associated gamma correction curve
Because of this the luma is passed thr ough an amplifier whose characteristic is equal but opposite to the characteristic of the CRT. This process is
called gamma correction. The effect of gamma correction on the luma staircase waveshape is shown in Figure 6.13.
Black
White
Figure 6.13 Staircase signal with gamma correction (exaggerated)
The gamma correction circuit could have been built into the monitors, but
when we consider this fr om the point of view of br oadcast TV where there
may be just one or two cameras whose output is being viewed on literally millions of receivers, then it does not make economical sense to put a circuit into
every TV. Therefore, the correction circuits are found in all video cameras.
Switching the gamma on and off on a CCTV camera appears to simply
alter the contrast level; however, remember that with the gamma turned
off the CRT will produce greater contrast changes at the lower and higher
brightness levels, leaving the mid-grey levels somewhat bland. As a general
140
Closed Circuit Television
rule the gamma correction should be turned on, offering a linear contrast
range. However, you may experience cases where the CRT phosphor characteristic does not match that of the gamma correction in the camera and the
contrast range appears too stark at the high and low levels. In such cases a
more linear appearance may be obtained with the gamma correction turned
off. The point to remember is, don’t simply use the gamma correction setting
as an alternative gain control.
Referring once again to the block diagram in Figure 6.11, the minimum
bandwidth of the luma process circuits must be 0–5.5MHz for a resolution of
around 550 TVL, although with modern technology many CCTV cameras
are able to exceed this resolution figure.
Once the luma signal has been derived, line and field sync signals ar e
added. These pulses are generated in the sync generator stage, which derives
an accurate signal by dividing down the crystal oscillator output.
The need for synchronization between the camera and monitor was discussed in Chapter 5, but at that point we only looked at a single camera and
monitor. Problems arise as soon as we try to send two or more camera outputs to a single monitor through a switcher or multiplexer because, unless
steps are taken to lock the cameras together in terms of synchronization, at
any instant in time the scan position of the two cameras would be totally
different. Each time the switcher toggled, the monitor would have lock to
the new scan position, the picture rolling and pulling as it did so.
Switch S2 is a thr ee-position switch with a common output terminal
connected to the sync generator. Position 2 is not connected to any circuit,
and when the switch is in this position the sync generator cir cuit simply
‘free wheels’ and generates an independent sync signal. This position is used
where the switcher or multiplexer contains circuits that can lock the camera
signals together, or where there is just one camera, or where the customer
doesn’t mind a lot of picture roll!
Position 1 is connected to an external socket (generally BNC), which is
usually labelled ‘genlock’. This is used where one camera, or perhaps the
switcher or multiplexer, generates the line and field sync, which is then sent
to each camera via a separate co-axial cable. The poor economics of using
this in CCTV are obvious when we consider that each camera requires two
co-axial feeds. Genlocking is normally only employed in broadcast television
where multicore cables are used, and cable runs are relatively short.
Position 3 is only available on cameras that ar e either 230 Va.c. fed, or
24 Va.c. fed. A sample of the 50 Hz mains frequency is fed to the sync generator which locks to the zero transit point in each a.c. cycle. If each of the
cameras is connected to the same phase of the mains supply, then they will
be synchronized. This method of synchronization is called line lock (LL).
Problems can occur on larger installations where cameras in different locations are fed from different phases of the mains supply; however, this can be
overcome by adjusting the V-phase control RV1, which is able to shift the
sync output from the generator cir cuit through 120°, which is the dif ference between any two mains supply phases.
The CCTV camera
141
Multiple camera synchr onization will be looked at in mor e detail in
Chapter 9.
The colour signal pr ocess block takes the R-Y and B-Y signals and
applies weighting to produce the u and v components. It also produces the
4.43 MHz sinusoidal subcarrier which, in the case of Figure 6.11, is done by
dividing down the master crystal oscillator, thus ensuring that the phase
relationship between the chroma, sync and CCD output is always correct.
The u and v signals are amplitude-modulated onto the subcarrier .
Normally with AM only one signal can be put onto a carrier, but in a PAL
colour TV signal a method called quadratur e amplitude modulation
(QAM) is employed which enables two separate signals to be placed onto
one carrier in such a way that they can be separated in the colour decoder.
To complete the PAL chroma signal package, a sample containing ten
cycles of 4.43 MHz subcarrier is positioned to coincide with the back porch
period of the line sync pulse. This signal, known as thechroma or colour burst,
is used to synchr onize the colour decoders in the monitors or r ecording
equipment.
The majority of colour cameras of fer a choice of either separate Y/C
outputs (S-VHS) or composite video. The relative merits of these were discussed in Chapter 5. All output connectors are terminated at 75 to provide
correct impedance matching to the 75 co-axial cable.
White balance
Up to this point we have assumed that the CCD chip or chips in a colour
camera are able to produce exacting proportions of R, G and B for any given
light input. Unfortunately this is not the case. T olerances between chips,
and within each individual chip, affect the R, G and B levels to such an extent
that the colour quality of the r eproduced picture would be at best poor ,
and in most cases totally incorrect.
Following power-up, the camera needs to be shown what white light
looks like. It then uses this information to derive correction values for the
processing circuits which ensure that the correct proportions of R, G and B
are produced for a white light input. In theory, if the camera is producing
the correct proportions of R, G and B for white, then all other colours will
be correct.
To perform white balance set-up a white card is held in front of the lens
to provide a white light input. The white balance function is then selected,
and the camera quickly adjusts its cir cuits and memorizes the corr ection
factors. As an alternative to using white card, if there is a large white object
within the field of view, the camera can be trained onto this.
This method is not always practical for CCTV applications and ther efore
the majority of modern cameras have an automatic white balance (AWB) function which uses pre-set correction factors. There is normally a choice of settings such as ‘Indoor’, ‘Outdoor’, ‘Fluorescent’ or ‘Auto’. The manufacturer
142
Closed Circuit Television
takes the average lighting conditions for these situations and designs suitable
correction factors into the camera. Apart from operating the selector switch,
there is no setting up r equired with this type of white balance, but ther e
are obvious limitations with this technology.
The main drawback with automatic white balance is the ef fect that the
different light colour temperatures have on picture quality. Take, for example,
an external camera which is to rely on halogen lighting during the evening. If
it is balanced using daylight as the light source, the colour may well be incorrect during the evening. Even during the daytime, difficulties may be experienced if ar eas within the field of view have dif ferent colour temperatur e
lighting. This situation can arise when fully functional cameras are employed
and their panning action takes them acr oss different lighting conditions.
These problems associated with AWB are one reason why for many years
colour cameras were often considered unsuitable for external applications.
Advances in technology have largely overcome the problems associated
with AWB. An automatic tracking circuit is included in many colour cameras, which ensur es that once AWB set-up has been performed, corr ect
white balance is maintained for all colour temperatures of light.
Back light compensation
A problem that is commonly encounter ed with CCTV cameras is wher e
the user needs to r ecord (clearly) a person when they ar e in a doorway.
During the evening this is not too difficult; however, during the day when
the area outside the doorway is brightly lit, the iris will close down to prevent overexposure of the outside ar ea. Unfortunately this r esults in the
person becoming underexposed and appearing as a silhouette. A similar
effect was discussed in Chapter 4 (see Figure 4.12a); however, in this case the
problem was resolved using the peak level adjustment in the lens.
Unfortunately many CCTV camera lenses do not have this facility and
therefore most cameras incorporate a featur e known as back light compensation (BLC). When turned on, the camera for ces the lens iris to open up
(assuming that an AI lens is fitted) to a point wher e the ar ea outside
appears overexposed; however, the person is at the correct exposure level
and is ther efore discernible. In practice this featur e performs better on
some cameras than on others, but for many years never r eally gave a satisfactory result. With newer technologies such as DPS and improved digital
signal processing within the cameras, the problem of filming objects under
these conditions is becoming much easier to resolve.
Colour/mono cameras
In order to ensur e faithful colour r eproduction in a colour camera it is
imperative that no infrared light is permitted to enter the image chip and, to
The CCTV camera
143
Figure 6.14a During the evening an acceptable image of the person is obtained
Figure 6.14b During the daytime the iris closes down in proportion to the light
level around the doorway, resulting in the person having a silhouette appearance
this end, all colour cameras have an IR filter in fr ont of the image chip.
Some monochrome cameras also have such a filter in or der to pr event
overexposure during the daytime caused by the lar ge amount of IR pr esent in natural sunlight.
The necessity of the IR filter in a colour camera did, for many years,
limit their use to locations where the lighting conditions are relatively stable
and where artificial lighting is both in the visible light spectr um and is of
a reasonable intensity. This is why for many years only monochrome cameras could be employed where artificial IR lighting was employed. However,
more recent developments have spawned a new generation of cameras
that are able to function as both colour and monochr ome, depending on
the lighting conditions. These colour cameras ar e able to automatically
switch to monochrome operation when the light level falls below a predetermined value. To enable such cameras to function with IR lighting, when
they switch to monochrome operation, the IR filter is mechanically pulled
away from the CCD chip.
Such cameras are able to take advantage of the much mor e desirable
colour picture when the light level is sufficient, but can revert to the more
144
Closed Circuit Television
effective monochrome operation when low light conditions pr evail. Note
that because in black and white mode the IR filter is emoved
r
and the CCD
chip is functioning in monochrome mode, these cameras have much higher
resolution and sensitivity figures for monochrome operation than they do
for colour.
Early versions of these cameras tended to be somewhat unstable under
some conditions – for example, where car headlights intermittently illuminate the viewing ar ea – but impr ovements in performance during r ecent
years have made these cameras a popular choice for external operation.
Camera sensitivity
There are a number of ways in which the sensitivity of a camera can be
measured and quoted, which is not helpful to the engineer when trying to
decide between cameras of different manufacture, because the figures may
not be comparable.
A true sensitivity figure is the maximum F-stop which produces a 1 Vpp
video output when the camera is trained onto a specific test car d under
specified lighting conditions. The AGC must be switched off , and the manual iris is pr ogressively closed down to the point wher e the video signal
begins to fall from 1 Vpp. At this point the F-number is r ead, and quoted
as the sensitivity.
However, this is not the figure that is often quoted; rather, the minimum
illumination figure is used. This is a measur e of the lowest level of light
falling onto an object from which a ‘recognizable’ video signal can be produced. Typically these specifications appear in a catalogue something like,
‘1 Lux – 80% – F1.2’. This implies a light level at the object of 1lux, where the
object has a reflectivity of 80%, and an F-stop of 1.2 was used.
It is the definition of ‘recognizable’ which causes some difficulty, because
it is left to each manufacturer’s judgement. This is by no means implying
that all manufacturers are attempting to pull the wool over installers and
specifiers, but it does offer a window of opportunity to some who wish to
market what in truth are low spec cameras, whilst quoting high sensitivity
figures.
An important consideration when looking at minimum illumination
figures is to see if the measurement was taken with theAGC turned on or off.
With the AGC turned off the figure is going to be mor e realistic because
the camera will not have been working ‘flat out’ to produce the recognizable
video signal. Be careful when comparing one camera which has been specified with the AGC off, with another with the AGC on. This is not a tr ue
comparison because it is almost certain that the model with the AGC ‘on’
will show a much better figure that that with the AGC ‘off’, yet it may well
be that the camera with the AGC ‘off’ is the better of the two. Unfortunately
not all specifications state whether the AGC was on or off when the sensitivity was measured.
The CCTV camera
145
Camera resolution
The theory behind the resolution of a video image has been dealt with in
Chapter 5 where we saw that ideally a 625-line system should be able to
reproduce 780 horizontal pixels (Figure 5.4), but that in practice, because of
limitations in technology, we opted for something less during the design
of the original broadcast system. For a 625-line television system with 585
active lines, the maximum number of horizontal pixels along one line is
780 0.7(Kell factor) 546.
The resolution of CCTV cameras is quoted in TVL (television lines).
This figure refers to the number of horizontal pixels a camera can produce along a
distance equal to the height of the screen. This point is illustrated in Figure 6.15.
x cm
Height
x cm
Aspect ratio 4:3
Figure 6.15 TVL is measured as the number of horizontal pixels along a distance
which is equal to the screen height
In Chapter 5 it was stated that, ideally, the horizontal resolution should
at least equal the vertical resolution. Therefore if the maximum number of
pixels along a line is 546 (PAL), then the number of pixels along a distance
equal to the height will be 546 divided by the aspect ratio, which is
546 (4/3) 409.5. Thus for a PAL CCTV camera to produce a picture which
matches that seen on a broadcast transmission, it must have a resolution of at
least 410 TVL. Similarly, for NTSC the horizontal resolution in terms of TVL
will be 462 (4/3) 346, although the practical figure quoted is 330 TVL.
In practice other losses in the system often require the CCTV camera to
have a much higher resolution than the figures quoted above if the system
is to match broadcast performance.
It should also be noted that whatever r esolution is quoted, it will not
apply to the entire picture area. Neither the lens nor the monitor CRT produce full resolution around the edges of the picture, and when looking at
TVL figures it is prudent to assume that this performance will only apply to
the centre portion of the displayed picture. This is of particular importance
when selecting equipment for a specific operational requirement where the
TVL for each image at the control room end has been specified.
The subject of camera r esolution is consider ed in further detail in
Chapter 13, when we look at methods of measuring system resolution.
146
Closed Circuit Television
Camera operating voltages
The three common supply voltages for CCTV cameras ar
e 230 Va.c.,
24 Va.c. and 12 Vd.c. Although at first glance the operating voltage may not
appear too significant, the choice of camera supply has a direct bearing on the
installation material costs and labour because some supply methods er quire
more wiring than others.
For 12 Vd.c. cameras the supply is separately fed, meaning that every
co-axial cable must be buddied with a d.c. supply cable. In some cases this isnot
a problem; however, where there is a need for all cables to be hidden, losing
the extra one can sometimes prove difficult.
Another problem associated with 12Vd.c. systems is that of voltage drop
on cable runs exceeding (typically) 100 m. This phenomenon, along with
methods of overcoming the problem, was discussed in Chapter 2.
The 12 V supply is derived fr om a power supply, the rating of which
must be suited to the number of cameras being installed. The curr
ent
drawn by a camera is typically between 350 and 500 mA and to avoid overrunning a power supply, a useful rule of thumb is to look at two cameras per
1 A supply. From this it becomes obvious that for even a modest system a
large power supply is r equired. There are two schools of thought on this
subject. On the one hand a single 12V power supply rated high enough to
power all cameras can be employed. Onthe other hand a number of smaller
units can be used, having the advantage that if one unit fails, the whole
system is not put out of action. In practice it is not much more expensive to
take the second option, and the problem of voltage drop can be reduced by
dispersing the power supplies around the site.
230 Va.c. systems do not suf fer the problems of voltage dr op and the
need for large power supplies. This supply is particularly suited to external
camera applications, and is essential if a pan/tilt unit is employed as 230V
is required to operate the motor.
Another advantage of using the 50Hz mains as a power source is that the
line lock feature on the cameras can be employed to provide synchronization
(see Figure 6.11).
The main drawback of the 230 V a.c. supply is the r equirement of a
fused spur at each camera location which can often make the installation
more involved, especially wher e cameras ar e mounted on towers and
must be supplied using an underground steel wire armoured cable. In the
UK, current regulations require this work to be carried out by a competent
person and final inspection and testing of the cir cuit must be performed,
with a certificate of compliance issued to the customer by the inspector, in
accordance with BS 7671. In other words, if the installer is not a qualified
electrician, the electrical installation work will need to be subcontracted out,
or the installer will run the risk of prosecution in the event of any mishap.
Another problem encountered with mains power ed systems is that of
ground loops, although this phenomenon is usually rectified using ground
loop correctors. This was discussed in Chapter 2.
The CCTV camera
147
For internal use, 24 Va.c. cameras are becoming increasingly popular.
Being defined as extra low voltage (ELV), 24 Va.c. does not come under the
same regulations as 230 Va.c., yet overcomes the problems of voltage drop.
As with 12 Vd.c. a separate a.c. power supply unit is required; however, some
switchers incorporate a limited a.c. supply which is sufficient to operate a
small system.
Some cameras are compatible for both 12 Vd.c. and 24 Va.c. operation.
These cameras generally operate at ar ound 9 Vd.c. internally and use an
internal d.c. to d.c. converter power supply to reduce the 12 Vd.c. input, and
an internal bridge rectifier and smoothing circuit rectify the 24 Va.c. input.
A variation of the low-voltage d.c./a.c. camera is the line fed system
where the power to each camera is fed down the co-axial signal cable. The
cameras are supplied from a 24 Va.c. supply contained within a dedicated
switcher/controller. The video signal is superimposed onto the a.c. supply
and the two are separated using filter networks within the controller. This
system makes installation very simple, and even lends itself to the DIY
market. However, the need for dedicated controllers and cameras means that
there is little flexibility in systemdesign, and the system cannot be extended
beyond its maximum camera capacity.
Specialized cameras
In addition to the vast choice of monochr ome and colour cameras available, advances in both CCD and digital technology have spawned a range of
high-performance specialized cameras, particularly in low light and covert
designs.
As we have seen, comparing like for like, monochrome cameras always
outperform their colour counterparts in terms of low light ability and resolution, and high-performance designs are available with minimum illumination levels well below 0.1 lux. Thus the installer looking for a camera
which will perform well in an environment with minimum illumination has
never had such an easy task. Even with colour cameras the situation is changing rapidly. For many years it has been generally accepted that colour
cameras are unsuited to external use because of their poor low-light performance, and the fact that they would only produce correct colours if the
area was illuminated in such a way that it resembled a TV production set! Yet
new designs in CCD image chips, along with the production of dedicated
digital processing ICs, have enabled surveyors to re-think.
Nevertheless there are still cir cumstances where a quality image is
required in situations where any form of artificial lighting, visible or otherwise, is unacceptable. In these circumstances a device known as an image
intensifier may be employed.
Image intensifiers are not new, and have been used for many years to
enhance the performance of cameras, including tube types which always
struggled to operate under low light levels. They function by ef fectively
148
Closed Circuit Television
amplifying the amount of incoming light using electr
basic construction is shown in Figure 6.16.
Light input
Lens assembly
Photosensitive
cathode
onic means. The
Phosphor coated
anode (15 kV)
Focusing
electrodes Electron emission
Camera
Figure 6.16 Principle of the image intensifier
The incoming light is focused onto a photosensitive cathode, the coating
of which has the property of emitting electrons when impacted by photons.
The electron emission from any point on the cathode is directly proportional
to the amount of light falling onto that point, and thus ther e is a constant
emission of electrons from all points on the cathode, representative of the
image produced by the lens.
The electron cloud is attracted towar ds the anode by a high-voltage
potential. As the cloud passes through the device it is focused electronically
onto the anode plate, which is phosphor-coated in much the same way as a
cathode ray tube in a monochrome monitor. Thus a picture of the original
image is produced in front of the camera imaging device; however, it is many
times brighter than that produced by the lens.
Use of an image intensifier makes it possible to obtain a clear pictur e
with nothing more than starlight illumination. However, the cathode and
anode materials have a limited operational life (typically 2000 hours), making image intensifiers a somewhat expensive option in any CCTV system. To
obtain maximum lifespan, a high F-stop lens should be fitted.
Although image intensifiers are still in use, owing to their limited lifespan
a more reliable alternative would be most welcome, and it is expected that
developments in CCD technology will make them largely redundant.
Covert cameras
Covert cameras may be employed for any number of er asons, and not solely
by the security forces. They are available for use by r etailers, employers,
entertainment centres and domestic homeowners alike. Some examples of
use are: recording the activities of an intruder once they have defeated the
overt CCTV system, pr oduction of video evidence which may be later
used by either prosecution or defence counsels, installation in preference to
overt equipment in order to maintain the aesthetics of the premises.
The CCTV camera
149
Until the advent of the CCD image chip covert cameras were relatively
bulky and difficult to hide. Nevertheless, covert tube cameras were successfully employed for many years. Of course it is not just the camera which
needs to be compact; the lens must also be unobtrusive, and it is this which
is perhaps the more difficult to achieve because the words ‘miniaturizing’
and ‘optics’ simply don’t go well together. As we saw in Chapter 4, ideally a
lens must be able to collect and process as much light as possible, and this
means using large optical components. Having said this, small lenses offering remarkable performance have been developed for covert use.
The most common type is the pinhole lens, although this name is somewhat misleading because these tend to have a diameter in the or der of
3.5–8 mm. They are generally available in 13 and 12 formats and have a CS
mounting making them compatible with any matching format CS mount
camera. Additionally, pinhole lenses ar e available in a right-angled construction to enable the camera to be hidden in locations wher e there is
restricted space. This is illustrated in Figure 6.17.
Another type of lens developed for covert use is the mini lens which, as
the name implies, is both short and narr ow. However, the small size is
achieved at the expense of certain components, and mini lenses generally
have no iris, meaning that they ar e only suited to applications wher e the
lighting conditions are reasonably stable. Note that a lot of mini lenses do not
invert the image at the output, so the camera must be mounted upside d
own.
When selecting a pinhole or mini lens, the installer should try to choose
one with the lowest possible F-number in order to attain the highest optical
speed, bearing in mind that the optical speed will be restricted by the miniaturized optics.
Fibre-optic lenses are available for use in situations wher e it is necessary to ‘see’ through thick walls or perhaps beyond a wide void. Both rigid
and flexible types are available. The rigid type simply fixes to the front of a
camera in the same manner as a conventional lens assembly,the light being
carried from the small front optical device to the main lens assembly via a
cluster of fibre-optic strands contained within a narrow tube. The flexible
type allows the front optical plate to pass thr ough more intricate spaces,
and the camera does not have to be fixed adjacent to the field of view.Due to
a number of types of losses within the fibre-optic strands, these lenses are not
as sensitive as glass pinhole types.
Of course, all of the covert camera/lens arrangements so far discussed are
expensive, which puts a lot of would-be users of f. However, with recent
developments in low-cost miniatur e covert combination camera/lens
assemblies, the number of covert installations has risen considerably
.
Popular off-the-shelf covert CCTV products are the passive infrared (PIR)
cameras (some of which include a working PIR), clocks, mirrors and domestic smoke detectors. ‘Bare bones’ PCB cameras are also available which can
be mounted onto almost anything. Many of these ar e designed to be connected using standa rd four-core intruder alarm cable, which is far mor e
flexible than co-axial cable and, despite the lack of scr
eening and the
150
Closed Circuit Television
Figure 6.17 Typical pinhole lenses for covert applications. A right-angled lens
may be used where a covert camera is to be installed in a restricted space.
(Photos courtesy of CBC (Europe) Ltd, manufacturers of Computar lenses)
mismatch in impedance, the transmission results are usually quite acceptable, although most manufacturers do not recommend cable runs beyond
100 m. The camera modules are generally designed to operate at 12V, which
is delivered along two of the four cores.
360° cameras
Another technology that has become available as a r esult of modern processing power is a camera that has a 360° fish eye lens fitted, but is able to
provide multiple corr ected 4:3 images as though it wer e a number of
separate cameras. The process includes a combination of high-r esolution
image sensors, complex mathematical algorithms and high-speed image
processing. The detailed 360° panoramic digital image from the camera is
‘de-warped’ (straightened) and corr ected to a 4:3 or multiscr een image,
depending on user and application preferences. The images may be transmitted as a digital signal to a PC or NVR, or in analogue format as a standard PAL CCTV signal which can be monitor ed and recorded as normal.
Figure 6.18 shows a typical camera head, along with the 360°panoramic and
a quad display of four corrected images from that same panoramic.
Using a keyboard, the operator is able to view the selected areas, change
the selected ar eas, pan ar ound the 360° panoramic, or include the 36 0°
panoramic within the screen multiplex. Alarm inputs are often available to
cause preset areas to be called up. In some cases it is also possible to pr ogramme the unit to follow moving objects.
The example shown in Figur e 6.18 shows a self-contained unit that
includes all of the pr ocessing at the head end. An alternative to this is to
The CCTV camera
151
Figure 6.18a Camera assembly with a 360° fish eye lens producing a panoramic
field of view. An internal processor in the head assembly performs image
correction and produces a number of multiplex output options
Figure 6.18b Panoramic view produced by the lens when the camera is ceiling
mounted
152
Closed Circuit Television
Figure 6.18c Typical example of the video output available from this camera.
Four corrected images in quad display created from the original panoramic
image. It would be easy to believe that there are four separate cameras in
the installation (Photos and images courtesy of Vista)
employ a conventional CCTV camera with a 360°fish eye lens fitted and pass
the panoramic image directly to a dedicated PC that is running a proprietary
software. Image correction and multiplexing is performed by the PC, where
the video information is also often recorded and stored. Although effective
in terms of their ability to produce multiple, corrected images from a single
panoramic, such solutions ar e not always as versatile as the head end
solution because too often they are less able to integrate with other CCTV
hardware.
The advantages of using these 360° cameras is that one unit can replace
a multiple of static cameras, installation is greatly simplified, one single dome
is aesthetically more acceptable than numerous static units, and the system
can be made to record the 360° panoramic with the ability to view anything
within that field at a later time. A typical application example is shown in
Figure 6.19.
Number plate recognition cameras
Often referred to as vehicle number plate recognition (VNPR) or automotive number plate recognition (ANPR), these cameras are actually standard
CCTV cameras. It is the softwar e associated with them that performs the
task of reading the number plate.
The CCTV camera
153
360 degree dome camera
Figure 6.19 A single 360 degree, image corrected dome unit. This can replace
multiple static cameras yet can achieve a greater range of image fields
These cameras may be divided into two gr oups: those that use a standard visible light spectrum camera, and those that have an IR filter fitted in
the lens.
Visible light versions simply provide a video feed to a PC that is running
a software package which picks out the distinguishing features of a number plate on a vehicle. The softwar e then displays (usually) a still shot
and/or a short video clip of the vehicle, a close-up shot of the plate, and a
computer-generated display of the plate. Pr ovided that the camera is set
up correctly, it is usually possible to identify the driver of the vehicle.
From the point of view of camera installation, for fixed installations the
specifier or engineer only needs to consider the viewing angle and the
required focal length of the lens, and (of course) the method of transmitting
the video signal back to the monitoring centre.
In the UK, the fluorescent background on a vehicle number plate reflects
infrared light very efficiently so, as daylight turns to dusk, the number plates
may be illuminated using an infrared lamp mounted with the camera.
The one problem with visible light VNPR is that, because the camera has
to look more or less directly at the vehicle, the vehicle headlamps usually
blind the camera causing the iris to close, making the plate indistinguishable.
This problem may be overcome by using an IR filter lens. The camera is
able to function in the daytime using natural IR light, and in the evening
an IR lamp at the camera location switches on to maintain functionality ,
without the camera being af fected by the headlamps because they ar e
blocked by the filter.
The one major drawback with using IR is that the system can only provide a number plate. There is no useable video footage relating to the vehicle type or colour, or an image of the driver. To overcome this issue, VNPR
systems normally take a feed from two cameras, one operating in the visible
154
Closed Circuit Television
light spectrum, the other in the infrared. The visible image is used (where
possible) to provide the video footage, and the IR input is simply used to
derive an image of the number plate.
Typical images derived from both visible and infrared VNPR cameras,
along with a digitally derived number plate, are illustrated in Figure 6.20.
The most common application for VNPR/ANPR is motor vehicle speed
measurement, although it is frequently used for controlling barriers in private car parks. However, the technology is really only a simplified version
of the more complex biometric video analysis softwar e solutions that are
becoming increasingly more accurate and ef fective, so it is clear that the
technology will continue to evolve, and uses beyond vehicle plate r ecognition will be developed. Already in the UK, the same technology is being
used to detect and read the vehicle tax discs on moving vehicles as well as
the number plate. Other systems ar e being developed to aid the policing
of the proposed policy of only permitting vehicles to use the ‘fast lane’ of
a motorway if they are carrying two passengers – the cameras would be used
to detect the presence of two persons (although this technology is, at the
time of writing, somewhat controversial as it would, for example, be very
difficult to prove the presence of a child on the rear seat). VNPR is also one
method being looked at for the administration of toll o
r ad charges, although
other ‘vehicle tagging’ methods are also being looked at in relation to this.
Figure 6.20a Image from a VNPR camera with a visible light filter, taken during
the daytime. The natural IR light provides sufficient illumination to derive an
image of the number plate.
The CCTV camera
155
Figure 6.20b The same vehicle filmed with a colour camera (located next to the
IR camera) fitted with a standard lens provides visible footage which may be
used to identify the vehicle and (possibly) the driver.
Figure 6.20c Vehicle number plate, digitally derived from the images from the
IR camera. (Photos courtesy of Vista)
7 Video display equipment
In Chapter 6 we saw how the camera converts light energy into electronic
signals relating to the luma and chr oma content. In this chapter we will
look at methods of converting these signals back into light energy, creating
the illusion of moving coloured pictures.
For many years the sole device for producing coloured pictures for both
the CCTV and domestic television industries was the cathode ray tube, or
CRT. The vision to have a flat panel display has been ar ound for many
decades (in the 1960s Star Trek series, look at the futuristic flat panel ‘main
screen’ that captain Kirk had on his bridge, a telling sign of wher e sci-fi
buffs thought we would be in terms of video displays); it was the technology that was lacking.
Manufacturers, in particular Sharp Electr onics, began working on LCD
display devices a few decades ago; however, it was not really until the turn
of the millennium that such display devices wer e becoming practical for
displaying video images. There is a rapid shift towards flat panel (and in
particular, LCD) displays because, we are told, they are ‘better’. However,
we must define the term ‘better ’. Compared with a CR T monitor, a flat
display takes up less space, consumes less power , produces less heat, is
lighter to transport, and looks modern. However, comparing like for like,
a CRT monitor will provide much higher resolution and brightness levels
for a fraction of the price. There is no doubt whatsoever that flat displays
will equal if not outperform CR Ts in the futur e. However (at the time of
writing), where a high-r esolution image is desir ed, a high-quality CR T
monitor is still the best choice.
In this chapter we will look at each of the curr ent display technologies
and, where possible, compare their strengths and weaknesses. As it is the
oldest form of TV display device, we shall begin with the CRT.
The cathode ray tube
People often refer to the thermionic valve as a thing of the past, something
that went out during the 1970s with the rapid intr oduction of the transistor and silicon chip. Yet the cathode ray tube is a thermionic valve, as the
diagram of a monochrome CRT in Figure 7.1 clearly shows. The operating
principle is quite straightforwar d. The cathode, which is connected to a
positive d.c. supply, is heated up. This causes the electrons present within
its molecular structure to accelerate to an escape velocity, creating a negative space charge in the area just in front of the cathode.
Video display equipment
157
Glass
face
Graphite coating
Line and field scan coils
Phosphor
Aluminium
layer
Second anode
Control grid
Cathode
Heater
Electron
beam
Space charge
First anode
Third anode
Fourth or final
anode
Figure 7.1 A monochrome CRT
A very high voltage, referred to as the extra-high tension (EHT), is connected via the final anode to the scr een. The size of this voltage depends
upon the screen size, but for a 30 (12) monochrome tube, the EHT is typically 10 kV. If there were nothing between the cathode and the screen, the
negative electrons in the space char ge would make their way towar ds the
EHT at the screen, and return to their power source via the final anode connection. Because electrons are moving from the cathode to the screen, we
can say that there is a current flowing through the CRT. It is important to
note that current flow only takes place because the tube has been completely evacuated of air . Otherwise the electr ons would impact with the
molecules of air and rapidly lose their momentum, causing them to fall
back to the cathode once again.
The problem is that the electrons would not be travelling fast enough for
our purposes, and furthermor e, they would be spr eading out as they travelled towards the screen due to the r epelling effect of their like char ges. To
overcome these problems, anodes 1, 2, and 3 are added. The first anode (A1)
is connected to a d.c. potential of around 100 V. This has the effect of attracting the electrons away from the cathode, giving them velocity in the dir ection of the screen. The electrons are travelling so fast by the time they reach
this anode that they pass straight through a hole in its centre, and move at
high velocity towards the EHT potential. Another name for the first anode
is the accelerator anode. Unfortunately it can also be r eferred to as the
screen grid. This is a name which goes back to the days of the valve, but
today it can cause confusion for the unsuspecting engineer who might
think that it makes reference to the screen on the front of the CRT.
158
Closed Circuit Television
Anodes 2 and 3 form an electrostatic focusing lens. Focusing is necessary
to counter the divergence of the electron beam caused by the electrons moving apart. Anode A2 is connected to the EHT, whilst A3 is connected via a
variable resistor (focus control) to a potential of around 200 V. These two
largely differing potentials produce an electrostatic field between anodes
A2 and A3. Altering the potential on A3 changes the shape and strength of
the electrostatic field, and hence the ef fect on the electr on beam passing
through the centre. The principle is illustrated in Figure 7.2. The focus control is adjusted until the angle of the converging beam is such that the electrons all hit the same point on the scr een. In other words, the picture will
appear at best focus.
EHT
Electrostatic field
A2
Diverging
electron beam
A3
Converging electron
beam
Focus control
200 V
Figure 7.2 Principle of electrostatic focusing. Cross-sectional view; the anodes
are three metal cylinders
The CRT screen is coated with phosphor, which glows when hit by the
high-velocity electrons. Behind the phosphor is a thin aluminium layer ,
which serves both to protect the phosphor from burn, and act as a reflective screen to ensure that all of the light produced by the phosphor is projected forwards.
The assembly comprising the heater, cathode, control grid, and anodes
is known as the electron gun assembly, although as we have seen, it does
not actually fire electrons but rather it emits them as a r esult of the high
EHT potential on the CRT viewing screen.
So we see how a CR T produces an electron beam current which flows
from the cathode to the final anode, pr oducing a small white spot in the
centre of the screen. We shall now look at how the brightness of the spot
can be controlled.
The brightness is dependent on the number of electrons hitting the phosphor, so by controlling the beam current we can control the brightness. This
function is performed by the control grid, the action of which is illustrated
in Figure 7.3.
Video display equipment
159
Cathode
Negative space charge
80 V
Electric field
Control grid
Brightness control
100 V
Cathode
potential
Grid
potential
Grid potential
with respect
to cathode
Effect
80 V
80 V
0V
80 V
50 V
30 V
80 V
20 V
60 V
None. Maximum current
through the grid aperture.
Size of aperture reduced
by the effective negative
charge (w.r.t. cathode).
Beam current reduced.
Grid is so negative that the
aperture is closed. Zero
beam current; no display.
Figure 7.3 Electric field produced by the grid/cathode potential effectively
closes the aperture in the grid
If the cathode is maintained at a constant potential, and the control grid
is set at the same potential, ther e is no ef fect on the electr on beam and
maximum brightness still exists. However , making the grid less positive
causes it to appear negative to the electr ons leaving the cathode, r esulting in those electr ons close to the sides of the apertur e being r epelled.
However, some electrons still pass through the centre of the aperture, and
thus the beam current and brightness are reduced. When the grid potential
becomes approximately 60 V lower than the cathode, the r epelling action
is so strong that the aperture in the grid is effectively closed. All beam current ceases, and there is no spot on the screen. Connecting a variable resistor to the control grid provides a means of brightness control.
The beam current can also be controlled by varying the cathode potential. Increasing the cathode voltage makes the grid appear more negative,
160
Closed Circuit Television
causing a reduction in brightness. Reducing the cathode voltage makes
the grid appear less negative, and the brightness incr eases. Thus, if the
video signal waveform is applied to the cathode, the electron beam will be
modulated by the rapidly changing video signal, pr oducing brightness
levels on the screen that are proportional to the video signal level.
For the CRT to produce a picture, the electron beam is made to deflect
both vertically and horizontally at high speed, causing it to scan the screen,
producing a blank white display known as a raster. Applying a video signal waveform to the cathode causes the beam current to be modulated, producing a picture as the screen is scanned.
Deflection of the electron beam is performed by the line and field scanning
coils which are placed around the CRT neck. Alternating currents are passed
through the scan coils which in turn set up alternating electromagnetic fields.
These fields interact with a magnetic field that exists ar ound the electron
beam, and thus deflection takes place. The line scan coils perform horizontal deflection. The field scan coils perform vertical deflection.
The graphite coating inside the CR T is used to form connections
between the final anode, the CRT screen and the second anode.
The colour CRT
In Chapter 3 we saw that white light is made up fr om the three primary
colours: red, green and blue. Thus, for a colour CRT to work effectively it
must reproduce these three colours, which it does by having thr ee separate cathodes, each driven by separate red, green and blue signals derived
in the colour camera.
The colour CRT screen is coated with thr ee different types of phosphor
which emit different frequencies (colours) of light when struck by electrons,
these colours being red, green and blue. The phosphors are laid on the screen
face in a tight pattern, illustrated in Figure 7.4. Each electron gun is targeted
at just one set of phosphors, and so in ef fect we can say that one gun is creating the red light output, another the green, and another the blue.
One triad of phosphor stripes 1 pixel
G
G
R
B
G
R
B
G
R
B
G
G
G
R
R
R
R
B
G
R
B
G
R
B
G
R
B
B
B
B
Figure 7.4 Typical phosphor stripe formation on a colour CRT screen
Video display equipment
161
When all three guns are emitting electrons, all the phosphors are illuminated and the eye r eceives red, green and blue light. However , because the
phosphor spots are so small, at normal viewing distance the eye is unableto
discern them and the brain is tricked into thinking that it is seeing a white
screen. The individual spots are visible if you move very close to a colour
monitor screen and focus your eye on an area of white picture content.
Colour can be introduced by turning the guns on and off. For example,
if the blue gun is turned off, the eye receives only red and green light and,
as we saw in Chapter 3 (Figure 3.3), the brain will interpret this as yellow.
The principle of the colour CRT is shown in Figure 7.5. The actual operation is a little mor e complex than this diagram r eveals because special
magnets are required around the tube neck, either external or internal, to
ensure that each electron beam only ever strikes its own colour phosphor.
This is called convergence of the CRT, but it is not necessary to look at this
in any more depth as it not something that the CCTV engineer will become
concerned with.
Control grid
3 first anode
(usually connected
together)
Single focusing unit as
shown in Figure 7.1
R
h
G
B
h
R
Cathode (3)
G
B
Three separate electron beams are converged
to illuminate just one triad of phosphor spots at
a time, producing a single pixel
Figure 7.5 Principle of a colour CRT
The other main dif ference between colour and monochr ome tubes is
the operating voltage. For a colour CRT it is generally higher than for their
equivalent size monochrome counterparts. The cathodes typically require
a signal drive voltage of around 80 Vpp (swinging between 70V and 150 V),
the first anode potential is ar ound 300–500 V, the focus is appr oximately
1 kV, and the EHT can be approximated as being equal to 1 kV per inch of
screen size, and so for a 24 colour CRT the EHT will be around 24 kV.
CRT monitors
Generally speaking these have two video inputs: composite (CVBS) orY/C,
the latter being included to accommodate an S-VHS input. BNC connectors
162
Closed Circuit Television
are normally used for the CVBS input/output, both because of their robustness and because they are designed to maintain the 75 impedance of the
transmission system, provided of course that the correct BNC type is used.
BNC may be employed for the Y/C inputs; however, SCART or four-pin
S-VHS connectors are commonly used.
Some monitors use RCA (phono) connectors for CVBS input/output,
but this can pr esent a problem when the monitor input cable is co-axial
because RCA connectors, which were originally intended for audio use,
do not fit onto co-axial cables such as RG-59. Thus a BNC to phono adapter
must be used which cr eates an added connection in the transmission
medium.
Owing to both manufacturing tolerances and viewing preferences, it may
be necessary to perform a number of adjustments to a monitor during installation. Therefore we will look at the typical adjustments and their function.
In practice most adjustments ar e performed in softwar e using a menu;
however, some are still performed the more traditional way using a terminal
screwdriver.
There is often much confusion as to the dif ference between brightness
and contrast. The brightness contr ol moves the level of the entir e image
such that dark and light ar eas of the image all become lighter or darker .
The contrast setting moves the light and dark areas towards or away from
each other. The action of the two controls is illustrated in Figure 7.6.
V
Bright
B
Dark
A
0
Figure 7.6 Effect of brightness and contrast control. Brightness alters the d.c.
level (A), moving the complete waveshape up or down without altering the
amplitude. Contrast alters the peak–peak value (B)
t
Video display equipment
163
For example, turning the brightness level up fully r esults in the black
areas becoming dark gr ey and the pale gr ey areas becoming white, with
the white areas becoming excessively bright. Turning the contrast level up
makes the dark grey areas become black and the light gr ey areas become
white. Users will often adjust these to suit their own preferences; however,
the service engineer should be aware that a CRT that is operating continually with the contrast set very high will suf fer CRT failure very quickly
(typically after two years). Also, in a CCTV control room the user may be
complaining that the same images appear totally different when viewed on
different monitors. This could be due to CR T ageing on some of the monitors, or the fact that they are of different manufacture, but it may be that the
brightness and contrast settings at each monitor are set very differently.
In Chapter 5 we discussed the need for synchr onization between the
cameras and the monitors. Although modern manufacturing tolerances have
removed the need for vertical and horizontal timebase adjustments, these
controls are still found on some models. Incorrect adjustment of the vertical hold will result in field roll or picture bounce. If the horizontal hold is
maladjusted, the picture may tend to pull to the right, or br eak up into a
series of horizontal black lines. These effects are illustrated in Figure 7.7.
Vertical hold maladjusted
Horizontal hold maladjusted
Figure 7.7 Effects of maladjusted timebase controls
164
Closed Circuit Television
The vertical and horizontal shift controls are used to centre the picture on
the monitor screen. The CRT first anode adjustment, also referred to as the
screen or A1, is usually made available through the rear of the monitor. It is
not recommended that this adjustment is performed by the CCTV engineer
as the correct set-up procedure is somewhat involved. It requires removal of
the cabinet, and a voltage measurement taken at the A1 terminal on the CRT
whilst adjusting the control. The A1 appears to behave as a brightness control; however, it is not to be used for this function, and maladjustment of this
can result in impaired picture quality, and a possible reduction in CRT life.
Adjustment of the focus control is also normally made through the rear
of the monitor. The monitor should be made to display a pictur e containing a high degree of detail, especially in the centr e, and the signal source
should be known to be sound; in other words, the camera focus should be
correct! The focus contr ol is then adjusted for the sharpest pictur e at the
centre of the screen.
Two other contr ols which may be available ar e raster correction adjustments. All monitors and TV r eceivers suffer from an effect known as pincushion distortion. This picture distortion, illustrated in Figure 7.8, appears at
the edges of the display and is caused by differences in the distance between
the point of origin of the electr on beam (i.e., the CRT cathode) and points
along the sides of the screen. Circuits within the line and field output stages
are included to correct the effect, and on larger monitors especially, external adjustments are made available.
The monitor requires very stable d.c. supply rails, and for all colour monitors these are provided by a type of power supply known as a switched
mode power supply (SMPS), which is employed because of its high efficiency.
Figure 7.8 Effect of pincushion distortion. Without raster correction, the edges
of the picture become bowed
Video display equipment
165
The 230 Va.c. mains supply is full-wave rectified and smoothed to produce
320 Vd.c. This d.c. is then switched at a high frequency (40–80 kHz) through
a transformer primary winding. Voltages induced in the secondary windings
are rectified and smoothed to produce a variety of high- and low-voltage
d.c. supplies.
The principle of the SMPS is shown in Figure 7.9, which illustrates a very
important feature of this type of power supply: mains isolation. It is essential
that the 0 V side of the d.c. power supply in a CCTV system is isolated from
the a.c. mains supply. In items of equipment such as cameras, contr ollers
and multiplexers, this isolation is performed by the action of the mains
transformer. However, in the monitor it is possible for the 0 V line to find
a return path to the a.c. mains supply thr ough the bridge rectifier. If this
were to occur, the chances are that the earth connections to the CCTV system would cause fuses in the monitor to r upture, as well as the RCDs to
trip. However, if an earth fault were to exist in the system, all metal parts
of the CCTV system may become live, with possible fatal consequences.
Chopper transformer T1
P
N
Rectifier
and
smoothing
320 Vd.c.
150 V
120 V
50 V
820 k
0V
12 V
regulator
Control
circuit
Electronic
switch
Feedback
‘Live’ area
Isolated area
Figure 7.9 Principle of the SMPs, illustrating the isolating effect of the chopper
transformer
By using the arrangement shown in Figur e 7.9, the 0 Vd.c. line is completely isolated from the ‘live’ area by the transformer action. Even the feedback from the output to the control circuit, which is necessary to maintain
a stable d.c. output, is isolated by means of an optocoupler . It is essential
that no attempt is made to modify the power supply section in a monitor as this isolation may be compr omised, and all servicing work in this
166
Closed Circuit Television
section must be carried out by a qualified television service engineer, whose
role and training is somewhat different from that of a CCTV engineer.
A word of warning: when a monitor is in stand-by , the SMPS is not
switched off; it simply reduces the 150V supply to such a level that the monitor is no longer able to function. This is done so that the 5 V supply to the
CPU chip is maintained in order that it is able to receive and process a command to come out of stand-by . If an engineer wer e to remove the monitor
covers whilst it is in the stand-by mode, ther e is the very r eal hazard of
receiving a fatal shock from the 320 V rectified d.c., or the a.c. mains input.
Monitor safety
Monitor servicing should be referred to a qualified television service engineer, and the CCTV engineer should only perform the picture adjustments
outlined in this chapter . However, there are occasions when r emoval of
the monitor cabinet is necessary, and in such cases the engineer must be
aware that this immediately exposes both him/herself and anyone in the
area to the possibility of a lethal electric shock. Supplies such as the 150Vd.c.
rail are able to deliver more than sufficient current to kill, and are especially
lethal if applied across the chest, which can easily happen if the engineer
has one hand on the chassis to hold it steady , and the other hand slips and
comes into contact with a component connected to the high-voltage supply.
This is illustrated in Figure 7.10.
Power
supply
Equivalent
circuit
Figure 7.10 Current flows from the positive supply (right hand), through the
body and back to the negative supply (left hand)
Video display equipment
167
In this illustration the engineer has his left hand on the negative metal
case (or chassis) of the equipment on which he is working. If his right hand
were to slip and touch a high-voltage contact, a curr ent path would exist
between his hands, taking a path straight acr oss his chest and ther efore
the heart. It has already been stated that this is a very dangerous situation.
Whenever working on a live monitor a ‘one-handed’ appr oach should
be adhered to. This means that the engineer never has both hands on or in
the monitor at the same time whilst it is connected to the mains supply ,
thus reducing the chance of electric shock acr oss the chest. This point is
illustrated in Figur e 7.11, which shows the engineer working with one
hand away from the chassis (his hand is on the bench; however, it could be
placed on the plastic or wooden cabinet of the equipment). In this case
there is no circuit and hence there can be no current flow.
Power
supply
Equivalent
circuit
Figure 7.11 There is no circuit so current does not flow, provided that the bench
top is insulated
Of course there is still a chance of shock from the supply to earth, which
brings us to the second important point on safety when working on a live
monitor. When working on live equipment connected to the mains supply , the
equipment must be connected via a mains isolation transformer . For a small
colour monitor a rating of 500 VA is sufficient; however, for larger screen
models a 750 VA transformer must be used.
It is not usually practical to work on a monitor on site. 230 Va.c. isolation transformers are not a usual item in a CCTV service engineer ’s kit,
one-handed working can be dif ficult in what are often confined working
conditions, and it can be difficult to ensure the safety of control room/area
168
Closed Circuit Television
staff whilst the monitor is exposed. Consequently defective monitors ar e
normally exchanged and serviced in a workshop.
In Europe, following servicing a monitor must be P AT-tested to prove
its integrity.
One point which the CCTV engineer must always test is the integrity of
the earthing on monitors which have a metal case. This includes not only
the earth in the mains connector plug, but also the earth in the mains supply
socket to which the monitor is to be connected. In a worse-case scenario a
faulty monitor without an earth can cause not only its own case to become
live, but the entire CCTV system including metal camera housings, mounting brackets, etc.
Liquid crystal displays (LCDs)
When it comes to an alternative to the smaller size CRT, the liquid crystal
display is without doubt the most popular at the moment, mainly because
it is much less expensive than its closest rival the plasma display (which
we shall discuss later in this chapter). For the computer world the LCD is
definitely a worthy contender to the CR T, but this is because computer
graphics are not fast moving, which means that the r efresh rate of the
pixels does not have to be very high, and the user does not normally desir
e
a very wide viewing angle. Once we r equire a viewing angle of ar ound
200°, a refresh rate high enough to reproduce fast-moving objects without
blur, and a high contrast ratio and brightness level, then only the mor e
expensive and elaborate LCD panels begin to come close.
Early display devices employed a liquid crystal type known as dynamic
scattering mode (DSM); however, these were not very efficient and were
soon replaced by the twisted nematic (TN) type. TN crystal display devices
are still the most popular , although other variants have evolved such as
STN (super twisted nematic) and TSTN (triple STN). The basic TN device
loses contrast as the number of scanning lines is increased, bearing in mind
that PC monitors may produce many more scan lines than the 625 employed
in television. STN and TSTN are capable of high contrast ranges at high line
rates, even on larger display devices, although TSTN was really developed
for large-screen monochrome display devices.
Liquid crystal is a substance that falls between a solid and a liquid. The
long crystal molecules form up into a helix structure, and this has the effect
that any light passing though the crystal experiences a polarity shift of 90°.
When a voltage potential is placed acr oss the crystal substance, the helix
structure breaks down and incident light passes through unaffected.
LCD panels may be divided into two types: reflective and transmissive.
Reflective panels use the control voltage applied to the crystals to determine
the amount of light that is r eflected off the crystals, whilst transmissive
panels control the amount of light that passes through them. An example
of a reflective display would be a pocket calculator wher e incident light is
Video display equipment
169
either reflected back off the crystals, or passes through them and is absorbed
by the black back plate. TV display panels employ transmissive technology where light from a back light is either permitted to pass thr ough the
crystals, or is blocked by their action. One important point to note is that,
unlike CRT or plasma displays, LCD does not pr oduce any light and its
operation is totally reliant on an external light source.
The principle of operation of an LCD panel is shown in Figure 7.12. If we
ignore the crystal cells for a moment, light from the back light is subjected to
the action of two polarizing glass plates. The vertical polarizing plate will
only pass light waves that are vertically polarized, and the horizontal polarizing plate will only pass horizontally polarized light waves. Thus, the light
output at the front screen will be zero because the first plate will block the
horizontal light waves, and the second can only pass horizontal waves.
Vertical light blocked
by horizontal polarizing
plate
Some light blocked
by horizontal
Front
polarizing plate
screen
Horizontal polarizing glass
(passes horizontally polarized light)
Data input
(processed video signal)
Cell charge store capacitor
a
b
c
Liquid crystal filled pixel cells
Vertically polarized light
Vertically polarizing glass
(passes vertically polarized light)
Non-polarized light
Backlight
Figure 7.12 Operating principle of an LCD display panel
Via the data input, the liquid crystal cells are individually charged to a
voltage potential that is relative to the video signal level for each pixel in the
display. Consider pixel cell ‘a’ in Figure 7.12. When no charge is applied to
the cell the helix structure will be intact and light passing through the cell
is subjected to a 90° polarity shift, making it horizontally polarized. This
means the light will be able to pass through the horizontal polarizing plate
and a bright light output will be observed at that pixel location. At cell ‘b’
170
Closed Circuit Television
in Figure 7.12, a maximum char ge is applied, the helix str ucture breaks
down, the light passes through the cell unaffected, and it is therefore blocked
by the horizontal polarizing plate. This equates to a black pixel location. For
cell ‘c’, a 50% charge level causes a partial breakdown of the helix structure,
meaning that the light is moved through about 45° and some of it will pass
through the front plate. This r epresents a 50% light output at this pixel
location.
In order for the cells to r espond to rapid changes in brightness level, the
switching devices that control the charge/discharge of the capacitors in the
cells must be capable of very fast switching times. The devices used for such
switching action are transistors that are fabricated on a thin film – hence their
name, thin film transistors or TFT. Earlier LCD panels that did not employ
TFT suffered a very slow r esponse time, which resulted in blurring of fastmoving objects and a reduction in contrast level. There are a number of variations of TFT but the most common is the amorphous silicon type because of
the relative ease of fabrication. Another type of TFT is polycrystallinesilicon,
which offers much greater refresh rates than amorphous silicon, although
it is very difficult to fabricate, especially for large panel sizes. These devices
have found a niche for use in LCD video pr ojectors where high resolution
and fast refresh are required, but the panel size need not be very large.
We have seen up to this point how an LCD display can produce a monochrome image. To produce colour we use a principle that is almost the
same as that employed in the colour CR T. Remembering that we r equire
the three primary colours (R, G and B) to r eproduce the colour spectrum,
the LCD pixels are arranged in groups of three, with a filter placed in front
of each pixel. This r esults in each pixel emitting just one of the thr ee primary colours. Figure 7.13 shows the dif ferent filter arrangements that ar e
employed; however, the striped filter is most commonly used for TV display
devices.
Mosaic array
Stripe array
Delta array
Figure 7.13 RGB pixel cell array patterns employed in LCD panel displays. The
stripe array is the most common type
Video display equipment
171
In Chapter 5 we saw that when an analogue video signal is converted into
digital format, it is generally represented by 8-bit binary words, which gives
us 28 256 contrast levels. In an LCD panel, the cells ar e addressed using
digital video signals, which means that the panels ar e (usually) capable of
reproducing 256 grey-scale levels. For a colour display, bearing in mind that
we have three different coloured pixels each capable of r eproducing 256
levels, the maximum number of colours that we can obtain is 256 256 256 16.78 million.
The number of pixels that a panel contains will vary depending on its
size and the cost of the device. For example, an SVGALCD panel would be
required to have a pixel size of 800 600. However, bearing in mind that
we require three LCD cells to make up one coloured pixel, the number of
cells required would be 800 600 3 1.44 million.
The back light is normally a cold fluorescent tube (although some larger
display devices employ tungsten-halogen or metal-halide) that is driven
by a high-frequency supply at around 500–700 Va.c. The drive frequency is
carefully selected to ensure that it does not cause a str obe effect with the
picture rate and is typically in the order of 40–60 kHz. The waveshape must
be a pure sinewave because any harmonic components at such a high voltage could cause interference with the surrounding processing circuits, and
even other equipment close by . Perhaps the most common failur e of an
LCD display is the back light because the tube itself has a limited life, and
the power supply r equired to produce the drive voltage r uns hot and as
such, is more liable to failure.
Two problems that have plagued LCD display devices for many years
have been the lack of contrast ratio and light output. In recent years manufacturers have devised some clever techniques to overcome these problems
and, although each manufacturer has their own methods, the general principle is as follows. To increase the light output range, the display processor
monitors the average light level for each pictur e frame. When it detects an
overall bright picture, the processor increases the brightness of the back light.
Conversely, for a dark picture content, the back light level is decreased. This
technique can also improve the contrast ratio considerably.
The contrast may be further improved by employing a dynamic gamma
control which also looks at the average brightness of each pictur e. When
the average brightness is high, the contr oller increases the amplitude of
the luminance signal and, conversely , for a dark pictur e the luminance
level is reduced. By altering the luminance level we are in effect altering the
contrast, but it is in relation to the average brightness level, so the changes
are not apparent.
Another issue with LCD panels is the poor viewing angle when compared with CRT or plasma displays. For many years the viewing angle
was very poor (and still is on some many less expensive panels); however,
methods have been developed which enable a form of a lens to be fixed
over each pixel, which disperses the light over a wider ar ea. Thus, viewing angles in the order of 140° are now attainable.
172
Closed Circuit Television
Plasma display panels (PDPs)
Plasma display devices have been in use for a number of years and, in the
domestic television market, have been challenging the LCD device. For a
few years the playing field levelled out into one where LCDs cornered the
smaller flat screen market (mainly because it was not possible to manufacture large size LCD displays economically) and plasma dominated the
large flat scr een market. In the early years plasma displays wer e very
expensive and yet suffered from a lack of definition, poor contrast and a
relatively short life (approximately three years).
In more recent years the field has changed; lar ger LCD displays ar e
now available (with much impr oved picture quality) and, thanks to
improved manufacturing techniques and image pr ocessing, plasma displays offer high resolution, improved contrast range and much longer life
expectancy. At the time of writing, it remains to be seen where the market
will go, and whether the LCD ultimately will manage to dominate the market as it currently threatens to do.
The principle of the plasma display can be seen when we consider a
conventional fluorescent tube which, when subjected to a high electrical
charge, glows because the gas inside br eaks down into a plasma. Each
pixel in a plasma display device is filled with a mixture of xenon and neon
gas and has electrodes fixed to it so that, when a voltage is applied across
the electrodes, the gas br eaks down and some of the electrical ener gy is
converted into electromagnetic waves, i.e., light. The light pr oduced is in
the UV spectrum; however, it is converted into visible red, green and blue
by the addition of phosphors inside each cell. The phosphors ar e excited
when hit by UV light and in turn emit visible light energy.
Each pixel is made up from three plasma discharge cells where each cell
contains one type of phosphor. Therefore each pixel contains a red, a green
and a blue light-emitting cell. The basic constr uction of a single pixel is
shown in Figure 7.14, where it can be seen that each cell actually has three
electrodes. We will consider the function of these in a moment. The phosphor linings can be seen in the diagram and, when the gas discharge takes
place, the phosphors emit visible light out thr ough the fr ont viewing
screen. The electrodes that are attached to the front glass are made from a
transparent material so that they are not visible on the screen.
To produce different coloured pixels it is necessary to pr event some of
the cells from firing. For example, if a certain pixel needs to appear yellow,
then only the red and green cells must fire. The data electrodes control the
firing of the cells by having their driver set to either a logic 1 or 0. Only
those cells that have been addr essed as 1 will fir e during the dischar ge
period.
The sustain electrode determines how long the cell will maintain its discharge. The longer the discharge period, the brighter that cell will appear.
The scan electrode works in conjunction with the sustain electrode to produce the dischar ge potential and pr ovide a dischar ge current path.
Video display equipment
Scan
electrodes
Gas filled
cells
Data
electrodes
173
Sustain
electrodes
Red
Green
Front glass
(screen)
Blue
Phosphor coated cells
producing R, G, and B
light
Figure 7.14 Pixal cell construction of a plasma display panel
For each TV frame, the display pr ogresses through a three-step cycle: set,
address and discharge (sometimes referred to as initialize, write and sustain).
Bearing in mind that the cells will still be illuminated fr om the last TV
frame, at the start of the next frame it is necessary to ensur e that all cells
are extinguished and ar e restored to a neutral state. T o accomplish this,
a set pulse is applied between the scan and sustain electrodes, which causes
the gas to produce a small, low-level discharge.
Following this short set period, the addr ess period begins. This is when
the digitized video signal is used to determine whether or not a cell should
fire by applying a logic level to its electrode control circuit. For each cell that
is set to logic 1, a negative pulse is applied to the data electr
ode and a positive pulse to the scan electr ode. This has the effect of charging the gas, in
which state it does not actually emit any light energy. The address period
lasts for approximately 2 ms, during which time all cells will have been
addressed and charged as necessary.
Following the address period comes the dischar ge (or sustain) period,
when all of the cells that have been charged are simultaneously discharged
to create the picture frame. Discharge is achieved by pulsing both the sustain and scan electrodes in such a way as to create a high potential across the
gas, causing it to break down and emit light energy. The brightness of the cell
is determined by the length of time that the sustain pulses are applied, which
in turn is determined by the binary value of the digital video signal corr
esponding to that pixel cell. As with LCD panels, the norm is to use 8-bit
174
Closed Circuit Television
video which gives 256 levels of gr ey and 16.78 million colours, although
some monitors may employ 10-bit video, which gives 1024 levels of gr ey
and 1.07 billion colours (in theory!).
Unlike the CRT where the picture frame is scanned over two consecutive fields, a plasma display produces all pixels simultaneously. Therefore
the incoming video signal must be de-interlaced befor e being applied to
the display control circuits, resulting in a progressive scan. Remember that
interlaced scanning was employed to eradicate the pr oblem of pictur e
flicker associated with the CR T. A plasma display does not flicker in this
way, so there is no need to retain the interlace.
Problems occur with higher-resolution display panels because the large
number of cells means that the addr ess period is too long. T o overcome
this, some manufacturers split the scr een into top and bottom halves (in
terms of addressing) so that two r ows of pixels can be addr essed simultaneously, reducing the address period by 50%. The principle is illustrated
in Figure 7.15. The problem with this is that the monitor requires two display driver circuits, adding to the cost.
Display
driver 1
Two rows are addressed
simultaneously, reducing
the address period
Display
driver 2
Figure 7.15 By dividing the panel display into two halves and employing
separate display drivers, two pixel rows can be addressed simultaneously
A problem that may be experienced with PDPdevices is a poor contrast
range, which is a result of their lack of ability to produce a true black level
in the image. This problem is due to the low-level light emissions resulting
from the necessary discharge of the cells (by the action of the scan pulse)
at the end of each frame period. In more recent years some manufacturers
of PDP devices have employed ingenious methods of modifying the scan
pulse action such that the cell discharge is more gentle and therefore does
not produce the same amount of UV energy.
Another problem that is inherent in PDP devices is an effect that is known
as false contours. This ef fect only occurs on moving images, and usually
only in brighter picture areas, where certain areas appear overexposed. The
source of this effect can be traced to the coding used to set the dischar ge
period in the cell drivers, wher e the light level content for adjacent cells
becomes added together by the human eye as it follows a moving image
across the screen. Once again, different manufacturers have found different solutions to this pr oblem, some of which perform better than others,
which is one reason why one PDP may appear to perform better than
another. In general, the issue of false contours is over come by controlling
Video display equipment
175
the discharge pulses in such a way that the eye cannot integrate the information in adjacent cells in bright moving objects.
Whether or not plasma display devices have any future in CCTV applications remains to be seen. Some control rooms have employed them as a
means of providing large screens for general monitoring or multiplex display; however, this was, per haps, before the availability of lar ge-screen
LCD. The future of PDP will be determined by its ability to develop and
improve at the same rate as LCD and ther efore remain competitive, both
in terms of image quality and price.
Projection systems
Video projection is by no means a new art. Back in the 1950s black and white
projection televisions were developed to provide a large-screen alternative to
the 12 television CRT which, at the time, was the lar gest available screen
size. For many decades to come, pr ojection monitors wer e large, heavy,
expensive units that produced low-resolution coloured pictures, and offered
a very narrow viewing angle and a poor contrast ratio. Har dly a contender
for a CCTV contr ol room! Advances in technology have moved pr ojection
monitors forward a long way, and in more recent years a number of models
have become available that do offer a reasonable picture quality when viewed
from a distance (which is how any large screen is meant to be viewed!).
If we compar e any high-quality LCD or PDP display monitor with an
equivalent-size projection monitor, the LCD/PDP models will always come
out best in terms of contrast ratio, resolution and (usually) price. Where projection TV still has the dominance is in the very lar ge screen market. Any
CCTV control room that requires very large displays for monitoring purposes may find a solution in a projection monitor. However, both the engineer and the owner must be aware that the maintenance cost for these units
tends to be higher than for LCD or plasma displays because of the lamp
life. It is wise to check the specified lamp life for one of these units as, when
lamp replacement becomes necessary, the r eplacement cost of the lamp
unit alone will r un into hundr eds of pounds sterling at least, and can
sometimes run into thousands of pounds sterling.
Projection monitors usually employ an LCD panel in the light engine to
produce the red, green and blue pixels. The operating principle is quite similar to that of the LCD display, except that instead of having a r elatively low
output back light as a light source, projection monitors employ a high brightness xenon lamp with a light output in the order of 1000–2000 lumens.
An alternative to using a self-contained projection monitor is to employ
a video pr ojector. Here there are two competing technologies: the LCD
and the digital light processor (DLP) projector, which is designed around a
digital mirror device (DMD) light projection engine that is manufactured by
Texas Instruments.
The basic DLP principle is illustrated in Figure 7.16, where the action of
four mirrors is shown. Each mirror can be rotated through 45° by application
176
Closed Circuit Television
of a drive voltage, and the drive voltage is controlled by a binary address
which determines the brightness of the pixel pr oduced by each mirr or.
With no deflection voltage applied, the mirr ors are in their r est position
and the light is reflected into a light dump area within the projector. When
a voltage is applied to a mirror, the mirror tilts and the light is reflected out
through the lens assembly . In Figur e 7.16a, the four mirr ors are shown
in their r est position and thus ther e is no light output to the scr een. In
Figure 7.16b two mirrors have been rotated, and therefore their light output is projected onto the screen to produce two pixels.
Cooling fan
Light dump
Micro mirror
Light source
Figure 7.16a DMD operating principle. With all micro mirrors set to their default
rest position, all light is reflected onto the light dump
Cooling fan
Light dump
Micro mirrors deflected
through 45 degrees
Light source
Figure 7.16b Deflecting the micro mirrors through 45 degrees results in the light
being reflected out through the front lens and onto the screen
Video display equipment
177
A practical DMD comprises at least 1.3 million tiny aluminium mirr ors
producing the pixel count required for a large-screen projected image. Grey
scale is achieved by oscillating the mirrors during each TV frame period, the
mark-to-space ratio of the oscillation determining the brightness of each
pixel. For example, a 1:1 oscillation rate would pr oduce a mid-grey level,
whereas as 3:1 oscillation rate would produce a dark-grey pixel because 75%
of the light landing onto the mirror would be reflected onto the light dump.
The light engine is the name given to the assembly containing the three
DMD chips and the associated optical assembly centring ar ound a large
prism (Figure 7.17). The function of the prism is to split the white light into
its red, green and blue components and focus the thr ee outputs onto the
corresponding DMDs. The light r eflected from the DMD mirr ors then
passes back through the prism and into the main optical lens assembly or
onto the light dump, depending on mirr or orientation. The principle is
shown in Figure 7.18.
Figure 7.17 The light engine comprises a light prism, three DMD devices,
three DMD driver cards, three thermal electric coolers (TECs), the light dump,
a mechanical shutter and a cooling fan
The main advantage of DLP over LCD projection devices is shown in
Figure 7.19, where it can be seen that the density of the mirr ors is much
higher than that of the LCD cells. This is due to the fact that ther e is no
need for any separator between pixel devices in a DLPengine. This higher
mirror density results in a very efficient light output. DLP projectors usually employ modified mirror drive modulation techniques similar to that
used in plasma displays to pr ovide increased contrast ratio and to counteract unwanted ef fects caused by the human eye integrating the light
from adjacent pixels.
Although video projectors are capable of producing high-resolution, largescale images, this is at some considerable expense. Furthermor e, because
178
Closed Circuit Television
Light dump
White light input
comprising R, G & B
components
Lens assembly
White light
output
Light prism
DMD mirror
plane
Red
Green
Blue
DMD devices
Light input path
Reflected light paths
Figure 7.18 The light prism separates the R, G and B light components and
focuses them onto their corresponding DMD device. The reflected light paths
then pass back through the prism where, depending on the angle of each micro
mirror, they pass on to the main optical lens, or into the light dump
LCD cells
DLP micro mirrors
Figure 7.19 DLP technology enables the mirrors to be packed more densely
than LCD cells, resulting in greater light output efficiency
of the noise and heat that they generate, plus the fact that they must have
a clear, unobstructed view of the screen, video projectors are not a popular choice for normal viewing applications in a CCTV control room.
Termination switching
In Chapter 2, Figure 2.6, we saw that for maximum power transfer to take
place between any two devices, the output impedance of the first device
must equal the input impedance of the second device. When the impedances
Video display equipment
179
are not matched, some signal loss will be evident and, in the case of a signal transmission system, signal reflections may occur. Figure 7.20a shows
a passive circuit arrangement for composite video input/output connection on any item of CCTV equipment. W ith this arrangement, the input
and output sockets ar e effectively connected in parallel acr oss a resistor,
meaning that changes in impedance at the output will alter that at the
input. Therefore the input impedance, which should be 75 in order to
match the co-axial cable, is dependent on whether or not another piece of
equipment is connected to the video output socket.
To monitor
circuits
75 S
Terminator
switch
Video in
Video out
Figure 7.20a Terminator switch arrangement on a monitor loop circuit
In Figure 7.20a, when an item of equipment is connected to the output
socket, correct impedance matching will be maintained by the input circuit of that equipment. When there is nothing connected to the output socket, a switched 75 resistor is made available to maintain correct impedance
matching.
For many years this selector switch, known commonly as thetermination
switch, was manually set by the installation engineer . The corr ect setup
method was to set every switch in a chain of equipment to the ‘out’ or ‘off’
position except for the one on the equipment at the end of the chain. In
more recent years, this switching action has been automated by making it
a part of the output BNC socket. When ther e is nothing connected to the
output socket, the switch is closed, connecting the resistor into circuit. The
action of connecting a BNC plug into the socket opens the switch contacts.
An alternative, and mor e effective, method for automatic termination
switching is shown in Figur e 7.20b, wher e active buf fer amplifiers ar e
employed at both input and output sockets.
180
Closed Circuit Television
To monitor
circuits
Z in 75 Video in
Z out 75 Video out
Figure 7.20b Use of buffer amplifiers to maintain correct impedance matching
Correct termination is important in a CCTV installation. Looking again
at Figure 7.20a it can be seen that if the r esistor is switched into cir cuit
when an item of equipment is connected to the output, the impedance will
fall below 75 . This will cause an attenuation of the signal, resulting in a
poor-contrast picture. If the switch is left open when there is nothing connected to the output, the input impedance will incr ease, causing a highcontrast display.
It is easy to think that a low- or high-contrast picture can be easily corrected by simply adjusting the contrast contr ol on the monitor; however,
there can be more serious consequences from an incorrectly set terminator
switch. In the case where the resistor is switched in when it shouldn’t be,
the reduced signal is passed onto the next piece of equipment in the chain,
and so on. Similarly, when the resistor is omitted when the monitor is the
last item in the chain, the excess signal level may r esult in frame roll due
to clipping of the sync pulses.
More serious consequences may become evident in systems where telemetry is passed down co-axial video cables (see Chapter 10). Remember that
co-axial cables behave as LC tuned circuits, and that standing waves exist
along their length. As we saw in Chapter 2, when the terminating impedance is incorrect, the standing waves alter , and reflections can pass back
down the cable. Such a condition can cause corr uption of the telemetry
data, resulting in intermittent or permanent loss of control of remote cameras, etc.
Occasionally an engineer may encounter a situation where there is neither an input buffer nor a terminator switch. In such cases a 75 terminator should be employed. These devices resemble a sealed BNC connector;
however, they have an internal 75 resistor.
Video display equipment
181
Resolution
This is generally quoted in TVL, the definition of which was given in
Chapter 6 (see Figure 6.15).
Monitor resolution is governed primarily by the quality of the display
device, although the design of the drive cir cuits does have some bearing.
A quick look through any CCTV supply catalogue reveals a range of monitors with resolutions from 300 TVL to over 1000 TVL, although as with cameras, high-resolution monochrome monitors are far less expensive than their
medium-resolution colour counterparts.
When selecting a monitor for a particular application, resolution is one
factor which must be taken into account.Yet it should be remembered that
for CRT monitors, the TVL quoted will not apply to the entire screen area
because a CRT suffers problems of beam de-focusing at the sides and, in
particular, in the corners. In addition, colour tubes often suf fer convergence problems in the corners of the screen. If we add to this the problems
of image distortion around the edges introduced by the camera lens, then
we see that it is highly unlikely that a monitor with a esolution
r
of 400 TVL
will be able to produce such resolution across 100% of its screen area. As a
rule of thumb, if you imagine a r ectangle in the centre of the screen with
an area equal to about one thir d of the screen area, then this is where the
quoted resolution will be achieved. Thus, if it is r equired to have a r esolution of 400 TVL at the sides of the pictur e, then a monitor of ar ound
450 TVL would most probably be required.
Ergonomics
Some CRT monitor designs are derived from domestic TV receiver models
where the tuning and RF signal processing stages have been removed, and
a range of video (and possibly audio) input/output sockets ar e added.
Being designed around a domestic receiver, these monitors frequently offer
high resolution. However, the moulded plastic cabinet does not make
them particularly robust, and it is not possible for them to be stacked. The
metal-cased rectangular design is often more suitable for CCTV purposes
as it is mor e able to withstand the industrial envir onment, and it is possible to stack units into banks.
Careful thought should be given to the positioning of monitors during
the initial site survey. Such things as viewing height, distance and angle,
ventilation, monitor size, display type and r esolution should all be considered at this point. Standar ds exist for the er gonomic design of control
centres (ISO 11064: Ergonomic design of control centres), and these have
considerable bearing on the design and layout of CCTV control rooms.
Generally speaking, monitors should be positioned so that they may be
viewed without causing discomfort to the operator over a period of time.
The distance of the monitor from the operator will depend on its size and
182
Closed Circuit Television
function; however, recommendations are laid down in the new standards
which are outlined in Table 7.1, along with the figures quoted in the BSIA
Code Of Practice for CCTV (BSIA No. 109: Issue 2: October 1991). In addition to the distance, the viewing angle must not be excessive, and 30°maximum is recommended for monitors located away from the operator, and
15° maximum for a monitor mounted on a desk directly in front of the operator. Also, site the monitor so that the display is not impaired by glare from
light sources, otherwise steps must be taken to remove or mask the source
of the glare. Alternatively, an anti-glare screen can be fixed to the monitor,
although this can reduce the light output.
Table 7.1 Note that some of the BSIA figures relate to slightly different monitor
sizes. Where this is the case, the monitor size is shown in parentheses
Monitor size
Viewing distance (ISO 11064)
Viewing distance (BSIA COP)
23 cm (9)
30 cm (12)
43 cm (17)
53 cm (21)
0.9–1.2 m
1.2–3.0 m
1.3–3.6 m
1.3–4.6 m
1–2 m
2–3 m (39 cm/14)
–
3–5 m (51 cm/20)
One point to consider when stacking or racking monitors is ventilation.
The reliability of a monitor may be reduced if it is made to operate continually at a high temperature, and in a large tightly packed stack the units in
the centre could easily become overheated. Of course it is not always up to
the installing or servicing engineer how the monitors ar e to be mounted,
but where a monitor is found to be over heating, the hazar ds associated
with this (unreliability, fire, etc.) should be pointed out to the customer.
The required monitor size is dependent on such factors as the amount
of detail which is to be discerned and the number of images that are to be
displayed. For example, it would not be practical to display sixteen
images on a 23 cm monitor.
Finally, make sure that it is possible to access the monitor after installation. It will be necessary for operators to routinely clean the screen, as well
as make occasional adjustments and, when a monitor fails, it will be necessary for the service engineer to r emove it and install a temporary or permanent replacement.
8 Video recording equipment
The ability to be able to r ecord CCTV images is of immense importance
because it not only enhances the ef fectiveness of systems which ar e at
times unmanned, but it also pr ovides supporting evidence for contr olroom operators who by natur e of their work become eyewitnesses to the
incidents they are monitoring. On the one hand the video evidence will
only be as good as the system performance will allow; on the other , if the
specification of the r ecording equipment does not match that of the system, or the equipment is incorr ectly set up or is poorly maintained, then
the quality of evidence might well fall far below the live pictur e quality
offered by the system.
It was not until the VHS (video home system) machine arrived on the
scene during the latter part of the 1970s that the security industry had anything like a viable r ecording system to complement CCTV installations.
Prior to that, the only r ecording medium available to the CCTV industry
was reel-to-reel machines which only r ecorded in black and white, and
were not particularly successful owing to their high cost, short r ecording
time, low resolution and high maintenance requirements. Further modifications to VHS r esulted in time-lapse r ecording and an enhanced VHS
format – Super VHS – which pr ovided the industry with equipment
that offered acceptable recording duration and pictur e resolution. However, the issues of relatively high maintenance requirements and tape management remained, and, as CCTV systems incr eased in size, these issues
became increasingly apparent.
Digital video recording has been available to the CCTV industry since
the early 1990s; however, for many years the large file size of the recorded
TV frames coupled with limitations in disk capacity and processor power
meant that the r ecording time was very much limited and the frame
update rate was poor.
Consequently these machines still relied heavily on DAT (digital audio
tape) recorders to archive footage and, as DAT was only a variant of VCR
technology (it actually evolved fr om Sony’s early Betamax technology),
many installers and users decided to stay with S-VHS.
A compromise
between analogue tape and hard disk recorders was found in the various
digital tape formats that became available in the latter part of the 1990s.
Two typical examples of these wer e the digital time-lapse (D-TL) format
developed jointly by Panasonic and Sanyo, and Sony’s digital video (DV)
format. However, despite the outstanding image quality and the fact that
the quality did not deteriorate over time, ther e was still the issue of tape
management to address.
184
Closed Circuit Television
By around the year 2000, computer technology had pr ogressed to a
point where more viable disk-based digital video r ecorders were available, and since that time ther e has been a rapid shift fr om analogue and
digital tape to disk-based DVRs (digital video r ecorders) and NVRs (network video recorders).
In this chapter we shall examine the principles of video r
ecording,
concentrating on disk-based r ecording. However, because at the time of
writing VHS is still not completely dead in the industry , at the end of
the chapter we shall overview time-lapse VHS recorders.
Digital video recorders (DVRs)
A DVR is basically a PC that has been designed to perform one specific
function, which is to take both analogue and digital video input signals,
digitize the analogue signals, compress all signals as necessary, and record
these signals onto a har d disk. The DVR enclosur e may be of a custom
design, or it may be a standar d PC tower, but either way it still utilizes
state of the art PC technology including a high-speed (often dual-cor e)
processor, a lar ge-capacity hard disk, a lar ge amount of RAM, and an
operating system such as Microsoft Windows or Linux. Because all video
images in the DVR ar e in the digital r ealm, most machines incorporate a
multiplexer and therefore, in reality, they are two units in one. Multiplexing
principles will be discussed in detail in Chapter 9.
It is interesting to track the evolution of the DVR fr om its early beginnings in the early 1990s when, because of the limited PC technology of the
time, it was very limited in terms of recording capacity, frame refresh rate
and picture quality. As processor speeds and memory capacity continued
to increase, hard disk size and reliability improved, and compression techniques evolved, then a steady impr ovement in DVR performance was
observed. And there is no doubt that this trend will continue for the foreseeable future as PC technology continues to br eak new boundaries, and
compression techniques continue to improve.
It is common to find manufacturers equating digital image quality with
analogue recording performance – for example, stating that a file size of
20 kB will produce an equivalent picture performance to that of a S-VHS
machine. Whilst it is understandable that they ar e attempting to use a
familiar image quality to help the installer and end user appr eciate the
sort of quality they should expect for a given file size, in truth it is difficult
to accurately equate analogue and digital images in this way , especially
with the compression method used in MPEG. This is because the losses in
digital recording appear somewhat dif ferent on the scr een from those of
analogue recordings.
Where MPEG video compression is employed, as the amount of compression is increased, the image begins to break up into pixel blocks, a feature that will never appear on an analogue r ecorder. Yet these blocks are
Video recording equipment
185
often only evident in areas of very fine picture detail which would never
have been recorded by an analogue machine anyway. Thus a comparison
of the performance of the two formats may r eveal that, although neither
format is capable of reproducing a particular part of an image, the loss is
manifest in quite different ways. In most cases a comparison of the ar eas
of picture that can be r esolved reveals that the digital images ar e cleaner
and have a gr eater contrast ratio. On the other hand, digital images ar e
proving more difficult to enhance using current techniques than their analogue counterparts, which is not good news for anyone involved with
forensic analysis of video information.
So what makes one DVR better than another? W ell, processor speed
and type, RAM capacity and hard disk capacity all make a difference. The
type of operating system can also affect the speed of the machine. But perhaps the most important consideration is the quality of the hard disk that
is to be used to r ecord the digital video information. One of the biggest
problems with DVRs is hard disk failure because the disk is working flat
out, 24/7. The lar ger manufacturers making custom DVRs will only use
high-quality, server-class, hard disk drives that are designed for continuous operation. However, be careful when choosing to use DVRs that ar e
based on standard PCs which have been adapted by the installation of a
video capture card and suitable application software. Such machines must
incorporate a high-quality hard disk drive if they ar e to function reliably
for any length of time.
The user interface for a DVR varies considerably . A number of manufacturers have chosen to mimic the buttons on a VHS machine, which
proves to be a very friendly interface for users who are not at all technical
and may str uggle to cope using a mouse and softwar
e application,
whereas everyone knows how to use the basic functions on a VCR.
Where the DVR is based on an adapted PC, the user interface tends to
be via the PC display . Some of these ar e more friendly than others, and
should be tested befor e a final decision is made to employ a particular
machine.
DVR principle
Figure 8.1 illustrates the operating principle of a digital video r ecorder.
For the machine to be able to perform ef fectively, the cor e PC section
should incorporate a high-speed dual-core processor, a large RAM capacity and a dedicated hard disk drive onto which will be installed the operating system, the DVR (and MUX) application softwar e, drivers, etc. T o
increase speed and reliability, a separate hard drive (or drives) should be
employed for the actual video storage. It is dif ficult in a textbook such as
this to specify actual processor speeds and RAM capacities because technology in this area is progressing so rapidly that the text soon becomes out
of date. As a guide, remember that video signal pr ocessing places a high
186
Closed Circuit Television
MUX
D/A
converter
Analogue
video
output(s)
Analogue
inputs
Analogue
input
capture
A/D
converter
Digital
compression
de-compression
Internal
hard
disk
Video storage
TCP/IP
NIC
Optical
disk
recorder
Control to
DVR section
Processor
BIOS
RAM
PC core
System
hard
disk
Operating system
DVR application
software
drivers
Figure 8.1 Typical DVR architecture, based on PC hardware and operating
system
demand on a CPU, and you cannot have too much RAM capacity available for buf fering purposes during compr ession/de-compression of the
video images. So, for the next few years at least, it will be safe to say that
the fastest processor available will be the pr eferred choice for any DVR,
and RAM capacities should be measured in Gbytes and not Mbytes.
The analogue input capture card shown in Figure 8.1 plays a key r ole
and may determine the performance of the DVR in terms of picture update
rate and image quality. The capture card will have a number of analogue
inputs ranging (typically) fr om 4 to 16. The speed and ef ficiency of the
capture engines will determine the rate at which the functions of
A/D
conversion and digital compression are performed, and therefore the rate
at which pictures are recorded.
In addition to the analogue inputs, the DVR will have a network connection (network interface car d, or NIC) via which the machine may
stream images from IP cameras and/or video servers that will have converted analogue camera signals into TCP/IP. The network interface is also
used to provide remote access to the DVR fr om other PCs running either
browser or administration software. Such access can provide remote configuration or reconfiguration, browsing of recorded images from a remote
site, and control of PTZ cameras.
Video recording equipment
187
The hard disk used for video storage may be a single internal drive;
however, this will have limited r ecording capacity, and offers no protection against loss of all video information in the event of catastrophic drive
failure. Many machines of fer the facility of r emovable, replaceable disk
drives, and the larger machines are generally able to support RAID arrays
(discussed later in this chapter).
An analogue video output is usually available, often in the form of both
composite and S-Video. This output is normally used to pr ovide a monitor facility, although it is sometimes used to download video onto a VCR
or other external recording device. This output is normally pr ovided via
an internal MUX.
The optical disk recorder shown in Figure 8.1 is not necessarily a standard feature, but has been included to r epresent the fact that the machine
must provide some means of extracting video information for the purposes
of evidence or possibly ar chiving. Programmable options are often available to select automatic archive routines. For example, following an external alarm input activation, the unit automatically copies to DVD the data
relating to the activation, beginning from a few minutes prior to the alarm.
The exact form that the archive device takes varies between machines,
from a standard DVD recorder using MPEG-2, to some form of data disk
recorder. However, it should be noted that in many cases the disks, including DVD recordings, cannot be replayed on the standard equipment incorporated in a PC without the special viewing softwar e. This softwar e is
provided by the manufacturer and is often free, but it may not be immediately available in every police station or courtr
oom, which is wher e
evidential material needs to be viewed.
Effects of compression
Without video compression, DVRs would not be viable because uncompressed image file sizes are very large, and even hard disk drives having
capacities measured in terabytes would only provide very limited recording times. The principles and effects of video compression are discussed in
detail in Chapter 5, but in this chapter , as we consider the DVR, we can
appreciate the practical implications of compression in CCTV.
On the one hand there is a desire to compress the video signals as much
as possible in order to extract the maximum recording time from the DVR,
and yet on the other hand we know that excessive compr ession results in
unrecoverable losses in pictur e information, leading to poor -quality picture reproduction. The ideal situation is one wher e we apply suf ficient
levels of compression to deliver a high recording capacity, but are able to
extract high-resolution images fr om the machine. Unfortunately it is
sometimes difficult to reach this position without spending a lot of money
on multiple DVRs, especially in systems that have a lar
ge number of
cameras and require a long archive period.
188
Closed Circuit Television
For many years the two principal compr ession methods employed in
CCTV DVRs have been MPEG-2 and Wavelet (now JPEG 2000); however,
since the implementation of MPEG-4, many machines ar e turning to this
compression format as it generally of fers improved image quality over
MPEG-2.
Recording capacity
In general, hard disk video r ecorders are set up such that they will continue recording until the disk is full and will then pr oceed to overwrite
the earliest recordings, thus providing continuous recording. Of course,
problems will arise when the disk capacity is insuf ficient to provide the
required archive period and thus the installer must ensur
e that the
intended equipment is up to the job. This is mor e difficult for disk-based
recorders than it is for VCR installations because, for VCRs all that is
required is to look at the recording time (i.e., 12 hours) and ensure that the
owner has enough tapes to cover the ar chive period. For example, for an
archive period of 31 days, VCRs operating in the 12-hour time-lapse mode
will require 62 tapes per machine plus, say , 5% to cover for tape extractions for evidential purposes.
In the case of hard disk recorders, the recording time is dependent upon
a number of factors. These are illustrated in Figure 8.2. First of all there is
Disk
capacity?
Number of pictures
(frames) per second
to be recorded?
File size
per picture?
Compression method
and radio?
Picture update
rate?
Number of
cameras?
Figure 8.2 Factors that govern the recording (archive) time for a DVR
the picture recording rate. This is equivalent to the time-lapse period in a
VCR where, on many DVRs, the installer or user can set the number of pictures per second that they wish to record. Secondly, there is the file size for
each picture that is recorded, which is a product of the type and amount of
Video recording equipment
189
compression that is applied. Because the images ar e digitized and compressed prior to recording, each picture in effect becomes like an individual file. The amount of compression can be set by the installer, or possibly
by the user. The more compression that is applied, the smaller will be the
file size and thus the greater will be the recording time. However, increasing the compression will reduce the image quality.
Another factor which determines the recording time is the picture update
rate, which is not to be confused with the pictur
e recording rate. The
recording rate is the number of pictures per second (PPS) that the machine
is recording, whereas the picture update rate is the rate at which each camera image is updated when reviewing the recording.
The number of cameras connected to the DVR af fects the maximum
recording time: the gr eater the number of cameras, the shorter the total
archive period, assuming that the pictur e update rate is maintained.
Finally, the recording time is determined by the capacity of the hard disk.
The update rate (in seconds) can be determined by dividing the
number of cameras by the recording rate. Or:
Pic update (in seconds) n ÷ PPS
where n is number of cameras, and PPS is the recording rate.
For example, for a system having just one camera, if the r ecording rate
is set to 25PPS, then the update rate will be 1 ÷25 40 ms. In other words,
the replayed images for the camera will be updated every 40 ms, which is
equal to the TV frame rate – that is, r eal time 25 pictures per second. If 25
cameras are now connected to the r ecorder and the r ecord rate is maintained at 25 PPS, then the update rate becomes 25 ÷25 1 second. So now
we see that when the recording is replayed, the images from each camera
will only be updated once per second.
Remembering that the PPS is a variable function of the DVR that is set
up by either the installer or the user, the choice of setting for any DVR will
be determined by two factors: the number of cameras connected to the
recorder, and the rate at which the user requires the replayed images to be
updated. Let’s consider a mor e extreme example. A DVR is required to
record the information fr om 30 cameras. The natur e of the security risk
requires a pictur e update of at least twice per second. Thus, fr om the
expression Pic update n ÷ PPS:
PPS n ÷ Pic update
30 ÷ 2 15 PPS
Thus, if the DVR’s PPS setting is made to be 15, then it will meet the performance requirement for that system. However, this calculation does not
take into account the archive period, and if the hard disk capacity is insufficient, the DVR will begin to overwrite video information too soon. Thus,
to ensure that a DVR will pr ovide a required archive period, the installer
must calculate the hard disk capacity. In practice, many DVRs incorporate
an internal calculation utility that will tell you the ar chive time for any
190
Closed Circuit Television
combination of PPS, picture update and compression settings, but in order
to use this you will have to have purchased the machine in the first place!
Being aware of this, many of the leading DVR manufactur ers offer calculation utilities on their websites, and/or through their technical support.
Using a longhand method, the required disk capacity for a system can
be calculated from:
Disk capacity 86 400 days PPS file size
where 86 400 is the number of seconds in one day, ‘days’ is the number of
archive days r equired, PPS is the number of pictur es per second to be
recorded, and file size is the size of each picture file (in kilobytes).
Note that the above expression and the examples given below assume
that the recorder has been set to r ecord the full frame size. Clearly, if the
DVR has been set to r ecord a r educed frame size, then ther e will be a
reduction in the file size for each pictur e, resulting in a corr esponding
decrease in required disk capacity.
For example: a CCTV system requires an archive time of 14 days. It has
been decided that the DVR will operate at a rate of 5 PPS and, for the proposed machine, the file size must be 20kB per picture in order to obtain an
image quality comparable to that of an S-VHS r ecorder. Thus, the DVR
would require a hard disk capacity of at least
86 400 14 5 20 000 121 gigabytes (GB)
This example, which is typical of curr ent digital r ecording equipment,
illustrates just how much disk space is required. And yet the figures used
in our example r eally only relate to a smaller installation because of the
low picture recording rate (PPS). Let’s now consider the larger system that
we looked at earlier. This used 30 cameras and r equired a picture update
rate of twice per second, which led to a required PPS of 15. Let’s now say
that the archive period for this system has to be 31 days. How much hard
disk capacity would be required if the recording quality were to be similar
to that of S-VHS?
86 400 31 15 20 000 803 gigabytes (GB)
If this system wer e expanded to 60 cameras, then the har d disk capacity
would have to be increased to
PPS n ÷ Pic update
60 ÷ 2 30 PPS
Disk capacity 86 400 31 30 20 000
1.6 terabytes (TB)
This capacity could be r educed by reducing the file size for each image;
however, this will result in a reduction in image quality.
One of the prime factors in the rapid shift fr om VHS to DVR in CCTV
recording was the availability of r eliable, high-capacity har d drives.
However, for many years such high capacities wer e only available by
Video recording equipment
191
employing SCSI technology, which enables a number of har d disk drives
to be daisy-chained in such a way that they behave as a single drive. DVR
manufacturers took different approaches to using SCSI: some machines
incorporated a number of internal SCSI drives, whereas others had external connections which enabled the installer to add drives as necessary
. The
problem with SCSI is that, although the drive speed is very fast, there is no
protection against loss of data in the event of any one of the har d disk
drives failing. When a disk drive fails, all data in the entire array is lost just
as with a single drive, except that the chances of a drive failur
e are
increased because there are more drives to fail.
RAID disk recording
An alternative to SCSI is RAID technology (r edundant arrays of independent disks – although the original ‘I’ term was ‘inexpensive’) wher e a
number of hard disk drives are grouped into an array, and the array acts
as a single drive. Although this might sound very much like SCSI, the difference is that, with RAID, it is possible to build in protection against data
loss in cases of a disk failure.
There are different levels of RAID; however, the principle is set out in
RAID 0, which is illustrated in the left-hand array of thr ee disks depicted
in Figure 8.3. Each disk is r eferred to as a column, and each column is
broken down into stripes. The stripe width refers to the amount of data in
each stripe which, in the illustration in Figure 8.3, is 128 kB.
Now, imagine that we are recording a 1 MB file which, taking 20kB as a
typical file size for one compr essed CCTV image, is equal to about 50
images. The first 128 kB will be written to column 1, stripe 1, the next
128 kB will be written to column 2, stripe 1, the thir d 128 kB to column 3,
stripe 1, the fourth to column 1, stripe 2, and so on until all of the file has
been written (which would be on column 2, stripe 3). Because of the parallel action of the disks, the writing speed can far exceed that of SCSI;
however, RAID 0 does not provide any immunity to data loss in the event
of a drive failure. Because the files ar e scattered across the disks, the loss
of a disk means that the system would have no way of r ecovering that
information. Therefore all of the data is in effect lost.
RAID 1 is, on its own, not r eally a RAID system at all; it is simply a
specification for arranging backup disk drives to mirr or the main drives.
In RAID 1, as data is being written to each main drive, it is also being written to a separate mirr or (backup) drive. In the event of a main drive failure, the mirror simply takes over and continues to run with the other main
drives. When the defective drive is r eplaced, the data on the mirr or is
copied over to the new drive, which then continues to work in the main
array.
The robustness of RAID only comes into play when we combine RAID 0
and RAID 1 to cr eate a RAID 1 0 array. This is depicted in Figur e 8.3
when we combine the thr ee-disk RAID 0 array with the thr ee mirror
192
Closed Circuit Television
Stripe 3
Stripe 2
Stripe 1
Stripe 3
Stripe 2
Stripe 1
Stripe 3
Stripe 2
Stripe 1
Disk 1
(Column 1)
Disk 4
(Column 1)
128 kB
128 kB
128 kB
128 kB
128 kB
128 kB
Disk 2
(Column 2)
Disk 5
(Column 2)
128 kB
128 kB
128 kB
128 kB
128 kB
128 kB
Disk 3
(Column 3)
Disk 6
(Column 3)
128 kB
128 kB
128 kB
128 kB
128 kB
128 kB
RAID 0 array
Raid 1 mirror array
Figure 8.3 Principle of RAID 10. The RAID 0 array (disks 1–3) has no
protection against drive failure. When used in conjunction with the RAID 1 mirror
array (disks 4–6), we have a RAID 10, i.e., a striped mirrored array
drives. This arrangement operates very fast and offers high data reliability
in the event of a drive failure. However, the cost and physical size is considerable when compared with SCSI because of the need for the mirr or
drives.
A less expensive alternative to RAID 10 is RAID 5, which is illustrated
in Figure 8.4. Again we shall take the example of a 1 MB file being written
into a series of 128 kB stripes. The first 128 kB block is written to columns
(disks) 1 and 2. The second block is written to columns 2 and 3, the thir d
to columns 3 and 1, and so on. In addition to each block of data, a parity
Video recording equipment
193
Column 2
Column 1
Data 8 parity
Data 7 parity
Data 5 parity
Data 4 parity
Data 2 parity
Data 1 parity
Data 7 parity
Data 6 parity
Data 4 parity
Data 3 parity
Data 1 parity
Column 3
Data 8 parity
Data 6 parity
Data 5 parity
Data 3 parity
Data 2 parity
128 kB
128 kB
128 kB
128 kB
128 kB
128 kB
128 kB
128 kB
Data 1
Data 2
Data 3
Data 4
Data 5
Data 6
Data 7
Data 8
A 1 MB data file, broken down into 128 kB blocks
Figure 8.4 A RAID 5 array requires a minimum of three disk drives to function.
This illustration shows how RAID 5 distributes a 1 MB file across three columns,
when the stripe width is set at 128 kB
check value is r ecorded, although the data block and the corr esponding
parity are not recorded onto the same disk.
Now, let’s assume that disk number two (column 2) has failed. No data
is actually lost because it has all been duplicated on the other two disks.
The system will continue to record data and parity (although now the parity will have to co-exist with the data) until a r eplacement drive is fitted.
The system will then reconstruct the stripes on the new drive.
The one disadvantage of RAID 5 when compar ed with SCSI is that
RAID 5 is slower due to the amount of duplicity in writing, plus the need
to write the parity information. However , it is the pr eferred choice of
many DVR manufactur ers because its speed is suf ficient for r ecording
CCTV images, especially as disk drive speeds have incr eased in mor e
recent years.
Many DVRs are sold as ‘RAID r eady’, which usually means that they
come with a single internal hard disk for use with smaller CCTV applications, but have a series of RAID 5 caddies into which the installer may add
two or mor e drives to build up an array (bear in mind that RAID 5
requires a minimum of three drives to function).
Where a RAID 5 drive has failed, the DVR will put out some form of
alarm to inform the user that there is a problem. It will also tell the engineer
194
Closed Circuit Television
which drive has failed. However, the engineer must be very careful when
replacing the drive that they pull out the correct one. Bearing in mind that
the system is limping along with only two other drives (assuming a threedrive array) and is not able to construct a true RAID 5 block parity pattern, removal of the wrong drive in effect means that a second drive has
failed and at that point it is almost certain that all data will be lost. This
action by the engineer is unlikely to amuse the user!
Digital video information extraction
DVRs have evolved to a point where they can easily match, if not exceed,
S-VHS in terms of image quality and r esolution (when comparing full
frame size recording). However, one important feature to consider when
selecting a DVR is its ability to produce a removable copy of the recorded
information that is equal in quality to the original, is in a format that can
easily be viewed, and contains a secur e watermark that can be used to
prove that the material has not been tampered with.
With tapes this was simple, because the original tape could be taken from
the source machine, sealed in an evidence bag if necessary and could easily be replayed on any other S-VHS machine via a suitable de-multiplexer.
The simplest way to r eplicate this process is to employ a DVR that has a
removable, replaceable hard disk. However, there are a number of issues
relating to this. First of all, the cost of a r eplacement hard drive is far
greater than that of an S-VHS tape. Then, once r emoved, the drive will
only function again when connected to a similar model DVR. Finally
, what
if the original material is spread across a nine-drive RAID 5 array?All nine
drives would have to be removed and replaced.
In reality, in the UK, where a DVR contains (or is even suspected of containing) evidence relating to a serious crime, the police have the option to
simply remove the entire DVR, possibly for many months. The pr oblem
with this is that in most cases the user has lost not only their r ecording
equipment, but their entire MUX as well, which renders their system useless. An example of this, per haps extreme but r eal nonetheless, is the
action that police in London took in the immediate aftermath of the terrorist bombs on 7 July 2005. Many thousands of VHS tapes wer e removed
from machines; however, these systems were soon up and running again
because all that was required was a replacement tape. Unfortunately, systems that employed DVRs proved to be much more of a problem because
many (not all, it must be stressed) did not have an acceptable archive facility, so the har d disk drive (and in some cases the entir
e DVR) was
removed. Larger systems that employed a multiple of DVRs all using
RAID disks suffered another problem: although archive was possible, it
would have taken many hours for this to take place, during which time
the machines would continue overwriting the oldest files. This was not
acceptable to the police authority whose job it is to secur e all evidence as
Video recording equipment
195
soon as possible. So, once again in some cases all DVRs had to be immediately powered down and removed from site.
The above example is of an extreme nature because we are dealing with a
very serious crime, at a level far gr eater than the average CCTV user is
attempting to deal with when having their CCTV system specified.
Consequently it need not be a requirement of every DVR to have some form
of hard disk replication or mirror facility, and many machines do of fer an
effective archive utility. What the installer/specifier must do is ensur e that
the facility will meet the expected requirements of the system.
One common archive method is for the DVR to have an in-built DVD
recorder. Video information that is deemed to be important may be copied
to a DVD disk (usually in data not MPEG-2 video format) and r emoved.
Many DVRs have the ability to automatically ar chive video to the disk
whenever an alarm activation occurs. The problem with some of these disks
is that they can only be played on a PC that is r unning suitable application
software containing the codec, which is not always convenient. T o avoid
this problem, many machines that utilize DVD archiving make use of standard browser or media player software applications, or include the reproduction software on the disk.
DVD archiving is often acceptable for smaller systems; however , what
options are there for larger recording systems? Where there is a r equirement to store large amounts of data, per haps in the gigabyte ar ena, perhaps the most obvious choice is to use one of a number of data tape
backup systems which are capable of storing many hundreds of gigabytes
of data on one single tape. However, where there is a chance that the information stored on those tapes may be r equired for evidence in a court of
law, the tapes must be logged and stor ed according to control room standards, similarly to analogue videotapes.
Another option for lar ge systems is for the user not to employ any
recording equipment at all and, instead, have all video str eamed from
their control room to a thir d party pr ovider. This thir d party will have
their own secure recording facility amounting to many terabytes of capacity and, for increased security, all video data will be recorded in tandem
in two separate locations. Using secure passwords and private fibre-optic
connections, the user may still access their video information just as if it
were stored locally. The advantages of this system ar e that the user is not
troubled with maintaining their recording equipment and keeping it up to
date, they still have access to all of their video content, and when an event
occurs on a scale of a terr orist attack, there would be ample time for all
data to be ar chived because the main servers could be pr evented from
overwriting, but could continue recording using their spare capacity.
At the time of writing this option is being consider ed by a number of
larger CCTV system operators in the UK, and at least one large city centre
CCTV system has taken this up using BT (British T elecom) services for
both signal transmission and video storage. T ime will tell how this
progresses.
196
Closed Circuit Television
VHS recording
In order to record high frequencies on a magnetic tape, a very high tape
transport speed is r equired. In all analogue video r ecorders this is
achieved by wrapping the tape ar ound a head dr um and r otating the
drum at a rate of 25 revolutions per second (i.e., a rate that is equal to one
TV frame period). The head drum is tilted with respect to the tape so that
the recording heads scan the tape following a diagonal path up the tape.
Two head cores are fixed onto the drum 180° apart so that there is always
at least one head in contact with the tape at any time. The principle is
shown in Figure 8.5.
Path of head during
180 degree rotation of drum
Video head drum
Video head
A
B
A
B A B
Tape
A B
Figure 8.5 Principle of helical scanning. Magnetic information is laid down on
the tape in the form of diagonal stripes. This is known as helical scanning
From this illustration it can be seen that:
●
●
●
head A records the first TV field;
head B records the second TV field;
each single 360° revolution of the head drum records one TV frame.
During replay the machine requires a reference signal to synchronize the
drum rotation with the tape transport. This is what is meant when we
refer to tracking. The r eference signal is pr ovided from a contr ol (CTL)
track which is derived from the 50 Hz field sync, having first been divided
down to 25 Hz before being recorded along the bottom edge of the tape
(for NTSC, the 60 Hz field sync pr oduces a 30 Hz CTL track). The CTL
track is also used as a reference for the real time tape counter, and to mark
the positions of alarm input activations in time-lapse VCRs to aid fast
searching for the alarm events.
Video recording equipment
197
A videotape, therefore, contains the diagonal video signal tracks positioned in its centre, and lateral audio and CTLtracks along the edges. This
track format is shown in Figure 8.6.
Stereo audio tracks
A
B
A
Video tracks
B
A B
A B
Control (CTL) track
Figure 8.6 Videotape track format
Mistracking occurs in a VCR when, during replay, the tape transport and
drum rotation motors are both running at the correct speed, but they are not
in phase with each other. In this case the heads will be scanning two tracks
at the same time, resulting in a low output signal from the playback heads.
Mistracking is a result of small mechanical tolerances between machines
and, to compensate for this, manufactur ers include a tracking contr ol circuit. Tracking problems can also occur when a tape has become str etched.
This alters the angle of the video tracks, making it difficult for the heads to
follow them. Similarly the machine may suf fer mechanical wear, making
tracking difficult. Where tracking err ors are caused by faulty tapes or
machines it is often difficult to correct using the tracking control.
VHS is only capable of r ecording luma signals up to 2.8 MHz, which
corresponds to a r esolution of ar ound 240 TVL. An S-VHS machine is
capable of r eproducing video signal fr equencies up to 5 MHz, which
equates to a TVL in the order of 400.
It is possible to r ecord in standard VHS on a S-VHS tape, but it is not
possible to record in Super format using a standar d VHS tape because of
the limitations in the tape oxide. To prevent this from happening by accident, manufacturers fit a sensor in the machine which detects a S-VHS
tape. If a standard format cassette is inserted into the machine, then even
if the record selection is set for Super format, the machine will r evert to
standard format r ecording. Engineers should make customers awar e of
this, pointing out that they are compromising the quality of the system if
they use the much less expensive standard cassettes.
Time-lapse recording
By making use of the still frame function, in the ercording mode time-lapse
machines record a single field, pause momentarily, and then record another
field. Whilst the machine is in the pause mode the head drum continues to
198
Closed Circuit Television
rotate; however, the heads ar e switched of f. Note that because the tape
stops regularly, it is not possible to r ecord an audio track whilst the
machine is operating in time-lapse mode.
By operating in time-lapse mode a thr ee-hour cassette can be made to
last much longer. Of course the tape only contains a series of still images,
but this is often sufficient for CCTV purposes, and if the tape is r eplayed
in a faster mode than that in which it was r ecorded, a form of fast motion
film can be viewed. The number of still pictur es available on one thr eehour cassette is quite impr essive. In the thr ee-hour mode a machine is
recording 25 pictures per second, and thus over thr ee hours this amounts
to a total of 25 60 60 3 270 000 pictures. In time-lapse mode each
of these frames is an individual picture, and thus the cassette may be considered to be like a photograph album containing 270 000 photos.
The method of recording one TV frame per time-lapse interval is in fact
very wasteful because, when a VCR r eplays a still frame it actually only
scans one field of the frame, and the other field is effectively ignored. This
point is illustrated in Figure 8.7. Therefore it was decided that, if the second field of each frame is ignor
ed, then why bother r ecording it?
Consequently, time-lapse VHS machines actually only record one TV field
out of each frame, meaning that we can now r
ecord twice as many
pictures, i.e., 2 270 000 540 000 still fields/pictures.
Path of video heads
A
B
A
B
A
B
A
B
Stationary tape
Figure 8.7 In still mode the tape is stationary, and thus both video heads scan
the same track
An important consideration with time-lapse operation is the loss of vertical picture resolution. In Chapter 5 we saw that a TV picture contains at
the very best 625 lines less 40 vertical flyback lines, equalling 585 active
lines. However, when we consider that these lines ar e contained within
two fields, clearly when a VCR is in still mode, although it is still producing two fields of 312.5 lines, they are in fact the same lines. In other words,
the vertical resolution is reduced by 50%. This loss may be crucial when a
recorded picture is to be used as evidence.
Not all machines have the same number of time-lapse modes, although
the more elaborate models offer a very wide range. Some typical modes,
along with the delay periods for both frame and field operation, are given
in Table 8.1.
Video recording equipment
199
Table 8.1 Time-lapse modes. Note that the actual recoding period
is always three hours longer than the stated period. This is due to the way
that VHS time-lapse recorders have to arrange their video tracks
Mode
Running time
Delay
12 h
24 h
48 h
72 h
120 h
240 h
480 h
720 h
960 h
15 h
27 h
51 h
75 h
123 h
243 h
483 h
723 h
963 h
0.10 s
0.18 s
0.34 s
0.50 s
0.82 s
1.62 s
3.22 s
4.82 s
6.42 s
Linear slow speed is a form of ‘continuous’ time-lapse wher e 12 or 24
hour recording is achieved not by pausing the tape but by running it continuously at a speed that is four or eight times slower than normal. This
was introduced solely to enable audio recording. Because the tape is running continuously, a lateral sound track can be laid down at the top of the
tape, albeit of very poor quality owing to the slow tape speed.
When alarm mode is selected the machine operates in time-lapse mode,
but when the alarm input terminal on the machine is active, the machine
reverts to the three-hour mode for a preset period (usually a few minutes),
thus ensuring that maximum information is r ecorded. The alarm input
may be triggered manually by an operator using a simple push switch, or
it could be connected to some form of electronic alarm system. On a larger
system where the machine is expected to record the output from a number
of cameras, the alarm activation signal can also be connected to the multiplexer or switcher. The normal recording pattern can then be interr upted
so that only the cameras in the activation ar ea are connected to the VCR
input during the alarm period.
Auto re-record is a featur e which is particularly useful for unmanned
systems. As the name implies, when the tape comes to the end, the
machine rewinds and re-records. Options are usually built into this feature to prevent erasure of important evidence. For example, the machine
may be set so that it will not auto er -record when an alarm input activation
has taken place.
VCR maintenance
Perhaps the biggest criticism of VHS when employed in the CCTV industry is the need for mechanical maintenance. And this criticism is justified
200
Closed Circuit Television
because mechanical parts in a VCR deck do wear out with constant use.
Numerous CCTV engineers can r ecite stories of customers who have
invested thousands of pounds in CCTV equipment, only to penny pinch
on tape replacement, and it is not uncommon for a customer to complain
to an installer that, following an incident, the tape has been replayed only
to produce a snowstorm. Upon inspection it is often found that the tape in
question is the one which was provided during the system handover a number of years previously! In such circumstances, not only will the tape be next
to useless (the oxide coating will be thin, and the backing material stretched)
but the machine will desperately require a major mechanical overhaul before
it is capable of reproducing anything like an acceptable picture.
In practice, a machine that has been operating continually for twelve
months will be in need of a service. Even if the picture quality still appears
to be satisfactory, there is no telling how much longer this will be the case.
When considering VCR servicing, video head dr um checking and
replacement is often singled out as the primary concern, yet this is not the
only item that may be showing signs of wear after a machine that has been
operated continuously for one full year (8760 hours). Items such as the
pinch roller, rubber belts, back tension band (if used) and head cleaning
rollers will all r equire checking and/or r eplacement, and the condition of
the lower video drum, tape guide posts and audio/control head should be
verified. Furthermore, the tape path must be cleaned of all dirt and tape
oxide deposits, and the alignment of tape guides checked and adjusted as
necessary following video head dr um replacement. In some cases, manufacturers provide service kits containing all of the mechanical components
which should be r eplaced after a period of no mor e than 10 000 hours.
Clearly the level of servicing being discussed here can only be performed by
competent VCR servicing personnel, and for this reason it is expected that a
machine will be removed from site and returned for (annual) servicing.
Video head cleaning
The frequency at which video heads r equire cleaning can be a matter of
opinion, and in reality is dependent upon a number of factors including the
quality of tapes being used, the number of times a tape is to be used, the
tape storage method and conditions, whether the machine has an automatic
cleaning facility, and a degree of luck! But before we consider the frequency
and methods of head cleaning it is important to understand what a ‘dirty
head’ is, and recognize the symptoms.
The ‘dirt’ we r efer to is not the airborne dust that we see collecting on
monitor screens and cabinets. It is particles of tape oxide which have
detached from the tape and have either adhered to the side of the video head
drum or become lodged in the head tip. This is illustrated in Figur
e 8.8.
Oxide on the side of the drum reduces the tape/head contact, which in turn
reduces the record/replay signal levels. The overall ef fect is poor tracking
Video recording equipment
201
symptoms. When the oxide is lodged in one of the heads, the ef fect on the
picture is the same as if one video head had failed.
Head tip
All tape oxide deposits on
drum must be cleaned off
Figure 8.8 Typical oxide build-up on a VHS video head drum
Head cleaning is the r emoval of the oxide fr om both the head gap and
drum surface. However, this is not as simple as it might first appear, because
the video head tips ar e extremely fragile, and the slightest pr essure in the
vertical direction will break them off, necessitating replacement of the entire
drum assembly. Oxide on the dr um surface can become well-adher ed, and
there is a temptation to try to scrape this of f. However, the smallest scratch
on the drum surface will alter the aer odynamics to such an extent that the
tape will ride off the drum, producing a permanent line across the picture.
Head cleaning cassettes are available, and some manufacturers supply
one of these with each time-lapse machine. During handover the customer
should be instructed in the use of these, but should be warned that excessive use may lead to early head wear , as some of these can be somewhat
abrasive. As a general rule, if the picture is perfectly all right, don’t bother
cleaning the heads!
Some machines employ an automatic head-cleaning facility . This usually operates on the principle of a felt wheel that is pressed against the head
drum each time the tape loads and unloads. Where this is the case, the customer can be instr ucted to load and unload a tape a number of times to
force this cleaning operation in the event of apparent dirty head symptoms.
The problem with both cleaning cassettes and automated systems is
that, when the oxide is lodged into the gap or fixed firmly onto the drum,
there is insufficient pressure to dislodge it. In this case the only option is to
clean the head dr um manually, and because of the delicate natur e of the
head drum, this operation should be performed only by those who have
been trained in the technique.
Tape management and care
Although this is primarily the responsibility of the operator, the engineer
should make their customers awar e of the implications r elating to tapes
during the initial system handover.
202
Closed Circuit Television
A tape passing through a machine operating in the time-lapse mode is
subjected to far more stress and wear than when it is used at normal speed
because each part of the tape spends much mor e time stationary against
the rotating heads. Thus it is necessary to use tapes that have a str ong
backing material which will not stretch, and a well-bonded oxide coating.
Not all VHS cassettes ar e of the same quality in this r espect, and this is
why only tapes marked as being of ‘pr ofessional’ quality should be used
for CCTV applications.
Another reason for using quality tapes is that not all oxide coatings have
the same magnetic qualities, the budget versions often suffering from highfrequency losses and lower signal-output levels (poorer S/N ratio).
The number of times that a tape should be used is a matter of some contention, but experience in the industry indicates that for tapes being used
in the 24-hour time-lapse mode, twelve passes through the machine is the
maximum. Thus, if a customer is going to operate the machine in the
24-hour mode, purchasing 31 tapes will allow a suitable r otation over a
twelve-month period. However , this does not allow for any losses
incurred when tapes are subsequently removed from the system because
they contain evidence.
The Data Protection Act 1998 requires that a detailed log of all CCTV
tape movement is maintained, and the engineer should encourage the customer to use an auditable tape logging and management system. For
smaller installations it is often possible to make use of systems that are provided with some of the videocassettes which are manufactured specifically
for CCTV use. However, where a manned control room is in operation, a
more sophisticated tape management system will be r equired. These may
be purchased ready made, and include a secur e storage cabinet which is
designed to aid tape rotation, and a complete set of logging sheets.
With regard to tape storage, advise the system owner to store tapes flat
and not vertical, to have a storage envir onment that is not too hot or too
cold (tapes are designed to live in ‘normal’ r oom temperatures, just like
ourselves!), and to ensure that that there are no stray magnetic fields in the
area which will cause at least some degree of erasure: video monitors leak
magnetic fields from all sides, so keep tapes well away.
Digital video tape
In addition to the poor r esolution offered by analogue video r ecorders,
another serious problem for the CCTV industry is the fact that the signal
degrades every time the tape is r eplayed or a copy is made. To overcome
all of these pr oblems a number of digital video tape r ecording formats
have been developed and, although digital video r ecording is by no
means faultless (it suf fers from digital noise and compr ession losses), it
provides a clean, quality picture which does not degrade with either time
or through generations of duplication.
Video recording equipment
203
One such format is the digital time-lapse VCR (D-TL) developed by
Panasonic and Sanyo. This digital recorder employs a normal S-VHS tape
cassette and in many respects functions as a traditional analogue machine.
However, the digital signal processing means that a horizontal resolution
of up to 520 TVL is possible. The D-TL combines the advantages of digital
video recording with the relatively low cost of S-VHS tapes. Per haps one
disadvantage is the fact that the recordings can only be replayed on D-TL
format machines meaning that, when data is r equired for evidential purposes, it will more than likely have to be copied down to S-VHS or even
VHS. Nevertheless, because the original material is of a high quality in
terms of r esolution, this does mean that the copies will be of a high
standard.
Another digital video tape recording format, known as ‘DV’, has been
developed by Sony with both industrial and domestic applications in
mind, and its resolution of up to 500 TVL makes it attractive to the CCTV
industry.
There are two versions of the DV cassette: DV and Mini DV . Both cassettes use a 6.35 mm wide (1/4) tape, but the Mini DV cassette is much
smaller, with a correspondingly shorter recording time. The Mini DV tape
is used in the majority of domestic digital camcor ders because it keeps
the size of the machine to a minimum; however , for CCTV applications
the 60-minute recording time is somewhat r estrictive, and many ar chive
machines take the full-size cassette which provides 4.5 hours of recording
time. Some models have the facility to take either size of cassette.
9 Camera switching and multiplexing
On larger installations it is clearly impractical to have one monitor foreach
camera, and therefore some means of image selection must be employed.
There are basically two options available: we can sequentially switch
between cameras, or we can display a number of camera images on a single monitor screen. Both of these methods have their strengths and weaknesses. Switching implies that the operator is only able to watch a single
picture at a time (although as we shall see later , more than one monitor
may be used) and therefore each area has a period during which it is not
being monitored. Multiple imaging per haps removes the blind spots;
however, where there are too many images on display, it becomes difficult
for an operator to maintain the concentration r equired to look at all of
these images, and the r educed size of the images often leads to a loss of
definition.
Of course, these ar e very much generalizations, and developments in
technology have done much to overcome the disadvantages, often by bringing the two concepts together.
Sequential switching
The principle behind video switching is shown in Figur e 9.1, where simple sequential switching is performed by the electronic switch S1. The rate
at which the switch scans the inputs, known as thedwell time, would be set
by the operator using a simple keypad input.
Timing
control
1
Video 2
inputs 3
S1
Output
4
Figure 9.1 Principle of sequential analogue switching
A more practical sequential switcher is shown in Figure 9.2. This arrangement makes provision for alarm inputs. Under normal circumstances S3 is
closed and the timing control circuit operates S1 and S2 in tandem. Hence
the video r ecording equipment is r ecording whatever the operator is
Camera switching and multiplexing
205
viewing. Upon r eceipt of an alarm, for example on input 3, switch S
3
opens, S2 is moved permanently to position 3 and the warning output is
activated. In this condition the main monitor continues to display all four
cameras sequentially; however, the operator can investigate the cause of
the alarm on a spot monitor whilst recording the activity. The operator can
be alerted to the alarm by a light and/ or buzzer connected to the alarm
output, which may also be used to activate the alarm input on a DVR/VCR
(see Chapter 8).
Before proceeding it is worth r eminding ourselves of the importance of
output termination, which was discussed in Chapter 7 (see Figure 7.20). Any
unused outputs should be correctly terminated using either the termination
switch on the back of the equipment (if fitted), or a 75
termination
device. In all modern CCTV equipment switching is automatic, the 75 resistor on an output being switched out of cir cuit when the BNC connector is pushed into place. The rules for termination apply to all of the equipment which we shall be looking at during the course of this chapter.
Timing
control
S1
Output to
monitor
1
2
Video
inputs 3
S3
4
S2
1
Alarm 2
inputs 3
4
Alarm
input
monitor
Figure 9.2 Sequential switcher with alarm inputs
Output to spot
monitor/video recorder
Alarm
relay
Alarm output
terminals
206
Closed Circuit Television
Sequential video switchers have no means of maintaining synchronization between cameras and ther efore, unless video signal synchr onization
is performed elsewhere such as in a DVR, the picture is likely to roll each
time the switch toggles because of the difference in the scanning positions
at each camera. This is illustrated in Figur e 9.3, where the phase relationship between the video signals fr om two unsynchr onized cameras is
drawn as they would appear if they were displayed on a dual beam oscilloscope. From this illustration it can be seen that camera synchr onization
is a matter of ensuring that each camera begins scanning the first TV field
at precisely the same instant. When this condition exists between every camera in an installation, there will be no picture roll when a switcher toggles
between cameras because the monitors will not see any change in the timing of the vertical sync pulses.
Field 1
Field 2
Output from
camera 1
Vertical sync periods
Field 1
Field 2
312 line sync pulses during
each field
Field 3
Output from
camera 2
Figure 9.3 Relationship between two unsynchronized cameras as they would
appear on a dual beam oscilloscope adjusted to the field rate
One common method of camera synchr onization is to line lock the
cameras using the 50/60 Hz mains cycle as a r eference. To achieve this a
sample of the mains frequency is passed into the sync pulse generator circuit
within each camera (refer to Figure 6.11). Of course, all of the cameras in
the system will require a mains frequency reference, which means that they
will have to be fed fr om either a 230/120 V mains or 24 Va.c. supply. The
sync generator is designed to trigger to one of the two points in the mains
cycle where it is passing thr ough zero, which in the example given in
Figure 9.4 is at each negative going transition. If every camera is triggering to this same point in the mains cycle, then the field sync pulses will be
aligned. Problems occur when cameras are fed from different phases of a
three-phase mains supply and, to compensate for this, cameras with the
Camera switching and multiplexing
207
line lock facility incorporate a vertical phase control which, when adjusted,
alters the timing of the sync generator cir cuit. The adjustment normally
has a range of at least 120°, which is the difference between any two phases
in mains supply. This contr ol may be set by trial and err or; however,
where this is done it is possible that some cameras are set ‘just on the edge’
and, when changes occur in the mains supply voltage, these cameras may
manifest problems of picture roll during switching.
To luma process and
CCD driver
Low voltage 50 Hz
from power supply
Divide by
2
25 Hz
Sync
generator
Phase control
Trigger points
Crystal
oscillator
Figure 9.4 Line lock arrangement in a camera
A more sure method of adjusting camera phase is to use an oscilloscope
to display the vertical sync phase as shown in Figur e 9.3. One camera
must be selected as a reference, and the output from every other camera is
then set against this, the V phase contr ol being adjusted until the vertical
sync pulses are aligned. Of course, this operation is not as simple as it may
sound because the oscilloscope may be located at the control room in order
to access the camera inputs, and ther efore the engineer will be unable to
view the scope display from the camera location. Using two people with a
mobile radio or mobile phone is one solution to the problem. A better solution is to take the oscilloscope out to the camera that is to be adjusted
(a hand-held scope is clearly the most practical device for this operation) and
connect the camera directly to the scope. The feed from the reference camera may then be pr ovided by linking the cables fr om the two cameras at
the control room. The principle is shown in Figure 9.5.
Another method of maintaining synchronization between cameras is to
use genlocking. This is where a vertical sync pulse is fed to an input at each
camera which is labelled ‘genlock’, ‘ext sync’, etc. (see Figure 6.11). Not all
cameras have this facility, and it is therefore important that suitable models are chosen for installations which are to employ genlocking. There are
two ways of deriving a master sync pulse. One is to take the video output
from one camera and distribute this to all of the other cameras in the system. In this case the sync selector switch at the first camera would be set
to the ‘internal’ position, and the switches on all subsequent cameras set
to the ‘ext sync’ position. This method is illustrated in Figur
e 9.6.
208
Closed Circuit Television
Camera 1
Control room
Normal connection
configuration
Matrix/MUX
Camera 2
Camera 1
(reference camera)
Control room
Link
Circuit modified to
facilitate genlock
adjustment
Matrix/MUX
Camera 2
(camera under adjustment)
Oscilloscope
Ref signal
Figure 9.5 Method for performing genlock. Both cables are disconnected from
the matrix/MUX and are linked together at the control room. Hence, the cable for
Camera 2 is utilized as a feed for the reference signal from Camera 1
The second source of a sync pulse is to use a master sync generator which
would be located in the contr ol room, the sync signal being distributed
around the system.
Genlocking is a very r obust method of synchr onization, but it r equires
the installation of two co-axial cables to each camera location, which incr
eases
the cost. The only way that this can be avoided is to employ camera/
switcher combinations which send a sync signal thr ough the video signal
cable and into the video output socket on the camera. Such systems tend to
be self-contained in that you can only use the cameras and equipment that
have been specially designed to operate together, which is not a problem as
long as you are not intending to extend an existing conventional system, or
are required to extend the system beyond its capacity in the future.
Genlocking works well in a television studio wher e all cameras ar e
linked using custom multicor e cables. However, because of all the extra
cable installation r equired, it was never r eally viable for CCTV and is
Camera switching and multiplexing
209
Sync select
INT
EXT
Video switcher terminal board
Video inputs
1
Video outputs
2
3
4
In
Outputs
Video distribution amplifier
Figure 9.6 One method of genlocking cameras. In this case all cameras are
locked to the vertical sync pulses generated by Camera 1
largely obsolete. Similarly, the need to line-lock cameras is now rare because
the majority of DVRs are able to perform camera synchronization.
Matrix switching
Small-scale switching is generally performed in the DVR/NVRs. However,
for large CCTV installations there is still a need for matrix switching. As
the term matrix implies, this equipment is capable of selecting any one of
a number of inputs and connecting it to one or mor e of a number of outputs. The principle is illustrated in Figur e 9.7, which shows an 8 in 2 out
matrix.
The electronic switches are controlled by the microcontroller chip. When
the operator selects output 2, input 4, the switch corresponding to these coordinates is closed, and the signal from input 4 will be present on output 2.
In this example the control console is simple and does not offer many features other than an alarm facility similar to that shown in Figure 9.2 (the
camera input corresponding to the activated input being made available
on output 2) and a sequential switching option on output 1.
210
Closed Circuit Television
Video inputs
Alarm input
(8)
1
2
3
4
5
6
7
8
1
Micro
controller
Outputs
2
Switching control data
lines
1
2
Sequential
3
Input select
4
5
6
1
7
8
2
Output select
In/Out selector
Figure 9.7 Principle of matrix switching. In a practical, modern switcher the user
interface is usually a PC software application
One essential feature of a matrix switcher is that it can be designed to be
expandable in terms of both inputs and outputs. T aking the example in
Figure 9.7, if the manufacturer adopts a modular design, then the number
of inputs can be increased to 16 by the insertion of an eight-channel input
card identical to the first. Further cards may be added to give 24, 32, 40, 48,
etc., inputs. Some larger units are capable of expanding up to many hundreds of inputs. Similarly, output expansion car ds enable the number of
available outputs to be incr eased. An expanded version of the unit in
Figure 9.7 is shown in Figure 9.8.
In this arrangement the expansion car d mimics the first input car d (in
this example an output expander has not been installed). The alarm input
facility may not be required in all systems; therefore, the alarm expander
may be separate from the video card so that it can be included as an option.
Beyond 16 inputs the switch selector type shown in Figur e 9.7 becomes
impractical because of the number of buttons r equired and, for many
years, a ten-digit keypad pr oved to be mor e practical, although modern
switchers employ a PC-based software user interface.
The type of equipment shown in Figur e 9.8 is a vast impr ovement on
the sequential switcher; however , as the number of inputs is incr eased,
certain limitations become apparent. Most noticeably, it would be difficult
enough for a single operator to switch thr ough 16 cameras, but to cope
with any more would make the task impossible. This problem is overcome
by enabling the system to have more than one control keyboard, and hence
more than one operator . However, this creates another problem: how to
Camera switching and multiplexing
211
Alarm input Alarm input
(8)
(8)
Alarm input
expansion
card
Video inputs
1
2
3
4
5
6
7
8
Buffer
Micro
controller
Video inputs
9
10
11
12
13
14
15
16
1
Outputs
2
8-channel expansion card
1
2
3
4
5
6
7
8
9
Camera
0
Monitor
Enter
Keyboard controller
Figure 9.8 An 8 channel matrix with one 8-channel expansion card installed
prevent a number of operators from trying to access the same inputs at the
same time. This can be taken care of in the system programming whereby
each keyboard is given access to only certain cameras (inputs), i.e., each
operator ‘patrols’ a given ar ea within the CCTV system. Furthermor e,
these inputs will only be placed onto the monitors (outputs) assoc iated
with that operator, thus preventing one operator from bringing images onto
another operator’s screen. The pr ogramming can be further enhanced to
allow two operators access to the same inputs; however , one of them will
always have priority so that in the event of both attempting to access the
same input, the operator with priority will always gain the access.
Another limitation of the simpler matrix systems is the inability to control the cameras. It is to be expected that a CCTV system with many dozens
212
Closed Circuit Television
1
2
3
4
5
6
7
8
40 video input
telemetry output
Matrix switcher and
telemetry control
Terminal 1
Terminal 2
Operator 1
(Supervisor)
Access to any input
and outputs 1 & 2.
Has priority over
all operators
On-screen text
generator
Terminal 3
Terminal 4
Operator 3
Access to inputs
17–24 & 33–40
outputs 5 & 6
Operator 2
Access to inputs
1–16
outputs 3 & 4
Operator 4
Access to inputs
25–40
outputs 7 & 8.
Has priority over
operator 3
Figure 9.9 A large matrix controller with multiple operator facility
of cameras must require some of them to have a pan, tilt and zoom (PTZ)
facility. This is why the lar ger matrix units have the capability to of fer
telemetry control. Because the number of PTZ outputs r equired will vary
from system to system, the telemetry capacity is usually expandable in a
Camera switching and multiplexing
213
similar way to the alarm input capability . Telemetry control will be discussed in more detail in Chapter 10.
A large matrix system illustrating many of the featur es we have been
considering such as multiple operators, levels of operator access and telemetry control is shown in Figure 9.9. Note the inclusion of a DVR at each
operator position, enabling incidents to be recorded whilst still being able
to continue monitoring. Also note the inclusion of an on-scr een text generator. This is essential for two r easons: in a lar ge system it would be
impossible for the operator to know what he/she was looking at without
some form of camera and/or area identification facility, and in the UK it is
a requirement for all video evidence presented in court to have a time and
date displayed on the screen.
The problem with such a system is that, despite the sophistication of the
switching control, each operator is still only able to view one scene at a time,
and for a large number of cameras it is simply not possible for a person to take
in a large number of changing images for any length of time. The problem
can be alleviated to some degree by making use of the alarm inputs in such a
way that each operator only scrolls through a small number of areas, selecting
the others only if and when an alarm signal is received. However, reliance
on an alarm detector may not be satisfactory for some situations, and the
system could never be made to work in a town centr e situation where all
of the alarms would be triggering all of the time! It is clear that ther e are
occasions when the operator needs to see a number of images simultaneously, and for this we need some form of screen splitting equipment.
The quad splitter
As the name implies, this equipment allows four camera images to be displayed on a screen simultaneously. In order to be able to achieve this, the
incoming analogue signals must be digitized and stor ed in a frame store.
Once in the store a central processor chip (CPU) has to r educe the size of
each image to one quarter of the original, and arrange the data so that when
it is clocked out and converted back into analogue form, the monitor will
display four images in four quadrants on the screen. To do this, each horizontal line will contain video information for one picture during the first
26 s, and for the adjacent pictur e during the second 26 s (remember
that, for UK television, one active line period equals 52 s). Likewise the
top and bottom halves of each field contain information fr om different
cameras. The process is illustrated in Figure 9.10.
The fact that each horizontal line now contains information which has
been derived from two horizontal lines belonging to two different cameras
implies that the horizontal sync pulses must have been r emoved by the
quad unit, and a new sync pulse inserted. The same applies to the vertical
sync pulse. This means that, because the quad is acting as a form of master sync generator (although it is not actually synchr onizing the cameras
214
Closed Circuit Television
4 analogue
video input
Operator
keyboard(s)
1
2
3
4
Replay
Clock
osc
A/D converter
Read/Write
control
CPU
Digital store area
Address
data
Output bus
1
2
3
4
D/A converter
Output control
Analogue outputs
Figure 9.10 Basic signal process in a quad splitter
themselves) there is no requirement for the cameras to be synchronized by
any other means. In other words, there is no requirement for line locking
or genlocking of the cameras.
The problem with compressing a 52 s line into 26 s is that the horizontal resolution must be reduced, in theory by 50%. In basic quad switchers this is precisely what happens and so, although the operator can view
four cameras simultaneously, it is at the expense of pictur e quality. Also,
because it takes time to process the information, it is not possible to show
four live action images; rather, the images move in a succession of frame
jumps. To enable the operator to view high-r esolution live images, the
quad normally has the ability to display any selected input in normal full
size, and alarm inputs may be included to pr ompt an operator to select
this mode.
In the basic quad the pr oblem of resolution loss becomes mor e acute
when the signal is r ecorded onto videotape because ther e is no way of
decompressing the compressed information. The implication is that, when
the tape is r eplayed, the four images ar e permanently compressed, and
even though the quad unit may have a VCR input facility which enables
any one image in the replayed quad format to be separated and displayed
at full screen size, the resolution is still only 50% of the original, both vertically and horizontally. In DVRs this is not usually a pr oblem because
each image will still be recorded at full resolution.
Earlier quad units wer e limited by the speed at which the CPU could
manage the data process. If we examine this pr ocess in relation to Figure
9.10, we see that when the unit is operating in the quad mode the CPU has
to clock four unsynchronized TV frames into the data stor e, strip out the
line and field sync pulses, compr ess the data into 26 s horizontal line
periods, add new line and field sync pulses, and clock the data out of the
Camera switching and multiplexing
215
frame store and into the D/A converter in the order that it is to appear on
the screen. For this to be done in real time takes a considerable amount of
processing speed and power , and the r elatively low clock speeds of the
early microchips meant they were simply not up to it.
Developments in microchip technology have brought about giant leaps in
processor clock rates, and this has enabled quad splitters operating in r eal
time to be developed. Some of these ar e capable of accommodating eight
inputs. In normal quad mode the images ar e arranged into two gr oups of
four, each gr oup display being alternated on the scr een, with a variable
dwell time.
By increasing both the memory capacity of the digital store area and the
CPU clock speed, the rate at which pictur es are processed is incr eased,
enabling a much faster rate of update. But per haps the most significant
improvement is in the fact that, although the r
esolution is initially
lost when the images ar e converted to quad mode, because modern
digital compression techniques ar e used (i.e., MPEG or W avelet), the
resolution can usually be r ecovered when the images ar e restored to full
screen size.
Video multiplexers
The multiplexer (MUX) can be said to take over from where the quad leaves
off. Taking advantage of the faster picture processing made available through
high-speed CPUs and digital compr ession techniques, multiplexers of fer a
whole range of facilities that greatly enhance the effectiveness of a CCTV system. A typical multiplexer will offer a range of screen displays; some typical
examples are shown in Figur e 9.11. However, as we shall see later in this
chapter, the multiplexer is not only capable of producing some clever screen
displays; sophisticated models are capable of delivering two completely different picture sequences or structures to the monitor and the DVR, a facility
which may be further enhanced when used in conjunction with alarm
inputs.
The operating principle of the multiplexer is illustrated in Figur e 9.12.
Following the process through, the analogue composite video signals coming from the cameras ar e immediately converted into digital signals and
stored in a temporary frame stor e area. Working at very a high speed, the
video compression encoder reduces the amount of data per TV frame before
placing this data into another store area. The video compression and expansion process is a complex operation r equiring a lot of processor power, and
to prevent the system fr om being slowed down, or even pr one to lock-up
(crashes), it is common for the compr ession circuits to have their own
control CPU, freeing up the main CPU to deal with the multiplexing and
other housekeeping operations. The compr ession circuits also have their
own dynamic RAM (DRAM) memory stor es to avoid clashes within the
frame store areas.
216
Closed Circuit Television
9 way
16 way
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
9 way with some images switching
1
2
3
4
5
6
7/8/9
10/11/12
1
2
3
4
5
6
7
8
9
Smaller images enlarged where required
1–4
selectable
13–16
5–8
selectable
1
2
3
4
5
6
7
8
Picture-in-picture (PIP)
5
1
Figure 9.11 Typical examples of screen layouts available with a multiplexer
Multiplexing takes place as the pictures are moved out from the digital
store area 1 by the main CPU, which clocks the data to the D/A converter
in an order that will pr oduce the desired picture layout on the monitor
screen. When the operator alters the layout via the keyboar d, the CPU
simply changes the or der in which the data is clocked out. Fast addr ess
and control buses are required between the CPU and the memory chips to
sustain such complex data control.
Where a camera input signal is lost, the MUX will usually detect the
loss of sync pulses and display a warning on the scr een. When a sequential switching mode is selected, the MUX skips any missing or unused
channels – a feature that is also found in many sequential switchers.
Camera switching and multiplexing
16 analogue
video input
User interface
Address
data
Main CPU
Control
data
A/D converter
Compression
control CPU
Digital frame
store
Control data
Video compression
encoder
(MPEG or Wavelet)
DRAM
Digital store
area
Video motion
detection (VMD)
Video compression
decoder
Integral DVR
217
Digital
store 2
Alarm
output
DRAM
Digital
store 1
D/A converter
Monitor
out 2
Monitor
out 1
Output buffers
(see text)
Figure 9.12 Signal processing in a multiplexer (simplified)
In units which have the pr ovision for dif ferent monitor outputs, a
second store area is required (Digital store 2 in Figure 9.12) so that data can
be output to the D/A converter in different sequences for each output.
The main implication for VCR r ecording is that, unlike with the quad,
the machine is able to r ecord complete TV fields. Replaying a tape with
this recording pattern dir ectly into a monitor would pr oduce unintelligible images; however, when the recording is replayed through the multiplexer, the images ar e able to be manipulated into whatever format the
operator desires, i.e., full screen, quad, 9-way, etc. Note that in full-screen
mode the movement may appear jerky because the VCR has not recorded
every field, and therefore the multiplexer is effectively having to display a
series of still frames.
Multiplexers are categorized as simplex, duplex or triplex. A simplex
MUX has just one multiplexer (Figur e 9.13), and therefore its capabilities
218
Closed Circuit Television
are limited to recording full-screen images whilst viewing live pictures in
a number of screen layouts. When the operator wishes to replay a recording through the simplex unit, both live monitoring and r ecording is lost
because the MUX is r equired to de-multiplex the r eplayed images. Note
that the unit is still able to of fer the same scr een layout options for
replayed video as it does for live inputs.
MUX
1
2
3
4
Integral DVR
Figure 9.13 Principle of a simplex MUX
The duplex MUX incorporates two multiplexers (Figure 9.14), enabling
one to handle the live information whilst the other takes care of the recording
requirements. Therefore, a duplex MUX is capable of displaying live images
in a range of screen layouts whilst recording full-screen images. It can also
replay images without interruption to either the live viewing or recording.
MUX 1
1
2
3
4
Note:
White screen colour indicates live display
Grey screen colour indicates replayed display
MUX 2
1
2
3
4
Figure 9.14 When used with two monitors, a duplex MUX can continue normal
operations whilst a recording is being viewed. Note that the DVRs may be
external to the MUX, or a single internal device
The triplex MUX principle is illustrated in Figur e 9.15. The unit does
everything that the duplex can do; however, it is possible to view live and
recorded images simultaneously on the same monitor screen, without interruption to normal recording. The operator still has all of the usual options
of picture enlargement or PIP. Alternatively, two monitors may be used to
display live and recorded images in various screen layouts.
Camera switching and multiplexing
Cam 1–4
Cam 5–8
219
Cam 1 Cam 2
MUX 1
DVR DVR
Cam 1 Cam 3
Note:
White screen colour indicates live display
Grey screen colour indicates
replayed display
Cam 5 Cam 6
MUX 2
Cam 7 DVR
Cam 8
Cam 9–12
Cam 13–16
MUX 3
Figure 9.15 A triplex MUX can display both live and recorded images on one
screen
It is not uncommon for an MUX unit to incorporate telemetry controllers;
this subject will be covered in Chapter 10.
On-screen information is essential for identifying the date, time and
location of video evidence, as well as assisting the operator in identifying
an area quickly. The MUX usually has an on-screen display (OSD) facility,
which is set up during installation.
Video motion detection (VMD)
Many CCTV installations rely on intruder alarm technology detectors such
as passive infrared, infrared beam and micr owave, to activate the alarm
inputs; however, an alternative to this is to use the images coming fr om
the video cameras as a means of intruder detection.
The principle of the VMD is to stor e a sample frame fr om the camera
input and then compare subsequent frames with the sample, looking for
changes in picture detail. Naturally it cannot look for a change in all of
the picture information before triggering an alarm, otherwise the intruder
would have to fill every part of the scr een. Instead, the VMD divides the
screen area into detection zones, and a change in one or more of these can
trigger an alarm. By using zones the VMD can be given a degree of intelligence. For example, if used in an outdoor location, a sudden darkening of
the entire scene caused by a cloud passing acr oss the sun can be ignor ed
by the VMD. Similarly, where one zone is equivalent to only a very small
220
Closed Circuit Television
area in the field of view , the unit can be pr ogrammed such that an activation of just one zone will also be ignored, as in all probability the would-be
intruder is more likely to be a bird or other small creature.
The unit is designed to detect changes in the contrast level in each
detection zone, and the sensitivity may be adjusted by altering the amount
of contrast change required to initiate an alarm.
Analogue VMD units have been available for many years, but their sensitivity was limited, making them pr one to false alarms due to sudden
changes in lighting level, movement of small animals, etc. Digital VMD
units take the analogue signal and immediately pass it into an A/D converter. The reference picture is placed into a frame store so that each subsequent frame can be compared with this image. Digital units are capable of
breaking the scr een up into many thousands of detection zones, giving
them a great deal of sensitivity, and the operator is able to control this sensitivity using two parameters: the amount of gr ey scale change and the
number of zones over which this change takes place. To make the sensitivity adjustment simple for the operator to use, the unit often has some form
of sliding scale on the OSD which is calibrated using arbitrary numbers.
The operator is normally able to set the detection ar ea(s) for each camera by using an array of rectangles, circles, etc., which are generated by the
OSD. This is illustrated in Figure 9.16. Note that these detection areas are
not to be confused with the much smaller detection zones used by the VMD
microprocessor for image analysis. Also note that the letters and numbers
in the illustration ar e for the purposes of r eference only, and would not
appear on an actual screen.
A
B
C
D
E
F
1
2
3
4
5
6
7
8
9
10
Figure 9.16 A video motion detector divides the image into a series of zones.
The operator (or engineer) decides which zones are to be active
In Figure 9.16, let us suppose that there is legitimate movement in front
of the building, but the operator is only concerned about intr uders moving along the flat roof area. The rectangles between points D2 to D10 (and
possibly C2 to C10) could be made active, thus causing an alarm to be generated when there is any movement in this area.
Camera switching and multiplexing
221
The majority of MUX units incorporate VMD as standard, making them
an attractive choice of CCTV control equipment in small- to medium-sized
systems, especially if telemetry is also included.
VMD can also be found in some cameras that have digital pr ocessing,
the alarm condition being present on a pair of switch contacts which can
be connected to a MUX or other alarm input.
An important point with VMD is that the camera must be r easonably
stable. If it is prone to movement in high winds, etc., the movement of the
image can result in false alarms. Some VMD processors have in-built compensation for this effect, but nevertheless every system has its limitations,
and a secure mount is recommended.
10 Telemetry control
Without some form of remote control, larger CCTV installations would be
of limited use. The operator needs not only to be able to adjust the angle,
zoom and focus of cameras, but also to perform other ‘housekeeping’ tasks
such as washing and wiping of the fr ont glass on external housings and
control of lights. Some of these commands ar e quite simple. T ake, for
example, the wash command; it is simply an on/ off situation. However,
other control commands ar e more difficult to send. Ther e is little to be
gained from simply ‘telling’ the pan motor to ‘pan’. Per haps if the motor
could answer back it might ask, ‘Which way? For how long? How fast?’
Operators of fully functional cameras have come to expect a lot from their
systems, and to some extent ar e attempting to emulate the actions of a
broadcast camera operator, but from a distance. The number of commands
needed to provide such complex control is substantial, and requires sophisticated telemetry to enable adequate communication. A list of the main
commands is given in Table 10.1. When the command includes speed, this
refers to controllers that have a dynamic (or ballistic) joystick wher e the
further the stick is pressed, the faster the motor moves. Thus, for example,
an instruction may not just be ‘pan left’ but rather ‘pan left at this speed’.
Table 10.1 Major telemetry control commands
Command
Number of instructions
Camera address (i.e., number)
Pan left speed
Pan right speed
Tilt up speed
Tilt down speed
Zoom in
Zoom out
Focus far
Focus near
Iris open
Iris close
Wash on/off
Wipe on/off
Lights on/off
Move to preset position n
1
2
2
2
2
1
1
1
1
1
1
2
2
2
Dependent on the
number of presets
Telemetry control
223
From the table we see that there are no fewer then twenty-one commands
for a fully functional camera, plus an additional number for the pr eset
positions. In this chapter we shall look at ways of communicating this
information to the various cameras in a system.
Control data transmission
Some of the earliest remote control CCTV systems relied on a hard-wired
link between the contr ol console and the PTZ head. However , these systems required a lot of multicor e cabling between the contr ol room and
each individual camera site and, in some cases, were prone to the effects of
voltage drop along the cables.
A much more effective alternative to this method of control was to send
digitally encoded PTZ commands along a twisted pair cable in the form of
RS 422 or RS 485 (see later in this chapter). The principle is illustrated in
Figure 10.1. Each command is encrypted into a data format and sent along
a two-wire link to a receiver. The receiver (site driver) contains a decoder
chip which interprets the commands and operates the appropriate relay(s)
via the relay driver chip.
RL2
To other
receiver units
Pan
M motor
Relay
driver
RL4
Controller
010101
RL1
RL3
010101
M
CPU
Chip
Tilt
motor
Power
supply
Figure 10.1 Control of motors using a separate data link such as RS 485 over a
twisted pair
There are two ways that the data links may be connected to the receiver
units. The first is like that illustrated in Figur e 10.1, where the controller
uses individual outputs to each receiver. The other way is to place the control data onto all of the lines simultaneously, enabling receivers to be connected individually to the controller and/or daisy chained. This is illustrated
in Figure 10.2.
224
Closed Circuit Television
Controller
Rx1
Rx2
Rx Receiver
Rx3
Rx4
Rx5
Figure 10.2 Simultaneous data transmission
Where a daisy-chain design is used, the encoder includes an address in
the command. For example, if the operator selects camera 5 to pan left, then
the encrypted data is effectively saying, ‘camera 5, pan left’. Note that the
first part of the command contains an address – ‘camera 5’. Thus, although
the signal is picked up by all of the receivers, only the one which has been
assigned as camera 5 responds. The receivers are assigned their addresses
during commissioning. Note that a controller designed for individual output transmission cannot be connected daisy-chain fashion becauseit will not
be sending any address data.
Twisted pair telemetry transmission is very effective and was employed
by a number of leading CCTV equipment manufacturers for a number of
years. However, from an installation point of view, the requirement to run
both co-axial and twisted pair cables to every camera was not ideal. For
this reason, manufacturers looked for ways in which both the video and
telemetry signals could be sent along the same co-axial cable. The trick is
to keep the video and data signals separate, otherwise the entir e system
will break down as data would be displayed on the monitors as noise, and
the telemetry receivers would become confused trying to decode a video
signal. The two common methods of multiplexing data and video ar e frequency division multiplexing (FDM) and time division multiplexing (TDM).
Frequency division multiplexing is illustrated in Figur e 10.3. The data
is modulated onto a carrier signal at a frequency in the order of 8–12 MHz,
which is well above that of the upper limits of the video signal which, for
PAL, is typically 5.5 MHz (4.2 MHz NTSC). Each telemetry receiver has a
demodulator circuit which removes the data from the carrier. The decoder
can then process the data to derive the telemetry commands.
Data carrier signal
Video signal
f MHz
0
5.5
8
Figure 10.3 Signal spectrum for frequency division multiplexing
Telemetry control
225
Although popular for a number of years, the majority of CCTV telemetry equipment today operates using time division multiplexing, which is
a take-off of the Teletext system used in the UK br oadcast television service. Here data is transmitted during the field flyback blanking interval,
during which time the electron beam in the CRT is cut off whilst the scanning circuits adjust themselves to begin scanning the following field from
the top of the screen. When viewed on an oscilloscope, the signal appears
similar to that shown in Figure 10.4. As with FDM, the telemetry receivers
each have a decoder cir cuit which picks out the data transmission and
deciphers the commands.
Data packets transmitted during
the field blanking period
Field sync
period
Field flyback period (20 lines)
Figure 10.4 Time division multiplexing. The data is sent at a different time to
the video information
The TDM signal is at a frequency somewhere in the region of 4.5 MHz,
and thus we see that in the case of both TDM and FDM, it is essential that
the bandwidth of the system is up to specification.Any faults in the cabling
or termination that would introduce a high-frequency filtering effect will
not only remove the high-resolution components of the pictur e, but may
also filter out the telemetry data. Such faults can be intermittent, or may
only affect certain camera locations.
Pan/tilt (P/T) control
The pan and tilt unit comprises two motors and a number of gears to convert the motor speed into torque. The motors may be 24 Vd.c., 24 Va.c. or
230 Va.c. Whilst a.c. motors are generally more efficient and often produce
greater torque than their equivalent-size d.c. counterparts, speed contr ol
of an a.c. motor is somewhat mor e complex than for d.c. motors. Thus,
single-speed 230 Va.c. motors are ideal where it is anticipated that high
winds are likely to exert a heavy load on the camera assembly, and a high
torque drive mechanism is r equired to over come this. However, where
dynamic joystick control is to be incorporated in the system, d.c. P/T units
will allow multi-speed (where the speed alters in incr emental steps) or
variable-speed operation. It is usual for the motor to be coupled to the
226
Closed Circuit Television
gears via clutch assemblies which will slip in the event of the mechanism
jamming, thus protecting the motors from stalling and burning out.
For simplicity the cir cuit in Figure 10.1 shows d.c. motors. Wher e a.c.
motors are connected to the driver cir cuit the arrangement is similar;
however, simply reversing the polarity of the a.c. supply to the motor will
not cause it to reverse. To achieve this action, the r elay changeover contacts
must reverse the connections to the motor field windings. A.c. motor theory
is not something that the CCTV engineer needs to become too involved with;
however, it is important when installing or replacing P/T units to ensure that
they are compatible with the drive voltage on the site driver. Applying 230 V
across a 24 Vd.c. motor is likely to have disastrous results! Some site drivers
have the facility to select the type of drive voltage (see later in this chapter).
A P/T unit contains at least four limit switches. These are a form of cutout which are adjusted to set the maximum points of deflection in both the
horizontal and vertical directions. Without these the motors could simply
drive the unit in cir cles, causing the inter connecting cables between the
housing and the r eceiver to become twisted ar ound the mounting until
they break, or until the housing becomes jammed. During installation the
engineer adjusts the limit switch positions to suit the particular location,
but in most cases the unit is designed such that it will not pan thr ough
more then 355° before stopping.
Limit switches are satisfactory if the camera is only going to be moved
occasionally; however, with permanently manned systems the need became
evident for a facility whereby a number of predetermined positions could
be set for each fully functional camera location. For these pr esets to be
effective it is not only necessary to fix the pan and tilt positions; the zoom
and focus settings for that position must also be fixed. Clearly mechanical
limit switches are not able to offer this amount of control, and some other
means had to be devised.
The most common solution is to place variable r esistors inside the
pan/tilt unit and the zoom lens assembly (this was discussed in Chapter
4, Figure 4.19). As the motors move each of the pan, tilt, zoom and focus
mechanisms, the associated potentiometer r otates, producing a d.c. feedback voltage which can be monitor ed by the contr ol circuit. During programming, the operator moves the P /T unit to the desir ed position,
adjusts the zoom and focus, and enters these into a memory store, allocating a preset number. At this instant the control unit measures the d.c. feedback from each potentiometer, passes this to an A/D converter, and stores
the digital values in memory . During operation, whenever the pr eset
number is selected, the control unit checks the position of the potentiometers, and moves the four motors until the feedback voltage fr
om each
potentiometer is equal to the value stored in memory.
With this contr ol method the pr esets and limit positions ar e not set
mechanically, but in the PTZ controller. Thus there is, in theory, no limit to
the number of possible preset positions. Depending on the controller, as few
as five, and as many as one hundred, preset positions can be programmed.
Telemetry control
227
A large number of positions is not a lot of use to a CCTV operator because
it is not feasible to expect him/her to remember every one. Consider a tencamera system where each camera has one hundred preset positions programmed: who could remember one thousand preset position numbers?
However, many control systems have an automated patrol facility whereby,
in the absence of any activity or per haps at times when the system is
unmanned, the operator can leave the cameras moving through a set pattern
of positions. It should be pointed out that where this facility is used extensively, the wear and tear on the P/T unit gears and clutches can take its toll.
Another use of multiple pr esets is in conjunction with alarm signals
(see Chapter 9). Upon r eceipt of a particular alarm signal input, the controller can often be programmed to move one or more cameras to a given
position and prioritize these video signals in the recording process.
Receiver unit
A typical telemetry receiver unit is shown in Figure 10.5. In this illustration
we can identify many of the features we have discussed so far in this chapter.
Relays
Control chip
Pan and tilt
outputs
Video in
(from camera)
BNC
Video out
(to control room)
Zoom, focus
and iris outputs
Memory chip
A/D converter
Feedback inputs,
pan, tilt, zoom,
focus, iris
Auxiliary
outputs
DIL switch bank
Fuse
Mains input
terminals
L
E
N
Mains
transformer
d.c.
power
supply
12Vd.c. 24Va.c.
LV camera supplies
Figure 10.5 Typical receiver unit
Twisted pair
data input
228
Closed Circuit Television
Where the data is multiplexed with the video signal on one co-axial
cable it is necessary for the camera signal to loop thr ough the receiver in
order that the data can be extracted (bear in mind that the video signal is
passing from the camera to the control room, whereas the data is going the
other way). Some receivers may have the option of a twisted pair telemetry input, and this brings us to the next point, the DILswitches. These are
used to customize the receiver to the system and location, providing such
options as co-axial or twisted pair data input, 4-wire or 3-wire lens wiring,
5 V or 12 V lens operation, PT motor drive voltage, auxiliary output functions, etc. The corr ect setting of these switches during installation is of
utmost importance as the unit will not function correctly (if at all) if any of
these are in the wrong position for the type of installation. Manufacturers’
instructions must be carefully adhered to.
For the driver in Figur e 10.5, the pan and tilt outputs ar e effectively
those we looked at in Figur e 10.1; however , the auxiliary outputs ar e
somewhat different. Because these may be used for switching either
mains-voltage or low-voltage devices, the terminals are simply taken to a
set of clean relay contacts on the PCB. The installer can ther efore connect
the required power for the device through these switch contacts. The DIL
switches are used to assign each output to a particular task (wash, wipe,
lights, etc.) so that when, for example, the operator pr esses the button
labelled ‘lights’ on the keyboar d, the auxiliary output assigned to the
lights changes states. If, when ‘lights’ is selected, the washer starts, it is
possible that a DIL is incorrectly set.
The control chip performs all housekeeping functions in the receiver as
well as data signal decoding, output switching, etc. The memory chip stores
the preset levels which come from the potentiometers via the A/D converter chip.
With all telemetry-controlled systems, the protocols used between the
controller and site drivers vary between manufacturers, and it is important
that equipment is not mixed within a system unless it is known for certain
to be compatible. Where equipment is incompatible, in most instances the
telemetry simply will not function. To aid compatibility, a number of manufacturers include the Pelco D set of protocols in their site drivers. Pelco D
has become something of a standar d within the industry and, because
Pelco permit the use of this pr otocol, items of equipment fr om different
manufacturers that might otherwise have been incompatible in terms of
telemetry are able to communicate.
The receiver requires a local 230 Va.c. supply, which must be earthed.
All cables between the r eceiver unit and the P /T and camera must be
made to be flexible, and for pr otection are normally fed through a Copex
flexible conduit. Note that to conform with IEE r
egulations on cable
segregation, any cables carrying mains voltages must not be passed via
the same conduit as co-axial and any other extra-low-voltage cables unless
the insulation of the EL V cable is rated to withstand mains supply
voltages.
Telemetry control
229
Dome systems
Pre-assembled domes containing a P/T unit, site driver and power supply
have largely overtaken the traditional pan/tilt unit. In some cases a camera is included as a part of the assembly , but this can be somewhat limiting as the installer then has no choice of camera specification. Ther e are
domes available where a camera and fully functional lens of the installer’s
choice can be fitted. Internal and external versions are available.
There are perhaps a number of r easons for the popularity of domes.
Very often they are preferred by the customer because they are overt, and
yet covert. That is, people can clearly see that the premises are monitored
by CCTV, but they never actually know when they themselves ar e being
observed. Aesthetically, domes often look less out of place than traditional
PT and site driver assemblies. Installers have less work because they ar e
ready assembled. But per haps one feature above all of these is the availability of high-speed domes, which are capable of 360°continuous rotation.
The high-speed function enables the PT unit to move the camera at
rates in the order of 300° per second, with a corresponding response from
the zoom and focus servos. The inclusion of multiple pr esets offers rapid
response to operator or alarm-activated r equests. The speed contr ol is
made possible by the inclusion of sophisticated servos on the site drivers
controlling either d.c. stepper motors or, in some instances, a.c. synchronous motors using a variable high-fr equency supply voltage. (Note: altering the frequency of the a.c. supply to a synchr onous motor changes its
speed.)
The 360° rotation is made possible by the use of slip rings to make all of
the electrical connections, including the video signal, between the camera
and the site driver. This principle is illustrated in Figure 10.6.
Zoom
Focus
Iris
Common Video
Slip rings
Brush contacts
Figure 10.6 Principle of using slip rings to couple a rotating camera to
fixed cables
230
Closed Circuit Television
Data communications
In Chapter 11 we will take a detailed look at network communications
appertaining to both CCTV signal transmission and r emote control of
CCTV devices. However , TCP/IP over a LAN /WAN is not the only
method for telemetry control of CCTV equipment, and a number of common standard interfaces are still employed, including RS 232, RS 422 and
RS 485 (RS recommended standard) to perform functions like telemetry
control and equipment connection. To enable equipment of different industries and manufacture to communicate, some means of standardization of
signal level, polarity and frequency is necessary. Also, the configuration of
connectors between equipment must be the same. Such standardization is
found in the recommended standards.
Perhaps one of the oldest communication pr otocols used in the computing industry is the RS 232, which was introduced in 1962. This is a very
common interface and has been used extensively for many years to link
computer-based equipment. In most cases the connector type known as
the D connector is used; the pin configurations for both the DB9 (9-pin) and
the DB25 (25-pin) are shown in Figure 10.7.
The voltage levels for RS 232 are much higher than for TTLlogic circuits.
In theory, logic 0 is between 3 V and 25 V and logic 1 is between 3 V
and 25 V. However, in practice, the majority of RS 232 drivers operate
Pin 1
Pin 5
Pin 6
Pin 9
Retaining screws
Pin 1
Pin 14
Pin 13
Pin 25
Pin 1
Pin 2
Pin 3
Pin 4
Pin 5
Pin 6
Pin 7
Pin 8
Pin 9
Data carrier detect (DCD)
Receive data (RXD)
Transmit data (TXD)
Data terminal ready (DTR)
Signal ground (SG)
Data set ready (DSR)
Request to send (RTS)
Clear to send (CTS)
Ring indicator (RI)
Pin 1
Pin 2
Pin 3
Pin 4
Pin 5
Pin 6
Pin 7
Pin 8
Pin 15
Pin 17
Pin 20
Pin 22
Pin 24
Frame ground (FG)
Transmit data (TXD)
Receive data (RXD)
Request to send (RTS)
Clear to send (CTS)
Data set ready (DSR)
Signal ground (SG)
Data carrier detect (DCD)
Transmit clock (DB)
Receive clock (DD)
Data terminal ready (DTR)
Ring indicator (RI)
Auxiliary clock (DA)
Figure 10.7 RS 232, DB9 and DB25 type connectors. (Male connectors, viewed
from front)
Telemetry control
231
between 12 V, and ther e are some drivers that enable RS 232 data to
swing between just 0–5 V, although the cable length may be limited.
The main limiting factor with r egard to RS 232 is the r elatively short
maximum cable length, which is specified as 15m (50 ft). The main reason
for this short length is the amount of capacitance in the cable which, as we
saw in Chapter 2 (Figure 2.2), integrates the square pulses corrupting the
data signal. Another limiting factor is the low baud rate (the speed at
which data is transmitted, expressed in bits per second or bps). The maximum specified speed for RS 232 acr oss a 15 m cable length is 20 kbps,
although rates of up to 115 kbs may be achieved over shorter distances.
Referring to Figure 10.7:
●
●
●
●
●
●
Frame ground is a protective earth which protects the sensitive microchips
from damage which would occur if static wer e allowed to build up
around equipment cases. It also serves as a screen, preventing RFI and
EMI from corrupting the data.
The transmit data and receive data connections are the main data links
from between the two items of equipment at either end of the bus.
Request to send is used for modem contr ol. It switches on the carrier
when the data terminal equipment (DTE) is ready to send data.
Clear to send: in response to a logic 0 on R TS, the modem takes CTS to
logic 0 as soon as it is ready to transmit. In other words, it is telling the
DTE that it is r eady. Data set ready: when the modem is active, it takes
this pin to logic 0 to let the DTE know that it is connected to a ‘live’
modem.
Signal ground is a common negative to which all signals are referred.
Data terminal ready: this is the complement to the data set r eady action.
The clock signals maintain corr ect timing/synchronization between
devices for the purposes of transmitting and receiving data.
The activity between R TS and CTS is known as handshaking, the term
describing the co-operation between devices that ar e exchanging data.
This protocol is necessary because it is not possible for equipment to send
and receive at the same time. A typical handshake sequence would be:
DTE asserts R TS – slight delay whilst the modem starts up – modem
responds with a reply on CTS – DTE begins transmission of data.
The limitations of RS 232 are overcome in standards such as RS 422 and
RS 485, both of which make use of low-impedance dif ferential signals
passing along a twisted pair cable. Much longer transmission distances
are possible (in the or der of km) and, because any external noise is
induced equally in both conductors, the noise is cancelled as a er sult of the
action of the balanced cable plus the balanced inputs at the r
eceiving
equipment. Special line amplifiers/drivers ar e required to launch and
receive the differential signals. The principle is illustrated in Figur e 10.8,
whilst the principles of twisted pair transmission wer e considered in
Chapter 2.
232
Closed Circuit Television
RFI
Data
Data
Data RFI
Twisted pair
Z 120 ohms
Differential line
RFI
driver
R1
Data
Differential line
receiver
R2
R1 and R2 60Ω
Figure 10.8 Illustration of RS 422/485 transmission. RFI is cancelled both by
the action of the balanced cable, and the differential inputs.
R1/R2 form a 120 ohm termination resistance across the line; note the reference
to earth (see text)
RS 422/485 ar e described as two wir e (single twisted pair) systems.
Whilst this is tr ue, the line drivers also r ely on a gr ound connection to
derive the signal voltage and, bearing in mind what was discussed in
Chapter 2 regarding ground potential differences (Figure 2.11), problems
may arise where the potential at each end of this ground connection is different. In order to provide a reliable ground connection between devices,
many manufacturers specify double twisted pair cable, where the second
pair are connected in parallel to the ground terminals. The doubling up of
conductors ensures a low electrical resistance, and the ground connection
is much more reliable than an electrical earth because there is no possibility of a ground loop current forming.
The RS 422 standard defines a maximum data rate of 10 Mbps at 15 m
(50 ft) maximum cable length, and 100 kbps at a maximum cable length
of 1.2 km (4000 ft). The signal level is between 0 and 5 V. One feature that
makes RS 422 popular for many applications is its ability to interface easily with RS 232. By connecting a simple line driver to a PC COM port, the
RS 232 signal can effectively be transmitted over a distance of some 1 km
by converting it to RS 422, and then using a second line driver at the
receiving end to restore the signal to RS 232 protocol once more.
The RS 422 bus can include up to five transmitters and up to ten receivers,
but only one transmitter may connect to the line at any one time.
Introduced in 1983, RS 485 is able to support up to 32 devices on one
twisted pair, although modified versions ar e able to support up to 256
devices. Devices are connected in parallel across the twisted pair in a daisy
chain, and each device may be configur ed to be both a transmitter and
receiver although, like RS 422, only one device may transmit at any one
time. The line drivers are tri-state devices, which means that their (transmitter) output may be at logic 0 or logic 1, or disconnected from the line.
In a CCTV application, the first device may be a telemetry controller, and
the other 31 devices site drivers. Wher e the system is to be administer ed
Telemetry control
233
by a PC, the RS 485 network may be connected to the PC via either an RS
232 (COM) port or USB port in a similar manner to that described above for
RS 422. An illustration of a PC administered system is given in Figure 10.9.
RS 232/USB
port
Up to 32 devices (i.e., CCTV site drivers)
N1 N2 E
N1N2 E
N1
N2
E
Admin PC
Line driver
(see Figure 10.8)
Figure 10.9 The line driver converts the RS 232 or USB output to RS 485.
Devices are then daisy chained across one pair, the second pair serving as a
ground wire
In order to prevent signal reflections in the RS 485 data bus, the two
ends of the bus should be terminated at the corr
ect impedance. The
termination impedance is typically between 100 and 120 , depending on
the type of cable employed, and the design of the dif ferential line drivers
(Figure 10.8).
Both RS 422 and RS 485 specifications exclude any actual transmission
protocol, meaning that the manufacturer is free to use or develop a protocol that best suits their application. This may explain why one CCTV site
driver that employs RS 485 is not necessarily compatible with other site
drivers that also use RS 485.
11 CCTV over networks
The combination of high-speed data transmission capability and effective
video compression techniques has resulted in an inevitable shift towar ds
using IT network technology to transmit CCTV video images, both locally
and over long distances. Consequently, the CCTV engineer, who in many
cases began his working life as a time-served electrician or security/fire alarm
systems installer, now needs to have reasonable knowledge and competence
in both computer-operating systems and networking. It certainly is not the
intention to turn security systems engineers into fully fledged network engineers by the end of this chapter, but the reader will have an understanding of
network topology, how data is managed across networks, the limitations of
any network, and how all of this applies in a practical way when connecting
IP cameras and network video recorders (NVRs) and the like onto a network.
Network topology
A topology is a structured wiring diagram, or map, of a network. In computer networks ther e are five topologies in use (although some ar e far
more common than others): bus, star , ring, mesh and wir eless. These are
illustrated in Figure 11.1.
Figure 11.1a Bus topology. All devices are connected to a single continuous cable
CCTV over networks
Hub
Figure 11.1b Star topology. Often used in conjunction with bus topology
Figure 11.1c Ring topology. Devices are connected in a loop, and data is
passed in one direction between devices
235
236
Closed Circuit Television
Figure 11.1d Mesh topology. Very reliable, but difficult to install and modify
owing to the large number of cables required
Of the four hardwired topologies, star is by far the one most commonly
employed in IT networks. It is simple to install and it is very easy to add
or remove devices on the network. Star topology is often employed in conjunction with a bus, wher e a hub is connected to the bus and devices ar e
connected in a star formation fr om the hub. Bearing in mind that CA T 5
and CAT 6 network cables need to be terminated at 100 and should not
exceed 100 m in length, the hub both terminates each individual cable (or
segment) and relaunches the signal onto all cables, enabling each segment
to stretch up to 100 m. The hub also facilitates simple addition of other
devices, provided that it is not already fully loaded. Another advantage of
star topology is that a break on one segment will only disable the devices
connected to that segment, making the system far mor e robust and reliable. This is in contrast to the bus where, in the event of a cable break, devices on one side of the break will not be able to communicate with those on
the other.
In ring topology, each device is connected to its neighbour via a loop
in/loop out arrangement and the data is passed around the devices in one
direction. This arrangement is seldom employed in networks today because
it offers no real advantages over a bus, and it is more difficult to add/remove
devices because the ring has to be broken in order to do so.
Mesh is a theor etical concept and would, in practice, pr ove far too
expensive to install, and it would be very difficult to add/remove devices.
CCTV over networks
237
The illustration in Figure 11.1d only shows four devices, but imagine the
problems in connecting fifty or a hundred devices in this manner! In practice, mesh topology is used to a limited extent on the Internet in or der to
provide reliable connection.
RF wireless networks, sometimes r eferred to as ad hoc networks, ar e
popular because of their versatility; all that is r equired to establish a connection is the presence of a signal between the device you ar e using and
the network you are connecting to. Wireless networks have a number of very
obvious advantages including portability of devices, simple addition/
removal of devices and limited installation costs. Nevertheless, because of
bandwidth limitations, wireless connections are generally slower than hardwired, and in some cases the signal strength may vary over time, resulting
in intermittent loss of connection.
Networks are divided into two types: local area networks (LANs) and
wide area networks (WANs). It is sometimes dif ficult to determine exactly
where one ends and another begins, and there are differing interpretations
of this terminology. In simple terms a LAN can be defined as a cluster of
PCs (or other network devices) that are able to pick up an Ethernet broadcast and, as we shall see later in this chapter , broadcasts are blocked by
routers. So a LAN is generally confined to one side of a router. By contrast
a WAN spans across routers and, although it does not necessarily have to
cover a wide geographical ar ea, in many cases it connects between sites
many miles apart. It should be pointed out that a W AN does not have to
include the World Wide Web (or Internet). It may be a private network
comprised entirely of private lines connecting the sites. Such a W AN is
often referred to as an Intranet.
Network hardware
A typical WAN comprised of two LANs that employ both bus and star
topologies is shown in Figur e 11.2. In this topology, the backbone connection is the main bus onto which all network devices connect. Each individual connection to the backbone is known as a segment. The backbone
usually has a much br oader bandwidth than the segments as it has to
carry all of the traffic on the network.
Any device that is connected to a network is known as a
node. Most
devices connect to the network via a network interface card (NIC), and most
of these are addressed on the network using an IPaddress (see later in this
chapter). A network device that uses an IP address is referred to as a host.
The NIC, also known as the Ethernet card, performs the functions of organizing the data into frames, managing the transfer of these frames between
hosts, and error correction. The NIC also plays an important role in determining the speed of the network. In Chapter 2 we looked at possible data
speeds for CAT 5 and CAT 6 structured cable networks. However, it is not
only the cable that determines the data rate; the NICs must also be capable
238
Closed Circuit Television
Servers
NVR/MUX
PC3
1
2
IP cameras
PC4
Switch
1
2
Switch
Hub
Backbone
3
Router
Hub
Segments
PC1
PC2
WAN
Server
PC5
3
Router
Hub
4
Figure 11.2 Primary network components, including CCTV devices. This
example shows a WAN comprised of two LANs
of processing data at the required rates. For example, connecting a device
that has a CAT 5 NIC to a CAT 6 network will not produce any improvement in connection speed as the system will always be forced to operate at
the speed of the slowest device.
Every NIC is given a unique media access control (MAC) address, also
known as the hardware address, or the Ethernet address. This takes the form
of a twelve-bit hexadecimal code, separated by hyphens. For example,
00-20-4A-32-F2-3C
The Institute of Electrical and Electronic Engineers (IEEE) has been assigned
the task of administering MAC addresses throughout the world, and any
manufacturer of NIC or other network devices will pur chase a unique
CCTV over networks
239
address comprised of the first six bits of the twelve-bit code; i.e., 00-20-4A
in the example above. By making the second six bits of the MAC addr ess
unique, the manufacturer is able to assign a unique MAC address to every
device they produce. As we shall see later in this chapter, the MAC address
plays a very important role on the network.
In a device such as a PC, the NIC is normally integral to the motherboar
d,
although it may be fitted to one of the PCI slots. In a device like a network
camera (or any other network security device such as an alarm collector
point or access contr ol door contr oller), the NIC is often incorporated
within the on-board RJ 45 socket.
A hub is perhaps the simplest of all network devices because it has no
intelligence and makes no decisions. It simply pr ovides a star connection
for (typically) four or eight devices, with a 100 termination for each connection. Most modern hubs ar e active devices, which means that they
incorporate separate amplifiers for each input/output connection to ensure
reliable data transmission without attenuation. The amplifier action also
serves the purpose of doubling the distance between one host and another.
In Figure 11.2, the maximum distance between the NVR and IP cameras 1
and 2 would be limited to 100m. The inclusion of the hub effectively doubles
this to 200 m. There is a limit as to how many hubs may be used as repeaters on one line to extend the cable length, because eventually the S/N ratio
will become so poor that connection between NICs will be lost.
In terms of data transmission, a hub does not r educe network traffic
because it simply passes all traffic to all hosts. For example, in Figure 11.2,
the images from IP cameras 1 and 2 may be intended only for the NVR,
but the data packets from these cameras will pass through both hubs and
will be present on the segments connecting the two PCs. This will slow
down that part of the network because the PCs will have to wait their turn
to gain access in order to transmit their information.
A much more effective way to create network segments is to use aswitch.
Although at first glance it appears like a hub, a switch is intelligent enough
to recognize the intended destination for any data packets and will only
place those packets on the corresponding output. Thus, a switch creates a
dedicated path between any two hosts for the duration of their communication. This action greatly improves network performance because it means
that, for much of the time, more than one pair of hosts can communicate at
the same time. Taking the example in Figure 11.2, if PCs 1 and 2 were connected via a switch rather than a hub, communication between these two
hosts would not place any traf fic on the backbone, and other hosts connected on the backbone would be able to communicate at the same time.
In some instances, the use of a hub may be pr
eferable to a switch
because the processing action in the switch does intr oduce an element of
delay, whereas a hub passes the data much more quickly.
The way that a switch identifies the destination of data packets can vary,
depending on the type of switch. Some switches look at the MAC addr ess
(layer 2 switches) whilst others look at the network addr ess for the data
240
Closed Circuit Television
packets. These are known as layer 3 switches, although in practice they are
really routers. The type of switch, ther efore, can affect the way that the
network functions and how certain devices (such as IP cameras) connect
to their hosts.
Similar to the switch is the bridge which, like the switch, intelligently
connects segments together. The main difference is that, whereas a switch
connects multiple segments (like a hub), the bridge only connects two segments. It is used to keep traffic on the two segments separate except when
it is necessary to pass it over . The principle is illustrated in Figur e 11.3.
There are two types of bridge, those that are able to connect dissimilar networks (i.e., ring to bus) and those that can only connect similar networks.
PC2
PC1
PC4
Segment 1
PC5
Segment 2
PC6
PC3
Bridge
Figure 11.3 The bridge only passes traffic across segments when it is necessary
to do so, for example, when PC1 is communicating with PC4. This action
improves the network speed because, at times when the segments are isolated,
hosts on Segment 1 can communicate at the same time as hosts on Segment 2
A router is used to send data from one segment to another or from one
network to another and, like a switch, it r educes network traffic by only
forwarding data when necessary . However, unlike either a switch or a
bridge, a router is an intelligent device that is on a par with a PC. Routers
constantly interrogate the network to find out what other networks and
routers are out there and then store this information in their routing tables.
When a host attempts to send data to a host on another network, the
router will use its r outing table to determine the shortest path thr ough
which to send the data, bearing in mind that the path may include a number of routers.
Another significant difference between a r outer and a switch is that a
router will not pass what are known as broadcast data transmissions (these
will be discussed later in this chapter). This action can sometimes affect the
setting up of network devices such as IP cameras because the connection
protocol used for the initial setup may be a br oadcast, meaning that the
router prevents the engineer from connecting to the device in order to perform set-up. This is why it is sometimes necessary to configure the IP camera
on the network local to the administration PC before deploying it on site.
The engineer may be requested to enter the gateway IP address when setting up IP devices. A gateway device is the one through which all data must
CCTV over networks
241
be routed in order for it to be passed on to another network (either local or
wide area). Consequently, a gateway is normally a r outer, but it does not
have to be. In some cases it may be a PC or a server that is acting as a gateway. Note that, where a device such as an IP camera only communicates
with hosts on the same subnet, it is not necessary to enter a gateway IP
address.
A network server often looks like any other PC tower , although it may
be a little larger. In truth, there is not a lot of difference between the system
architecture of a PC and a server. A server usually has a much higher specification than a normal PC in terms of processing power and is designed
to operate 24 /7 over a period of years. However , what r eally makes a
server perform as such is the operating system that is installed, for example
Microsoft 2003 Server, Redhat Linux, or Sun Solaris.
The function of servers on a network is to provide services to the client
PCs and other devices connected to the network. Such services include
DHCP and DNS (discussed later in this chapter), mail services, web connection, application software, print services, file management, etc.
Network communications
The network communications pr otocol most commonly encounter ed is
TCP/IP, or transmission contr ol protocol/Internet protocol although, as
we shall see later , there are other protocols (or languages if we want to
view it from a human perspective) in use. The IP address most commonly
associated with TCP/IP is IP Version 4 (IPv4), which has been with us for
a considerable number of years now, having been devised in the 1970s.
A protocol is simply a set of rules, so TCP/IP are two sets of rules that
are applied to the function of transmitting data over a LAN/W AN. The
transmission control protocol is responsible for ensuring that data is transmitted across a network accurately, that an acknowledgement is sent by
the recipient host, that data is re-sent if no acknowledgement is received,
that data is checked for errors at the receiving end, and that the data packets
are re-assembled in the correct order. It achieves this by first breaking the
data into packets and then adding a TCPheader to each packet, which is a
bit like placing your letter into a stamped and addressed envelope before
posting it. These data packets with their TCP headers are known as datagrams, and each datagram must have a minimum length of 576 bytes and
a maximum length of 65 536 bytes.
The TCP processes are perhaps best understood by examining the
packet header, which is shown in Figure 11.4. The source and destination
port information ensures that data is sent between the intended processes
in each host. The sequence number contains the necessary data to ensur e
that the packets ar e in the corr ect order for reconstruction. The acknowledgement number is used by the receiving host to acknowledge to the source
host that the packets have been r eceived. If no such acknowledgement is
242
Closed Circuit Television
Source port
Destination port
Sequence number
Acknowledgement number
Offset
Reserved
Flags
Window
Checksum
Urgent pointer
Options
Padding
User data
0
16
32 bits
Figure 11.4a A datagram with its TCP header
Source
port
(16)
Dest
Port
(16)
Seq
No.
(32)
Ack
No.
(32)
Cont
Window
Chksum
(16)
(16)
(16)
Urgent
pointer
(16)
Options/
Padding
(variable32 total)
User
data
Figure 11.4b Transmission sequence for a datagram. The numbers in brackets
indicate the number of bits
received after a certain period, the packet will be er -transmitted. The checksum data is used to check received data for errors. When errors are detected
in a packet, that packet is discarded, the acknowledgement number is not
incremented and the sender will re-transmit the data. The other data in the
packet header, although important, is beyond the scope of this book.
Once TCP has created the datagrams, they are sent to the Internet protocol for final packaging befor e transmission. IP has the task of adding the
destination address, which will be used by all r outers and gateways to
determine the optimum transmission path in terms of length and speed.
IP is connectionless, meaning that ther e is no handshaking or acknowledgement of receipt, which is why it is used in conjunction with TCP in
order to guarantee reliable data transmission. The IP header also contains
information relating to the IP version, the length of the datagram, its own
checksum, a hop count, and other information. The hop count is normally
set at 32, and this determines the maximum number of network r outers
that the datagram may pass before it is discarded as undeliverable. If there
were no hop count, undelivered data would simply continue to pass around
the Internet indefinitely like some sort of r efugee, and by now the web
would be unusable because it would be congested with such data. The
complete datagram containing data plus TCP and IP headers is shown in
Figure 11.5.
An IPv4 address is made up from four 8-bit words known as octets, each
octet being separated by a dot. An 8-bit binary wor d can have up to 256
CCTV over networks
243
IP header
TCP header (see Figure 11.4)
Data packet
Figure 11.5 A complete datagram containing its IP and TCP headers
possible permutations and ther efore, rather than denote the octets in
binary, which would be dif ficult for us to r emember, they are denoted in
their denary equivalent values. For example, the binary octet 1 1 000 000
would equal 192 when converted to decimal, so we show it as 192 in the
IP address. Thus, all octets in an IP address will have a value between
000 and 255, and the complete IPv4 range of possible addr esses will be
between 000.000.000.000 and 255.255.255.255.
IPv4 classes
In the original conception, IPv4 addresses were broken down into classes.
The classes that we are interested in are classes A, B, C and D. The primary
difference between classes is the way that the network and host addresses
are denoted, which is illustrated in Figure 11.6 for classes A to C. Here we
can see that some octets contain the network addr ess, whilst the remainder contain the host address. It is often said that the network address can
nnn.
nnn.
Network
address
nnn.
nnn.
Class A
Host address
nnn.
Network address
nnn.
nnn.
nnn.
nnn.
nnn.
Class B
Host address
nnn.
Network address
nnn.
Class C
Host
address
Figure 11.6 Each IP class uses different octets to denote the network and host
addresses
244
Closed Circuit Television
be likened to a str eet name and the host addr ess can be likened to the
house address. This is illustrated in Figur e 11.7, where two separate networks are connected via a r outer. PC1 and PC4 both have the same host
address (.11) but their IP addresses are different because of their differing
network addresses: PC1 has an address of 194.176.2.11, whilst PC4 has an
address 194.176.3.11.
.11
.12
.11
.12
PC1
PC2
PC4
PC5
192.176.2.0
192.176.3.0
.13
.13
PC3
PC6
Router
Figure 11.7 This illustration depicts two networks that have different addresses;
however, the hosts have the same address range. Note that the network address
uses a zero to denote the host octets, which is why all network addresses end
with a zero. The IP address of PC1 is 192.176.2.11, whilst the IP address of
PC4 is 192.176.3.11
The IP class was originally denoted by the value of the leftmost octet.
These values are shown in Table 11.1, where it will be noticed that value 127
is not included. This is because it is reserved as a test address, known as a
loopback address. Class D IP addresses are reserved for what are known as
multicast, which are single transmissions that can be destined to groups of
hosts on a network. Multicast is normally only employed on research networks; however, it is appearing increasingly on dedicated CCTV networks
where images from a large-scale MUX or matrix switcher need to be accessed
by multiple operators. By employing multicast, network traf fic is greatly
reduced because each image need only be sent once to all recipients. Finally,
the class E IP address range is mainly used for experimental purposes.
Table 11.1 IPv4 class characteristics. The value of the leftmost octet denotes
the IP class. For example, address 192.176.2.2 would be a class C IP address
Address class
Address range
(first octet)
Number of
networks
Number of hosts
A
B
C
D
E
1–126
128–191
192–223
224–247
248–255
127
16 385
2 097 152
–
–
16 777 213
65 533
253
–
–
CCTV over networks
245
Table 11.1 also shows the maximum number of networks and hosts that
are available for each IP class. In many respects class C, which is the most
common, is the most practical because it permits a lot of networks, but the
number of hosts is limited to 253 (in theory this value would be 2 8 256,
but the maximum number of available host addresses is always reduced by
three, because three addresses are reserved for special functions on the network). Class B may appear mor e useful because it permits so many mor e
hosts on each network, yet still provides more networks than any institution
would require. In practice, to place 65 533 hosts on a single network would
result in so many data collisions and subsequent re-sends that the network
would prove to be unworkable. This is because Ethernet operates on a principle known as carrier sense multiple access/collision detection (CSMA/CD),
which in simple terms means that the system allows data collisions to occur
and, when they do, the datagrams are simply re-sent. Before sending a datagram, the NIC checks to see if the network is clear and, if it is, the datagram
is transmitted. Should another host transmit at the same time the datagrams
will collide, and both NICs will back off and wait for a period that is determined by a random timer on each card before attempting to re-send. Therefore one card will re-transmit before the other.
CSMA/CD works well as long as we do not overload the network with
too many (busy) hosts. With a very high rate of data traffic, the number of
collisions will be very high, which will in turn r esult in a lot of r e-sends,
which results in an even greater amount of traffic, resulting in more collisions, and so on.
Reserved addresses
Within the class A, B and C ranges ther e are ranges of IP addresses that
are reserved. These addresses are never used on public networks (i.e., the
Internet) but are intended for private networks. For class A the reserved
address range is between 10.0.0.0 and 10.255.255.255, for class B it is between
172.16.0.0 and 172.31.255.255, and for class C it is between 192.168.0.0 and
192.168.255.255. Another address range that is reserved and will therefore
never appear on the Internet is 169.254.0.1 to 169.254.255.254. These ar e
autoconfiguration addresses that a network device will assign to itself in
the event of it being unable to obtain an address from a DHCP server (see
later in this chapter).
It is important to recognize that it is only essential to adhere to the IPv4
class rules for public networks and that, for a private network, any IPv4
addresses may be employed. At first glance this may appear to be unworkable because we would end up with numer ous hosts from different networks communicating over the Internet, all having the same IPaddress – a
bit like having twenty people called John in the room and someone calling in
through the door, ‘I have a message for John’. However , a private network
will never connect directly to the public network but, rather, it must do so
246
Closed Circuit Television
through some form of interface. The most common interface is network
address translation (NAT), where a network device (usually a r outer or a
proxy server) takes datagrams on a private network that ar e intended for
the Internet and removes the local IP (source) address, replacing it with an
IPv4 class C or , perhaps, IPv6 addr ess. The principle is illustrated in
Figure 11.8. Without NAT we would have run out of IPv4 addresses a long
time ago whereas, by using NAT, we are able to use duplicate IPaddresses
across private networks without conflict.
Sender address
= 10.170.3.2
PC2
PC1
PC4
PC5
NAT devices
Router
PC3
Proxy
server
PC6
Recipient address
= 10.170.3.2
WAN
Network address (defined by
NAT devices) = 193.169.2.2
Figure 11.8 Function of a NAT. By coincidence, PCs 1 and 6 have the same
IP address. However, because they belong to different networks, no conflict
occurs because the NAT devices change the IP address as the data passes
to/from the WAN
Another advantage of employing NAT is that the proxy server or router
makes the original host device invisible to both the Internet and the er cipient. This increases the network security although, sadly , in this modern
world of hackers and virus script writers, a NAT device alone is woefully
inadequate to keep intr uders out, and other means of security must be
applied such as firewalls, anti-virus software, regular Windows updates, etc.
Subnetting
We have already come to see that we may use IPv4 addresses of any class
for private networks; however, we can (and often do) take this a step further.
Another data string known as a subnet mask is included in the datagram.
This data, when overlaid on top of the IP address, is used to identify the
CCTV over networks
247
network and host parts of the addr ess, instead of using the value of the
first octet. So, by using a subnet mask, we can make any IPaddress fit into
any of the classes A to C. For example, by adding a suitable subnet mask
to the class A reserved network address 10.172.2.0, we may identify the
first three octets as network address and the fourth octet as host addr ess,
making it a class C IP address.
One advantage of subnetting is that we ar e able to cr eate a very lar ge
number of new network addresses for private use or, alternatively, we can
increase the number of network addr ess by using the subnet mask to
move the dividing line between network and host (see Figur e 11.6) to the
right by one or two places.
With subnetting it is possible to independently r oute networks. This
makes much more efficient use of the available bandwidth by r educing
network traffic, reducing the size of the routing tables, providing a simple
way of isolating one network fr om another, and enhancing the ability to
provide network security.
The subnet mask is rather like an IPaddress in that it contains four 8-bit
words. It works by assigning a logic value of 1 to each bit in the IPaddress
that is to be designated as network address, and a 0 to each bit that is to be
designated as host addr ess. So, taking the pr evious network addr ess
example 10.172.2.0, the binary equivalent value of this address is 00001010
10101100 00000010 00000000. To make this a class C IP address the subnet
mask would have a value of 11111111 11111111 11111111 00000000. The dotted
decimal equivalent to this would be 255.255.255.000. When overlaid on the
IP address, we have:
class C IP address 00001010 10101100 00000000 00000010 10.172. 2.000
class C subnet mask 11111111 11111111 11111111 00000000 255.255.255.000
So we can see that wherever the subnet mask value is 255, the corresponding part of the IP address is designated as network address, and wherever
the value is 000, the corr esponding part of the IP address is designated as
host address. Notice that because all eight bits of the last octet ar
e
designated as host, ther e will be 256 – 3 253 possible host addr esses
available.
Following this to its conclusion, a class A IP address would have a subnet mask value of 255.000.000.000, and a class B subnet mask has a value
of 255.255.000.000.
Now let’s see how the subnet mask may be used to divide a network
into four equally sized segments. Staying with the example 10.172.2.0, we
know from Table 11.1 that we will have 2097 152 possible network addresses
with 253 hosts on each network. Suppose we now add the following subnet mask to this network address:
IP address 00001010 10101100 00000010 00000000 10.172. 2.000
Subnet mask 11111111 11111111 11111111 11000000 255.255.255.192
248
Closed Circuit Television
This has the effect of releasing two bits from the host address to the network address, increasing the number of networks to 8 388 608. However,
there are now only six bits left for the host addr esses on each network,
meaning that we may now only have 64 – 3 61 hosts on each network.
The question must now be asked, why would we want to do this? When
would an institution ever feel that they would r equire over eight million
network addresses? However, the reason for subnetting is not to increase
the number of available networks, but rather , to deliberately r educe the
number of host addr esses on each network. The r eason why we would
want to do this is because in practice the user may only have a few hosts
connected to any one network, so the r est of the available addr esses are
unusable and are thus wasted. By breaking the original address down into
four, with each subnet only having 61 host addr esses, there is less chance
of having unused addresses, and the other groups of 61 host addresses can
be used on other networks that have been allotted the other subnets.Also,
by reducing the number of hosts on each network, we further r educe the
network traffic and therefore improve the efficiency.
Subnetting is carried out by the IT administrators and may be applied or
modified at any time. The implications for CCTV are that, once cameras and
other network devices have been set up in r elation to any current subnets,
changes to these subnets without modifying the network settings for each
CCTV device will immediately cause them to dr op off line. In practice, the
CCTV engineer who has been called out to discover the reason for the sudden loss of CCTV images is rar ely told about such changes and can ther efore spend a lot of time looking for the cause of the trouble. A wise engineer
would consult the IT department at an early stage in his investigation of the
fault. Alternatively, he may check the curr ent subnet mask against either
historical records or the settings in the cameras/NVRs. Methods for performing this check will be discussed later in this chapter.
Assigning IP addresses
Every time a new host is connected onto a network it will need to have an
IP address assigned to it. As we have seen, this addr ess must be in the
range for the class or subnet that is being used on that particular network,
and it must not already be in use by another host or else ther e will be an
IP conflict on the network. When this occurs, the new host will fail to communicate either with the other hosts, or the router/proxy server that is acting as the gateway to other networks. It is clear that to manually assign IP
addresses would mean a lot of work for the IT administrators. Furthermore,
on a large network, it would be very dif ficult for them to keep a track of
the IP addresses, especially when devices such as laptops ar e constantly
being connected, removed and moved around on the network.
Although it is possible to manually assign an IP address to a host, most
networks have a dynamic host configuration protocol (DHCP) server which
will perform this function automatically. DHCP functions as follows.
CCTV over networks
249
Remembering that a host connects to the network via its NIC, and that
every NIC has a unique MAC address, the moment that a host is connected
to a network, it begins to br oadcast its MAC addr ess. This broadcast is
picked up by the DHCP server (assuming that one is available), which
immediately assigns an IPaddress to that NIC/host. The DHCPserver also
issues a lease time on the IPaddress, for example, eight days (Figure 11.9a).
The host will continue to use that IP address for as long as it remains connected to the network, but after 50% of the lease time has expired, the host
will automatically request a renewal of the lease time, which will normally
be granted. If the DHCP server is not available when the r enewal is
requested, the host will continue to use the IP address until 87.5% of the
lease time is expir ed. It will then attempt to r enew the lease time again,
only now it will accept a lease from any DHCP server that responds.
Figure 11.9a IP addresses and lease expiry dates in DHCP. The right-hand
column shows the MAC address for each host. This particular screen is taken
from Microsoft Server 2000
If the renewal is successful, the process simply repeats when 50% of the
lease time has expired. If the host is unable to obtain a lease r enewal, when
the lease time is expired the host must discontinue using the IPaddress, forcing it to begin the pr ocess of obtaining a new IP address. If, for any reason,
the host cannot contact a DHCP server to obtain a new address, after a short
period of time it will assign itself one of the r eserved autoconfiguration IP
250
Closed Circuit Television
Figure 11.9b Setting the IP address scope (range) in DHCP
addresses in the range 169.254.0.1 to 169.254.255.254. The host will then
attempt to contact a DHCPserver every five minutes in an attempt to obtain
an IP address that is valid for that network.
When a host is power ed down, on r e-connection it will automatically
attempt to renew the lease for the IP address that it had before it was shut
down. If unsuccessful, it will follow the autoconfiguration process described
in the previous paragraph.
In a case where a host is removed from one network and connected to
another, the new DHCP server will assign its own IPaddress. Upon returning to the original network, the first DHCP server will simply assign a
new IP address. It is also not uncommon for a DHCPserver to re-assign all
hosts a new IP address every eight days, although this period may be
changed by the IT administrator. The purpose of this r e-shuffle is so that
the server may determine which IP addresses are no longer being used, so
that they may be returned to the pool for re-issue to another new host. An
address may no longer be in use because, for example, the host may have
been something such as a laptop which was connected by a visitor for a
short period of time. For devices such as PCs and laptops this r e-shuffle
every fourteen days is not a problem, but it can have serious implications
for devices such as CCTV cameras, IP door controllers, IP alarm collector
points and such like, as we shall see in a moment.
So how does the DHCP server know which IP address range to pass out?
This is determined by the network administrators who set up the IPrange
using the DHCP screen in a server application such as Microsoft Server 2003.
A screen shot for the DHCP address pool is shown in Figure 11.9b. DHCP
is also where the network administrators will set up subnets, should they
choose to use them. The DHCPserver can also manage gateway addresses,
assign DNS servers (see later in this chapter), and perform many other tasks.
Manually assigned IP addresses
It was stated earlier that it is possible tomanually assign an IP address to an
NIC/host, but why would we want to go to this trouble? Well, one problem
CCTV over networks
251
with using dynamic IP addresses with security devices is that these IP
addresses may have to be manually entered into the software application
that is used to administer the security devices; in the case of CCTV cameras, this is fine until DHCP changes the IP addresses of the cameras.
When (or per haps, if) this happens, unless the PC/NVR r unning the
administration software is able to r esolve a DNS name for each camera
(see later in this chapter), it is possible that we may simply lose all communications with the IP cameras. Whether or not communications are actually lost will depend on the ability of the camera to pr ovide a host name
for DNS, the availability of a DNS server or , where NetBIOS is used,
whether the network has been set up to r esolve host names. Where there
is a r eal possibility of communication failur e caused by a change of IP
address, the manufacturer will specify the use of static IP addresses for
their equipment. Manually assigned IP addresses are referred to as static
or, sometimes, reserved IP addresses.
To assign a static IP address to something like an IP camera we first of
all need to know which address we are going to assign. There is no point
in assigning an address that is already in use as this will simply r esult in
an IP conflict on the network and a failure of the camera to communicate.
Where the cameras are being connected to an existing LAN (that is, they
do not have their own dedicated network), the only person who can pr ovide the IP address for each camera is the network administrator . The IP
address scope in DHCP may be set up such that a number of addresses are
excluded from the normal DHCP range. These excluded addr esses may
be manually assigned to hosts r equiring static IP addresses such as print
servers, network printers and security devices such as IP cameras, door
controllers, etc. Because these addresses are included in the DHCP scope,
they will still be recognized for DNS purposes.
The next step is to assign the IPaddress to the camera, which is the tricky
part because we cannot send the addr ess via the network using TCP/IP
since the camera does not yet have an IP address – a bit like the chicken
and egg situation. Different manufacturers have different solutions to this
problem. The simplest way to resolve this issue is to let the camera pick up
an address from the DHCP server the moment that it is connected to the
network, and then use pr oprietary software to probe the network for the
cameras. The cameras will then reply using their dynamic IP addresses, at
which point the commissioning engineer can use the software application
to replace the dynamic addresses with static addresses.
The one drawback with this method is that it assumes that ther e is a
DHCP server on the network, which is not always the case. Suppose, for
example, that the IP cameras were to reside on their own private LAN;
there would be no need for a DHCPserver apart to assist in the initial set-up,
which is a somewhat expensive luxury . And yet, without an IP address,
how can the commissioning engineer communicate with the cameras to
assign static IP addresses? The solution is to communicate using a pr otocol other than TCP/IP. Remember that every NIC or other network device
252
Closed Circuit Television
has a unique MAC address, so it is possible to communicate using this in
the first instance in order to assign the IP address, after which it can communicate using TCP/IP. In practice we are only doing manually what DHCP
does automatically.
Address resolution protocol (ARP)
This is the method most commonly employed for communicating using
the MAC address. It is the same type of broadcast that the NIC sends out
when first connected to the network when it is looking for a DHCPserver,
except in this case the broadcast is being sent from the administration PC,
prompted by the engineer . ARP performs a number of important functions on the network, all of which ar e beyond the scope of this text; however, for our purposes we can view the
ARP broadcast as basically a
message that travels over the network which says, ‘Hey , MAC addr ess
00:20:4A:32:F2:3C, your IP address is 192.176.2.13’. Being a broadcast, this
request is hear d by all hosts, but only the host with that unique MAC
address will update its IPaddress. This is illustrated in Figure 11.10 where,
in this instance, camera 1 would respond.
NVR MAC address
= 00:20:4A:D3:A7:44
Camera 1: MAC address
= 00:20:4A:32:F2:3C
Camera 2: MAC address
= 00:20:4A:C2:DB:A4
Switch
Hub
ARP
Broadcast
Router MAC address
= 83:A4:C3:0D:FB:26
Camera 3: MAC address
= 00:20:4A:BB:1C:5C
WAN
PC MAC address
= 83:A4:C3:C8:FF:3B
Router MAC address
= 83:A4:C3:0D:FB:A4
Figure 11.10 Every network device that incorporates an NIC has a unique MAC
(hardware) address. The ARP broadcast sent out by the PC would not pass
beyond the router
Once communications have been established, the engineer may assign the
IP address to the device. Of course, he will need to know the MAC addr ess
in the first place in order to enter it into whatever software application he is
CCTV over networks
253
using to perform the operation, which is why the MAC address is usually
written on the item of equipment.
In practice the engineer will often need to do more than tell the camera
its IP address. Bearing in mind that the engineer is performing the task of
the DHCP server, he may also need to assign the gateway IP address (if
there is a gateway between the camera and the NVR/admin PC) and possibly the subnet mask. Scr een shots for typical utilities used to perform
these functions are shown in Figure 11.11b and c.
One problem that may be encountered when using an application such
as that in Figure 11.11b is that it may not be possible to connect to the ermote
device because of the presence of a router in the network path. Remember
that the initial connection is being made using an ARP broadcast and, as
we saw earlier in this chapter when discussing r outers, a router will not
pass a broadcast. This is illustrated in Figure 11.10, where an ARP request
to connect to 00:20:4A:BB:1C:5C (camera 3) would be blocked by the o
r uter.
The engineer, therefore, would be able to configur e cameras 1 and 2, but
would receive a ‘Failed to connect to host’ message every time he attempted
to connect to camera 3. Note that in this example, camera 1 is connected
via a switch, which means that it is on the same subnet as camera 2, so configuration is possible via a broadcast.
Figure 11.11a TCP/IP set-up screen from Microsoft Windows XP, used to set
up a static IP address on the PC itself
254
Closed Circuit Television
Figure 11.11b A typical proprietary utility for assigning a static IP address to a
device. Such utilities normally use ARP broadcast to communicate with the
remote device
Figure 11.11c Typical Telnet software utility used to set up a remote network
device such as an IP camera, once an IP address has been assigned using ARP
CCTV over networks
255
The problem is that in a lar ge premises, the commissioning engineer
often has no way of knowing the network topology , so one way ar ound
this problem is to preconfigure each security device local to the administration PC by connecting it via a hub or directly to the network port on the
PC using a network crossover cable. Once configured and able to communicate using TCP/IP, the device may then be installed at its r emote location. An alternative is to install the device set-up application software onto
a laptop and configure the device locally, again by connecting via a hub or
crossover cable.
Autoconfiguration
The more sophisticated security devices such as NVRs usually have an
operating system such as Micr osoft Windows, Windows CE or Linux.
Where this is the case, the absence of a DHCP server is not such a problem
when trying to assign a static IP address. Windows-based network devices
are configured so that, after a few minutes of attempting to locate a DHCP
server, if no server is found the device will autoconfigur e and assign itself
a reserved IP address in the range 169.254.0.1 to 169.254.255.254. As long as
the commissioning engineer is prepared to wait for a few minutes following power-up and network connection for this process to take place, then it
should be possible to probe for the device using the manufactur er’s setup
software, bearing in mind that routers should no longer present any problem because communication is now via TCP/IP. Once communication has
been established, it should then be possible to assign the static IP address.
Domain name service (DNS)
One problem with using MAC addr esses and IP addresses is that these
long numerical strings ar e not very easy for humans to r emember. For
example, every website has an IP address, but imagine trying to remember
the names of all of your favourite websites by using their IPv4 addr esses,
and those that use the mor e recent IPv6 addresses would prove impossible because they are 128 bits long!
To make networks user (human) friendly , another server is employed
which looks at the IP addresses and r esolves them into user friendly
domain names. This server, known as the DNS server, may be separate from
the DHCP server, although quite often on smaller networks one server
may be configured to perform both functions. Domain names ar e hierarchical, the top names being the familiar .com, .uk, .ac, etc. The names that
we would be concerned with when setting up network security devices
are lower down the hierarchy.
Every host is given a name, usually by the network administrator
,
although in the case of devices such as IP
cameras, the domain name
256
Closed Circuit Television
normally incorporates some form of alphanumeric ID which the engineer
may be able to modify during commissioning. When the DHCP server
assigns the IP address to a new host, it also tells that host the IPaddress of
the DNS server. Normally the host will then pass on its domain name to
the DNS server, which will add this to its DNS r ecord. Now, when a person working on another host wishes to communicate with that particular
host (normally through some form of browser such as a web browser) that
host will, in the first instance, only know the domain name that the person
has typed in. Therefore it will put out a DNS request, and the DNS server
will resolve (convert) the domain name into the IP address of the host to
which they wish to connect. This means that users ar e able to work with
user-friendly domain names rather than IP addresses.
Looking again at Figure 11.9a, the DNS names for the six hosts in that
screen shot are listed in the second column.
There is another (older) name resolution service that is still used, called
Windows Internet naming service (WINS). This service was designed to
work on simpler networks (like the majority of domestic networks) where
there is no DHCP or DNS server. These networks generally use a protocol
called NetBIOS (also known as NetBEUI), which is simple to set up and
administer; however, it is only suitable for small networks because it cannot be routed and, in terms of bandwidth, is very inefficient. Where CCTV
cameras are running on their own private network it is possible that they
are using NetBIOS, in which case ther e is every probability that WINS is
employed for domain name resolution.
Note that WINS cannot be employed on any network that is intended to
connect to the Internet because TCP/IPrelies on DNS for address resolution.
On the other hand, both WINS and DNS can co-exist on the one network.
Ports
For TCP/IP, a port is an address which defines the association between the
data being transmitted and the applications on the sour ce and recipient
PCs for which the data is intended. For communication to take place, two
ports are required, one at the sour ce device and one at the r ecipient. The
source port identifies the application that sent the data, whilst the recipient
port identifies the application for which it is intended. Each port is assigned
a number, some of which are well known because they are associated with
common applications. For example, web servers use the pr otocol HTTP
which has been assigned port 80, some email servers use a pr
otocol
known as POP3 which uses port number 110, and FTP servers communicate on port 21.
IP cameras must communicate via a port, the port number being dependent on how the images ar e being extracted. If a standar d web browser is
used, then communication will be via port 80, whereas an NVR running its
own proprietary software would have a port assigned by the manufacturer.
CCTV over networks
257
A PC that is connected to the Internet may have any number of ports
opened, depending on the applications that ar e installed. It is thr ough
these open ports that the all too familiar vir uses, trojans, worms, adware,
etc., come in. Those who perpetrate such annoying and often destr uctive
software can set up their PC to ‘snif f’ hosts that ar e connected to the
Internet to see if there are any open ports and then, upon finding a victim,
the offending script is downloaded. Firewalls are used to combat this activity and they function by closing down all ports apart fr om those that are
absolutely needed (port 80, for example). The problem for network CCTV
is that, wher e there is a fir ewall between the cameras and the NVR or
other image processing equipment, it will not be possible for the cameras
to connect because their dedicated port will have been blocked. Ther e is
only one way to r esolve this pr oblem, and that is to have the network
administrator open the r equired port on the fir ewall. However, in most
cases they are reluctant to do this because it compr omises their network
security.
The problem of fir ewalls is unlikely to occur when all of the CCTV
equipment is operating on one local network because ther e will be no
need to connect via the Internet. However, more and more large organizations are wanting to network their CCTV systems nationally or even globally. Of course, it should always be possible to achieve this by using a web
browser because port 80 is always open; however , the image quality is
often not very good, and the frame refresh rate can be slow at times when
the Internet is busy. In truth, the Internet was never intended to pr ovide
every large organization with a high-quality global CCTV system and such
systems would operate far more effectively if private network connections
were employed, but usually the cost of this proves to be prohibitive.
Other network protocols
TCP and IP are not the only protocols that may be used on a LAN or WAN,
although in practice they ar e the most common. We have already looked
at an example of another protocol in NetBIOS. Here we will briefly look at
a number of other protocols that are used on modern networks.
User datagram protocol (UDP) is what is known as a connectionless protocol. When two hosts communicate using TCP , a session is opened between
them, which means that a continuous connection is maintained until the end
of the data transmission. By contrast, when UDPis used no such session is
created. The datagrams are simply sent across the network to the recipient
host where they may arrive out of order, and are not subject to any errorchecking or correction.
On the face of it this all may sound very risky; however, it does depend
on the importance of the data being sent and whether or not the application itself will initiate a re-send if no reply is received. For example, DNS
uses UDP where, if the datagrams are lost, DNS will eventually try again.
258
Closed Circuit Television
The advantages of UDP are reduced data and increased speed, which are
a direct result of omitting the session set-up and error detection.
Like TCP, UDP uses IP to send and receive the datagrams.
Some IP cameras employ UDP because of its clear reduction in network
overhead, which is consider ed a worthwhile tradeof f for the occasional
lost datagram which, in reality, represents only a very small portion of one
TV frame. Where such cameras could begin to fall down is during periods
of high network traffic, where collisions will be more prevalent and therefore more datagrams will be lost. But then it can be ar gued that cameras
using TCP/IP will struggle to maintain image integrity during such periods,
and their incr eased network over head will actually exacerbate the network traffic problem.
File transfer pr otocol (FTP) is a part of the TCP protocol suite and is
used as a means of transferring files between two hosts. Originally command line driven (i.e., using the DOS scr een), user interface softwar e
applications now exist to simplify the use of FTP. To open an FTP session
the user must know the IP address or domain name of the host with
which they wish to connect and, depending on the host setup, they may
have to enter a passwor d before they ar e permitted to download or
upload files.
FTP connections are made via port 21 although, once a connection has
been established, a second TCP connection is set up on port 20 over which
the actual data transfer takes place. The FTP connection is used for command and control data during the session.
Telnet is also a part of the TCP protocol suite and is a virtual terminal
service which permits a user to connect two hosts and take full contr ol
over the remote host. All mouse and keyboard activity is passed dir ectly
to the remote host so that applications may be opened and r un remotely.
This may sound very grand, but one drawback is the latency in the system, especially when the network is busy, where keystrokes may take several tens of seconds to propagate through. Another frustration when using
Telnet is its tendency to time out if the session is interr upted for too long
when the network is busy. There are other proprietary software packages
that perform the same function as T elnet, but with far less latency , and
offering a far more reliable connection.
Telnet is, however, used by a number of security systems network device
manufacturers to perform device set-up. Using a softwar e application
provided by the manufacturer, a Telnet session is established between the
administration PC and the device to enable such things as subnet mask
and gateway IP address to be programmed. An example of this is shown
in Figure 11.11c.
Hypertext transfer pr otocol (HTTP) is well known to anyone who has
used a web browser, even if they do not know exactly what it is. A typical
web page comprises a mixtur e of text, graphics and links to other documents or websites. HTTP is a pr otocol that manages communications
between the web browser and the web server and ensures that a requested
folder, link, etc., is opened.
CCTV over networks
259
HTTP forms part of the universal resource locator (URL), for example,
http://www.sitename.com. When considering CCTV network equipment, the URL may look something like http:/ /devicename:urlpath.
Sometimes the URL contains https rather than http. The ‘s’ indicates that
an encrypted communication channel must be set up between two hosts,
which usually results in a prompt for a name and password. In the case of
devices such as cameras and NVRs, the password is generated and established between the hosts automatically during initial network connection.
This explains why devices that have been removed from one site and installed on a completely different site fail to communicate despite all attempts
by the engineer during the commissioning stage. Most devices have some
means of restoring factory default conditions so that a new password can
be established between the new hosts.
Simple mail transfer protocol (SMTP) is responsible specifically for sending and receiving email between hosts. As it does not play any direct role
in network CCTV, we will not discuss this any further.
IPv6
At the time of development of TCP/IPin the early part of the 1980s, no one
could have predicted the unprecedented growth of the Internet and the
number of network hosts that would be in existence today. Whereas the original concept was to have a large number of computers communicating on
the network, today we ar e looking towards having a multitude of other
devices such as PDAs, mobile phones, wireless network devices of all types,
building management equipment, and domestic appliances such as refrigerators and heating/air -conditioning, to name just a few . This explosion
in demand for network access leaves us with a very serious shortage of
IPv4 addresses which, at best, can only provide 232 4 billion (approx.) IP
addresses.
The issue is not just the shortage of IP
addresses. As network sizes
increase across the globe, the IPv4 system is becoming less efficient due to
the large size of the routing tables required to maintain information about
the network.
To satisfy the future demand for IPaddresses, IPv6 has been developed.
Unlike IPv4 addresses which comprise 4 8 bits 32 bits per addr ess,
IPv6 addresses comprise 8 16 bits 128 bits per addr ess. This makes
provision for 2 128 over 340 billion, billion, billion, billion dif ferent IP
addresses. Working on the basis that there are approximately 6 billion people living on the planet, then each person could have around 56 billion, billion, billion IP addresses to use for their own personal items of network
equipment, so we should have enough to be getting on with for a while!
An IPv6 address looks nothing like the familiar IPv4 addresses. Having
eight sets of four 16-bit (hexadecimal) digits separated by colons, an IPv6
address looks like:
6AF4:0230:005A:0000:0000:0000:2D7A:437C
260
Closed Circuit Television
To simplify both the writing down and entering of these addr esses, some
rules have been applied that facilitate a possible shortening of the address.
First of all, it is not necessary to show leading zer os in each group, so the
address above could be written:
6AF4:230:5A:0:0:0:2D7A:437C
Secondly, a group or chain of groups that contain all zeros can be replaced
by a double colon. So now the above address can be written:
6AF4:230:5A::2D7A:437C
There is one limitation to this second r ule, which is that the addr ess may
only contain one double colon. Therefore, in an address that contains two
or more separate groups of zero, only one group may be removed.
The migration from IPv4 to IPv6 is not simple and will take a number
of years to implement. During the transition period, har dware must be
capable of functioning on either network to facilitate the intr oduction of
IPv6-ready equipment onto existing IPv4 networks.
Network diagnostics
When hosts fail to communicate on a network, there can be any number of
causes, and it is up to the engineer to logically work thr ough each possibility until the precise cause is located. However, logical fault diagnosis is
only possible when the correct tests can be applied, but unfortunately the
multimeter and oscilloscope ar e of little use in r esolving communication
problems on a network. There are network diagnostic tools available, but
these are more likely to be used by dedicated network engineers rather
than security systems engineers who, when it comes to network diagnostics,
are mainly concerned with first line fault diagnosis. The sort of tools equired
r
for this type of fault diagnosis are found in a PC rather than a toolbox.
To begin with, don’t forget to check the obvious. For example, if an IP
camera is failing to connect to its host PC, NVR or MUX, check such things
like: Has an IP address has been assigned to the camera? Is the IP address
that is assigned to the camera being used by another device? Is the patch
cable between the camera and RJ 45 socket corr ectly wired and not a
crossover cable? Is the RJ 45 socket actually connected to the network?
As well as making a visual check, these points can be checked using the
PING (packet internet groper) utility. PING is normally executed from the
command line (i.e., in the DOS scr een) and it provides a simple means of
testing if it is possible to reach a particular host on the network and obtain
a reply from that host. When PING is executed, the PC simply sends out a
message on the network which says, ‘host so-and-so, are you there?’. If the
message reaches the device, it will attempt to send back a r eply which is
then displayed on the PC. If a r eply is received, it confirms that all har dware and connections between the PC and the host in question ar e good,
CCTV over networks
261
that the host is able to communicate acr oss the network, and that the IP
address is valid on that network.
To execute the PING command you must first open the DOS scr een,
which is usually done from the Start button on the bottom toolbar (assuming the operating system is Micr osoft Windows™ 95 or later). Fr om the
Start menu, select the Run option and, for Windows 2000 or Windows XP
type ‘cmd’ (for Windows 95/98 type ‘command’) – the DOS screen should
open with the flashing cursor at the CMD pr ompt. At this prompt, type
‘ping’ followed by either the IPaddress or the hostname for the device you
are looking for – in this case the camera in question. A typical ping command will look like:
ping 192.176.5.10
A few seconds after pressing the Enter key, a response will be obtained. If
a device is present, the response on the screen will look something like:
pinging 192.176.5.10 with 32 bytes of data:
Reply from 192.176.5.10:
Reply from 192.176.5.10:
Reply from 192.176.5.10:
Reply from 192.176.5.10:
bytes 32
bytes 32
bytes 32
bytes 32
time 10 ms
time 10 ms
time 10 ms
time 10 ms
TTL 128
TTL 128
TTL 128
TTL 128
If no device is pr esent, the r esponse ‘Request timed out’ will appear .
Bearing in mind that, among other things, you ar e trying to ascertain
whether or not there are two devices sharing the same IP address, disconnect the CCTV camera in question and re-execute the Ping command. If a
response is now obtained you will know that you have been given an IP
address that is alr eady in use and the system administrator will need to
provide you with an alternative. Another consideration when no r eply is
received is that the network administrator has suppressed ping commands
on his network. Either ask him, or try pinging a known working host.
Another useful tool for finding out information about a PC or PC-based
DVR that uses a Microsoft Windows operating system is the ipconfig utility. There are a number of variations, called switches, to the basic command; however, simply typing ‘ipconfig’ in the command line will reveal
the IP address of the PC, its subnet mask and the default gateway IP
address (if there is a default gateway). A typical response is illustrated in
Figure 11.12a.
The ‘/all’ switch pr ovides much mor e information about the PC on
which the command is executed. As shown in Figure 11.12b, in addition to
the information derived fr om ‘ipconfig’, we can now see the physical
(MAC) address, DHCP configuration, the IP addresses of the DHCP and
DNS servers and the expiry date of the DHCP assigned IP address. This
command is executed by typing ipconfig/all in the command line.
Two other switches that may be used with ‘ipconfig’ are ‘/release’ and
‘/renew’. These can be very helpful if you have just changed the IP
262
Closed Circuit Television
Figure 11.12a Typical response to ipconfig on a PC
Figure 11.12b Typical response to ipconfig/all
address configuration of a PC fr om static to dynamic (see Figur e 11.11a).
Having assigned the dynamic addr ess, the PC may not see the new
address until it has been r e-booted, which may take some time. T yping
ipconfig/release in the command line forces the PC to release any existing IP
address. If you then follow this by typing ipconfig/renew, the PC will look
for the new (dynamic) IP address.
The traceroute utility may be used to discover the path fr om a PC to
another host on the network, either local or acr oss the Internet. T yping
tracert at the command line produces a list of every r outer (including the
IP address of each router) through which TCP/IP packets pass in order to
reach a particular host. It also shows the time (in milliseconds) that it takes
for the packets to pass through each router. An example is shown in Figure
11.13, where the path to the web server for www .pactechinfo.co.uk is
shown. The trace reveals that the packets pass through twelve routers after
hop number one which, as we can see from the information in Figure 11.12b,
is not a router IP address but the address of the local DNS server.
CCTV over networks
263
Figure 11.13 Typical response to tracert command. In this case the path from a
PC to a website named www.pactechinfo.co.uk has been traced
Earlier in this chapter we saw how Telnet may be used to connect two
hosts in order to control one host from the other. However, you may also
use this utility to test the connectivity of any port on the network, which
can prove useful for checking if a host is capable of responding via a particular port. By typing ‘telnet’ in the START – RUN menu in Windows XP,
a Telnet session may be opened. Typing ‘help’ in the command line reveals
a list of all the functions within this utility , and how to execute them. T o
execute a port test on a host, type ‘open hostname/ipaddress’ [port number].
For example, typing ‘open 192.176.2.2 [23]’ will create a session on the host
having that IP address, on port 23. Clearly, if the host name or IP address
of a CCTV camera plus the port number used by that camera is known,
then it will be possible to check to see if the camera is communicating on
that port. It should be pointed out that, if the connection is successful,
there will be no notification on the scr een; however; if the connection is
not successful, then a ‘Connection failed’ message will be displayed.
A failure to communicate with the IP cameras could, of course, be due
to something as simple as a defective or incorr ectly set network interface
card on the administration PC. Wher e the PC is connected to a main network, it is easy to verify if this is the case by simply trying to connect to
another host, such as a local file server or a well known and, ther
efore, reliable Internet site. However , where the CCTV system is operating on its
own dedicated network, with no connection to other LANs/WANs, another
means must be found to test NIC functionality and network set-up. One
such test is the softwar e loopback test, which is performed by entering, at
the command line, the r eserved IP address 127.0.0.1. This test does not
actually produce any network traf fic; however, the r esponse (shown in
Figure 11.14) does confirm that the NIC is functioning.
264
Closed Circuit Television
Figure 11.14 Pinging 127.0.0.1 produces a similar response to any other ping;
however, the responses are coming from the host’s own network interface card
CCTV over a network
In Chapter 2 we consider ed network cabling, examining the bandwidths
and bit rates that are available from the different cable categories and the
factors that may r educe those bandwidths. In Chapter 5 we looked at
video compression techniques, and saw how the file size of each pictur e
frame may be reduced to improve the efficiency of transmission over networks, as well as reduce the required storage capacity of digital recording
equipment. We also saw how excessive compr ession may r esult in poor
image quality due to inabilities of the decompression codec to recover all of
the lost information. In Chapter 8 we looked at digital recording and the different options available for storage, retrieval, downloading and archiving
of video information. Having now looked at the basics of networking and
seen how devices communicate over a network, we are able to bring all of
these factors together and consider the practicalities of using a network to
transmit digital video signals.
The implications of utilizing networks for CCTV (and indeed security
alarm and access control systems) are enormous. Where structured cabling
is already installed in a premises, or where it is possible to utilize the existing network for CCTV , installation is gr eatly simplified because of the
reduced cable installation. Furthermor e, by utilizing the network, we
introduce the possibility of remote administration of the system from any
point on the globe. However we cannot, and must not, ignore the practical
considerations.
First of all ther e is the issue of bandwidth. Even a compr essed video
signal requires a r elatively large amount of network bandwidth, and
the more cameras that we place on the network, the gr eater will be the
CCTV over networks
265
bandwidth required. However, it is not always simply a case of knowing
that there is sufficient bandwidth on a network to handle the video traffic.
Where the cameras ar e operating on an existing LAN that alr eady has a
large number of network hosts, we have to consider the amount of available bandwidth. Where a network is already operating almost to capacity,
the addition of just one camera may well take it over the top, r esulting in
intermittent loss of connection of some devices, slow network functionality, picture freezes on CCTV cameras, etc. In cases such as these, it may not
even be the camera that goes off line, but other devices such as PCs, printers, scanners and, mor e important, servers dr opping. In such cases it is
often the CCTV engineer who is blamed for the problem because, when all
is said and done, everything worked until the cameras wer e put onto the
network!
As a general r ule, when considering using IP cameras on an existing
network, the specifier or engineer must first ascertain the limitations of
the existing network and ask questions like: How much available bandwidth is ther e? What is the available bandwidth at peak times such as
early morning and early evening when everyone is either logging onto the
network, or is logging off? (Remember, if there is an IP access control system on the network, these times will also represent the peak traffic periods
for the access control because everyone is swiping in or out of the building.) How many cameras will be r equired? What image quality does the
user require? What frame update rate does the user r equire from each
camera? What is the bandwidth requirement of each camera for this image
quality and frame rate? And finally, how well has the network performed
to date? Before connecting your CCTV system to an existing network, it is
prudent to find out if ther e has been a history of r ecurrent communications problems, otherwise you may find yourself attempting to r esolve
problems that have nothing at all to do with the CCTV installation.
For larger IP CCTV systems where it is felt that an existing network would
be unable to support the system, it is very common to have a dedicated
network installed solely for the use of the CCTV system (and possibly
other security systems such as intr uder and access contr ol). In these circumstances it is usually necessary to install new cables (unless there is sufficient spare capacity in the network cable bundles) but at least the CCTV
system installer retains complete control of the CCTV IPnetwork. He does
not have to worry about the IT department changing subnet masks,adding
firewalls between networks, altering the network security such that he is
no longer able to make changes to the administration PC, or adding
another bandwidth-hungry application such as voiceover IP.
And, of course, having a private network for the CCTV system does not
mean that the images cannot be viewed fr om other parts of the network,
or even the world. The CCTV network may still be connected to another
LAN via a bridge, switch or r outer and, using suitable softwar e and a
password, users are still able not only to view CCTV images but interact with
the system, moving cameras, stream video from NVRs, etc.
266
Closed Circuit Television
Network CCTV example
A typical network CCTV system is illustrated in Figur e 11.15. In this
example the CCTV system enjoys the luxury of having its own private
CAT 6 LAN and, with only seven cameras on the network, pr ovided that
the network is installed to CAT 6 standards, the system will operate very
efficiently. The NVR will constantly record the images streaming from the
seven cameras whilst at the same time forward those images to the administration PC, which is also serving as the main viewing monitor . The only
Servers (incl.
DHCP)
Customer’s main
network
Administration PC
NVR
CAT 6
Hub
Switch
CAT 6
Hub
Dedicated CCTV
network
To 100 other
hosts
PTZ Dome
Router (incl.
firewall & NAT)
Proxy server (incl.
firewall & NAT)
WAN
CAT 5e
Hub
Customer's
remote LAN
To 200 other
hosts
Figure 11.15 Example of a CCTV network, connected to an existing network via
a network switch
CCTV over networks
267
other network traffic on the LAN will be the telemetry between the administration PC and the three PTZ dome cameras.
A network video recorder functions like any other network device in
that it will have a MAC address, it usually has a static IP address assigned
to it, it connects to the network via a NIC, and it uses either TCP or UDP
to perform data transport. For the majority of the time an NVR is simply
recording data (video images) from the IP cameras on the network, although
it may also be pr oviding a stream of multiplexed images to one or mor e
viewing monitors, depending on the operational r equirement for the
CCTV system.
The network traffic in our example may be r educed if cameras having
in-built motion detection ar e used. In this case, wher e there is no movement within the field of view for a pr edetermined time, the camera will
cease to send its images across the Web. However, this does have one disadvantage over using motion detection in the NVR in that, when an alarm
event occurs, the NVR may be unable to retain the video information prior
to the alarm event because it may not have been transmitted by the camera.
One of the prime advantages of network CCTV is that the images may
be viewed from anywhere in the world, provided that a suitable network
link is available. There are a number of methods by which the images may
be accessed, but not all of these ar e practical in all situations. Looking
again at the example in Figure 11.15, the user would be able to view any of
the camera images from any of the PCs on the local network using a standard web browser, provided that the IP address of each camera was known.
Alternatively, with the appropriate manufacturer’s software installed on a
PC, the user would not only be able to view images but may actually control the PTZ domes. Another alternative would be to access the NVR and
stream the most r ecently recorded images for any camera, thus in ef fect
viewing almost live images without having to know the IP addresses of
every camera.
Possible drawbacks of streaming images across LANs (i.e., via the network switch in the example in Figur e 11.15) are, firstly, that the images
may appear at the viewing monitor in a mor e highly compr essed state
than if they were viewed directly at the NVR. Secondly, if there is a lot of
network traffic on the user’s main network, the frame update rate may be
much lower because of the time that it takes to receive all of the data packets
required to construct each frame.
In Figure 11.15, the main network includes a DHCP server. This can
simplify the initial setting up of the IP cameras by assigning a dynamic IP
address the moment they ar e connected to the CCTV network. Once a
dynamic IP address has been assigned, it is a simple task to manually assign
a static IP address, because the camera is able to communicate using TCP/IP.
In the absence of a DHCP server the camera will be unable to obtain an IP
address, and the only way to assign one manually would be to use the
ARP command, either directly or via a software application like that illustrated in Figure 11.11b.
268
Closed Circuit Television
Note that if a router had been used in place of the switch, the cameras
may not be able to obtain an IP address from the DHCP server because,
unless the router has been set up to do so, DHCP will not function across
a router. Remember how DHCP operates. The moment that an IP camera
is connected to the CCTV network, it will br oadcast its MAC addr ess in
the hope that a DHCP server picks up the broadcast and, in turn, assigns
a dynamic IP address. However, a router will block the br oadcast, so the
DHCP server will never r eceive it. It is possible to configur e a router to
pass the MAC broadcast in order for DHCP to operate, but if the network
administrators have not done this, the IP cameras will have to have their
static IP addresses assigned using the ARP command, even though there
is a DHCP server on a near-by network.
Another way of overcoming the problem of DHCP being blocked by a
router is to install a relay agent on the CCTV network. This device uses
TCP/IP to communicate with the DHCP server and requests IP addresses
and leases on behalf of the hosts on its network. So, the DHCPserver is still
issuing the IP addresses and leases, but it is doing so acr oss the router by
sending them not directly to the host, but to the relay agent using TCP/IP.
The example in Figure 11.15 includes a connection acr oss the WAN to
another LAN owned and administered by the same user. NAT devices are
used to connect the two LANs and, as would be expected, these devices
include firewalls. It will also be noted that the r emote LAN is CA T 5e,
which will be much slower than the CAT 6 local network, not only because
of the lower specification, but also because of the number of hosts ersiding
on the CAT 5e LAN.
Trying to view CCTV images on the r emote LAN could pr ove to be
problematic, although in practice it is usually possible when everything is
set up correctly. First of all there is the issue of connecting to the CCTV network where, as when viewing images fr om the local LAN, the simplest
method is to use a web br owser to connect to the NVR. However , there
will be a lot more latency in the system as a result of the action of the router
and the proxy server, plus the narrower bandwidth and lower speed of the
CAT 5e LAN.
Connecting to the NVR fr om the r emote LAN using manufactur er’s
proprietary software will almost certainly be out of the question because
the port will be blocked by both fir ewalls. Loss of this option will mean
that the user will be unable to operate the PTZ domes fr om the remote
LAN. Only by having the IT administrator open the ports used by the
NVR on both fir ewalls will this method of communication be possible
and, for good reason, the IT administrator is usually reluctant to do this as
it will make the network more vulnerable to attack.
Integrating analogue cameras
The example in Figure 11.15 makes the assumption that every camera is a
network camera which, in the case of a new installation, may be the case.
CCTV over networks
269
However, for the foreseeable future there will be a requirement to maintain
existing analogue cameras and integrate them into network CCTV systems.
One possibility is to employ an NVR that accommodates analogue inputs
as well as network cameras, which is all right as long as the analogue inputs
terminate in the same r oom as the NVR is to be located, and as long as
none of the camera locations requires telemetry.
An alternative solution is to use a device that digitizes and compresses
the analogue video signal prior to transmitting it over a LAN or W AN.
The principle is illustrated in Figure 11.16, where a four-way video server
is depicted. Servers such as this simplify the task of integrating analogue
and network systems by enabling the analogue cameras to be connected to
the network at a location much closer to the camera. This may well mean
a considerable reduction in the length of the co-axial cable, which in turn
should provide improved image quality.
Video servers like that shown in Figur e 11.16 are available in different
forms, ranging from a single analogue input device that may actually be
located inside the camera housing, to lar ge rack mount arrays. The rack
mount array would more than likely be located in the control room, which
removes the advantage of co-axial cable length r eduction. Nevertheless,
the existing analogue cameras ar e still being made available on the network, and may be accessed fr om a much wider ar ea. A typical rack
employs separate four-way servers, making the system completely expandable. Each individual server has its own RJ 45 network connector , which
means that a network switch would also be r equired in order to connect
each of these to the LAN.
Figure 11.16 illustrates another advantage of employing video servers,
which is that PTZ telemetry signals may be passed via the Ethernet, thr
ough
the server and on to any PTZ or dome cameras via their co-axial connections.
In order to achieve this, a server with telemetry capability must be used, and
it must have appropriate drivers installed in order for it to produce the correct protocol. In most cases drivers forthe majority of the popular PT/dome
cameras are available, and the required driver may be downloaded fr om
the administration PC to the server.
Administration
PC
Analogue
cameras
Video server
LAN/WAN
NVR
Figure 11.16 Four-way video server converting four separate analogue video
signals into a TCP/IP network signal
270
Closed Circuit Television
Summary
In this chapter we have examined the basic operating principles of Ethernet
networks. By no means can we hope to cover the detail of this subject in
such a small chapter, bearing in mind that most networking textbooks are
at least 5 cm thick! However, what we have done is to focus on the issues
that are particularly relevant to security systems engineers who are having
to install and set up network security devices such as IP cameras, alarm
collector points, access control door controllers, and the like.
Network CCTV is by no means going to completely r eplace analogue
systems in the short term, and there is a firm role for both technologies for
the foreseeable future. Nevertheless, it is perhaps as short-sighted for security systems engineers to ignor e network technologies as it would have
been for shorthand typists to ignore the advent of the word processor. As
network performance and technologies continue to improve, security systems in general will continue to take more advantage of networks. Network
CCTV images will continue to impr ove, frame refresh rates will become
faster, even at peak network traf fic periods, and mor e reliable methods
will be derived for extracting and ar chiving video evidence over a network. For all of these reasons it is imperative that security systems engineers
have a grasp of network technologies, just as they have had to acquire a
grasp of PC hardware and Microsoft Windows operating systems.
12 Ancillary equipment
Up to this point we have examined each of the primary components in a
CCTV system, looking at their operating principles, their function, variations in technology, setting up and adjustment methods, and have identified typical fault symptoms and causes. In this chapter we shall deal with
those components in a CCTV system which, although at times may appear
mundane, nevertheless play an essential role.
Camera mountings
CCTV cameras are both expensive and delicate and in many cases need to
be protected from the elements, vandals, thieves, or a combination of
these threats. A range of protective mountings are available and the choice
for any given application may be determined by such factors as thedegree
of protection required, size of camera, aesthetic r equirements, mounting
location and whether it is to be a covert or overt installation.
The simplest form of camera mount is a wall bracket and, although not
offering protection against any threat, in indoor locations where the threat
is minimal, brackets offer an effective and inexpensive means of fixing a
camera. All cameras have a standard mounting, which somewhat simplifies the choice of bracket, and the only points left to consider ar
e the length
and weight of the camera. For obvious reasons weight is an important factor, but it is not only the load-bearing ability of the bracket that must be
considered – the surface to which the bracket is to be fixed must also be taken
into consideration. For example, where an internal camera is to be fixed to
a thin plasterboard surface, it may be advisable to employ a toggle or gravity bolt fixing; however, not all internal brackets have mounting holes large
enough to accommodate such fixing devices. The length of the camera
must be taken into consideration when choosing a bracket to ensur e that
there is enough clearance behind not only to allow fr ee movement when
setting the camera angle, but also to pr ovide sufficient room to allow for
the cabling.
Some brackets, both internal and external, incorporate a cable management system where the video and power cables pass thr ough the bracket.
This is an important featur e where cables are prone to vandalism or malicious attack, and this feature also improves the aesthetics of the installation.
When selecting a suitable housing for a camera the engineer should
consider such points as internal or external use, the size of camera, the
method of fitting and r emoving the cover(s), wash/wipe and heater
272
Closed Circuit Television
requirements, overt or covert design, the degr ee of risk of attack (i.e.,
should an anti-vandal type be used?) and whether ther e any ergonomic
requirements. The traditional housing design incorporates an elongated
shell with an option for P/T mounting or a wall-mounting bracket, perhaps incorporating integral cable management. The metal casing should
be protected from corrosion by galvanization, anodizing or coating with a
weatherproof paint. Ther e will be a footplate inside the housing onto
which the camera will be fixed and, in the case of housings intended for
mains voltage operated cameras, power terminations at the r ear of the
assembly. When used in cooler climates, external units should have a thermostatically controlled heater to prevent misting of the front glass and the
lens. These heaters usually take the form of a 15–20W wirewound resistor
connected in series with a thermistor to the mains supply voltage. For
anti-vandal versions the front glass will be shatterproof.
The shell will have some means of opening to allow camera installation
and future maintenance and for this there are a number of design trends.
For larger housings, the top cover is generally held in place by clips and
lifts off for access, allowing ample r oom to work on lar ge camera/lens
assemblies. One problem that may be encountered is what to do with the
housing top once it has been r emoved. If working from a hydraulic platform there should be r oom to place it on the floor , but if the engineer is
only required to perform a simple operation which may be accessed via a
ladder (which, in the UK, is often not permitted under the regulations for
working at heights), then grappling with a lar ge housing cover may
become something of an issue. In some cases the manufactur er provides
for this eventuality by fixing a steel wire between the cover and the housing bottom, allowing the top to hang from the assembly when removed.
Smaller housing designs have a variety of access methods but some of
these may pr ove problematic for certain installations. For example, in
some cases the camera is fixed to a footplate which is then inserted into the
rear of the housing. In theory the engineer should be able to swing the
housing around to give room for this operation, but this is not always possible. Furthermore, once installed, it is very dif ficult for the engineer to
then perform adjustments to the lens and back focus. Other designs have
the top cover slide on and of f which, again, is not always a simple oper ation in some situations. For most situations the best form of housing cover
is where the top is either hinged, or is lifted off and allowed to hang from
a secure cable. For higher security situations wher e the camera might be
accessed by unauthorized persons, securing the cover using anti-vandal
screws is a viable option.
Another design feature is the pr ovision of a sun visor . Direct sunlight
on the lens may cause the iris to close down, r esulting in a very dark
image, or it can cause multiple r eflections within the lens that pr oduce
undesirable light str eaks or halo ef fects on the pictur e. Similarly, when
sunlight falls dir ectly onto the fr ont glass, unwanted haloes or bright
patches may appear on the picture. To eliminate these problems, external
Ancillary equipment
273
housings should incorporate some form of sun visor . This is generally
formed into the top cover and provides shielding to both the top and sides
of the front glass.
Another factor that may determine the choice of external housing is
whether or not a wiper is r equired, as not all housings ar e designed to
accommodate this. The wiper and associated motor are often specified by
the manufacturer as an optional item, allowing the installer to only use
these where it is felt necessary and ther efore reduce the installation cost.
When considering these costs it is important to r emember that it is not
only the wiper har dware that must be included but also the cost of pr oviding telemetry to control the motor. In the case of a fully functional camera this is not a pr oblem because telemetry will alr eady be pr esent;
however, for fixed cameras where there is no need for telemetry the added
cost of a wiper plus telemetry will raise the cost of that camera installation
considerably, begging the question: does that particular camera r
eally
need a wiper facility?
A wash facility may be installed at camera locations wher e there is
excessive dust or dirt, but car eful consideration must be given to this. It
will be necessary to accommodate a reservoir somewhere close to the camera in a place where it will not be tampered with. In the example shown in
Figure 12.1 the r eservoir is housed in the base of the camera column, but
this will not be possible on a wall-mounted camera installation. W
ash
reservoirs designed for location at the camera head ar e available, but in
many cases it may pr ove very difficult to replenish them. In the case of
wall-mounted installations, some installers have attempted to locate the
reservoir on the roof of the building; however, because it is above the outlet nozzle the water is prone to siphon away. In the majority of cases a wash
facility is not essential because a well-maintained wiper will be able to cope
with light deposits of dust and dirt, assisted often by the action of the rain.
Dome housings have rapidly overtaken the mor e traditional fixed and
pan/tilt (P/T) housings like that in Figur e 12.1, partly because they ar e
much more pleasing aesthetically, but also because the fixed versions offer
easy 360° camera positioning, and the powered versions offer much more
rapid 360° panning movement than their traditional counterparts (this
was discussed in Chapter 10). Another advantage of the dome housing
when used externally is the fact that it negates all of the problems of wind
drag which, as we shall see in a moment, af fect housings mounted onto
P/T units. Perhaps one drawback with the external dome housing when
compared with a traditional design is that, at the time of writing, no-one
has come up with an ef fective design for a wash/wipe mechanism other
than a mop on a long pole! In particularly dirty environments this should
be given serious consideration because, wher e the camera is not easily
accessible, the lack of a wipe facility can mean that the pictures very quickly
become obscured by splashes of dirt on the dome.
Dome housings ar e particularly useful wher e it is desirable not to
reveal the direction in which the camera is pointing. By coating the dome
274
Closed Circuit Television
Visor
Wiper
Wash nozzle
Wiper motor
Wash reservoir
Pump
Figure 12.1 External housing incorporating common features such as wash
and wipe and sun visor
with chrome or aluminium, or by using a smoked cover, the camera can be
concealed. The drawback of using such dome covers is that the light input
is attenuated – by as much as two f-stops in the case of a coated cover and
as much as one f-stop in the case of smoked covers. A further drawback is
that the curvature of the dome cover will inevitably introduce a degree of
optical distortion, and in some cases this can be quite noticeable. When
choosing a dome the issues of light attenuation and optical distortion
should be considered.
Although some manufactur ers of camera housings simply state the
intended applications for each of their units, the majority use Standard BS
EN 60529: 1992, which is the corr ect method of specifying any form of
enclosure within the electrical industry. The degree of protection afforded
by an enclosure is indicated by a rating known as anindex of protection (IP).
The IP codes are summarized in Table 12.1. In the code, the first digit indicates the level of protection against ingression by solid objects, whilst the
second digit indicates the level of protection against liquids. Where an ‘X’
appears in place of the first or second digit, this indicates that ther e is no
guarantee of any protection in the corresponding area. An additional letter may be placed at the end of the IPnumber in cases where it is necessary
to indicate a level of protection of persons that is higher than that afforded
by the first digit.
As can be seen, the codes ar e intended to include a wide range of pr otection ratings ranging fr om minute particle ingr ession to whole body
ingression, and from full immersion in water to no moistur e protection
Ancillary equipment
275
Table 12.1 Outline of IP (index of protection) codes
IP
number
First digit
(touch & solid objects)
Second digit
(liquids)
0
1
No protection
Protection against contact by
large part of body or objects
50 mm diameter
Protection against contact by
Standard Finger and objects
12 mm diameter
No protection
Vertical dripping water shall not
impair equipment operation
2
3
Protection against solid objects
2.5 mm thick
4
Protection against solid objects
1 mm thick
5
Protection against dust ingress
in quantities capable of
interfering with equipment
operation
Total protection against
dust ingress
6
7
–
8
–
Dripping water shall not impair
equipment operation when
housing is tilted to 15° from
vertical
Spray or rain falling 60° from
vertical shall not impair
equipment operation
Spray or rain falling from any
direction shall not impair
equipment operation
Low pressure water jets from
any direction shall not impair
equipment operation
High pressure water jets from
any direction shall not impair
equipment operation
Immersion of housing shall not
permit water ingression (within
specified time & pressure)
Indefinite immersion of housing
shall not permit water ingression
Additional letter: A Protection against contact by a large part of the body
B Protection against contact by a Standard Finger
C Protection against contact by a tool
D Protection against contact by a wire
whatsoever. From the point of view of CCTV this may appear somewhat
extreme; however, it must be borne in mind that these ratings apply to any
containment intended for electrical installations, including high voltage.
Example: A camera housing has a rating of IP50. In this case the housing
would offer a high level of protection against dirt and dust, but absolutely
none against moisture. Such a housing would be suited to indoor use only.
Example: A camera housing has a rating of IP66. In this case the housing
would offer a high level of protection against dirt, dust and moisture. Such
a housing would be highly suited to outdoor use.
Example: An electrical equipment housing has a rating of IP4X. In this
case the housing would offer a reasonable level of protection against small
276
Closed Circuit Television
objects but give no guarantee regarding moisture ingression. Such a housing would be suited to indoor use only, and even then only in a clean and
secure environment.
Example: An electrical equipment housing has a rating of IPXXA. In this
case the housing would offer protection against accidental human contact
but give no guarantee regarding moisture ingression, and does not protect
against dust, dirt and deliberate ingress by persons. Such a housing would
be suited to indoor use only , and even then only in a clean and secur e
environment.
Towers and columns
Mounting a CCTV camera on top of a tower is common practice, but there
are a number of things to take into consideration such as positioning, possible planning permissions, installation of supply and signal cables to the
site, the physical installation of the tower , cable management, protection
against attack, installation of site drivers and other ancillary devices and
access for servicing. Let’s look at each of these in turn.
Positioning of a CCTV camera tower can be very important and the
specifier who looks at the problem purely from the point of view of obtaining the best field of view may find him/herself coming unstuck. For
example, companies that have installed cameras directly in front of a private dwelling may quickly find themselves having to disconnect them
when the owner of that dwelling takes out a court order on the grounds of
it infringing his/her human rights. It may also be that the location of the
tower is determined by the location of existing under ground cable ducts,
the cost of excavating for new cables being pr ohibitive. In ar eas where
building development is still taking place, the specifier must try to take
into account the erection of structures that may impair the camera once it
has been installed; it would not be the first time that a lamp post or hoarding has been erected directly in front of a CCTV camera!
Town councils ar e very particular about what can and cannot be
erected in their centres, and special planning r ules may apply; for example, the style of the tower may have to blend with that of the str eet lighting and, in the UK, most local councils will only permit the use of dome
housings in town or city centres because of their aesthetic qualities.
The laying of under ground cables is always expensive, so if a costeffective means of laying the power and signal cables out to a camera site
can be found, it is always worth considering even though it may not be the
best location from a field of view perspective. The client has to decide if a
marginal improvement in camera performance is worth a considerable
increase in installation cost.
Tower structures require a solid anchor and should be installed by
skilled civil engineers. Manufacturers will provide guideline instructions
on such things as the minimum depth and ar ea of the foundation, but
knowing the correct mix and the minimum curing time of the concr ete is
Ancillary equipment
277
something that only qualified persons will know, not to mention that digging holes in built-up ar eas often unearths such things as high-voltage
cables, sewers, and gas and water mains.
CCTV camera towers ar e generally of two designs: the solid column
structure and the lattice tower (Figure 12.2). The lattice tower is by far the
strongest structure because its low surface area offers less wind resistance
Housing mount
(options to suit installation)
Anti-climb
frame
Ladder bar
(optional)
Lattice structure
(a)
(b)
Conduit or
cable tray
Access cover
Concrete slab
Ground level
Conduit
Cables
(c)
Figure 12.2 Typical camera towers and columns. (a) Fixed column – may be
cylindrical or hexagonal. (b) Lattice tower structure – may also have a wind down
mechanism. (c) Wind down column
278
Closed Circuit Television
and the triangular lattice pr oduces a very rigid constr uction. However, an
open lattice means that the cables must pass largely unprotected up through
the centre of the tower, a feature that is not very desirable for most installations. Some protection can be af forded by using steel conduit, but this will
not prevail against a determined attack. Solid columns ar e, by comparison,
far more suited to the task because they of fer considerable protection of the
vertical cables. With regard to their strength and rigidity, a hexagonal solid
column will be mor e stable than a cir cular design, which means that ther e
should be much less of a tendency for the column to wobble in high winds.
This added stability can be of r eal advantage in high columns wher e the
camera is fitted with a high-magnification lens.
Physical attack against CCTV camera installations is becoming more commonplace as those who would wish to indulge in a life of crime find CCTV
increasingly cramping their style. What these people ar e rapidly learning is
that once a camera is ‘taken out’, it may be some time before the owners have
found a budget to have it r epaired. This said, it is up to the specifiers and
installers to do their utmost to pr otect cameras, and mounting them atop a
high tower is one effective method of achieving this, provided that the access
cover at the base of the tower is secur e! Manufacturers have for a long time
employed anti-vandal bolts or other methods of deterring would-be attackers; however, in recent times a number of access cover designs have pr oven
to be woefully inadequate against a lar ge crowbar, and once a fire has been
lit inside the tower ther e is usually little left of the fibr e-optic cables above
ground afterwards. Such damage is costly and time-consuming to put right,
which is precisely what the criminal fraternity like. Ther efore, when selecting a column for situations where such attack is likely (housing estates, town
centres, etc.), look carefully at the design of any low-level access covers and
be convinced that it will hold out against a sustained physical attack.
With respect to attack, also be aware of the ‘monkeys’ who are capable of
climbing columns. If a person is able to get up to the camera head, then they
can very quickly inflict extensive damage and put the camera out of action.
To cater for this eventuality , manufacturers offer a range of optional anticlimb devices which generally take the form of a series of spikes mounted
around two thirds of the way up the tower. Beware when fitting these; when
working from a hydraulic platform these spikes ar e quite capable of taking
out the eye of the installer whilst he is securing the fixing bolts. Eye pr otection is advised.
Remember that for fully functional cameras it will be necessary to house a
site driver somewhere atop a tower. Many designs have optional brackets or
other mounting devices to accommodate these items, but make sur e that
the tower you are having installed has such facilities – preferably before you
install it!
It is one thing to install a tower that of fers a high degr ee of protection
against attack, but the service engineer must be able to access it at a later
date. In the case of lower towers having only a fixed camera, ladder access
may be adequate, but this does mean that a ladder bar should be fitted to the
Ancillary equipment
279
tower before it is erected. These bars usually drop down over the tower and
are then secured. The bar offers a solid point against which a ladder can rest
and be secured, and a well-designed bar will have flanges, one each end, to
prevent the ladder fr om sliding of f. This said, when working in the UK,
engineers must always be mindful of the r egulations that apply when
working at heights; and in many cases working fr om a ladder, even if
secured at top and bottom, would not be acceptable.
For higher towers with mor e complex installations on top, access will
either have to be via a hydraulic platform, or a wind-down tower may be
installed. There are a number of designs of winding mechanism available,
depending on the weight of the tower . When installing such towers, take
into consideration the direction in which it must wind down; for example,
having it fall acr oss a motorway carriageway might not be the best idea!
Also, be aware that you may require special training in the use of the winding mechanism before being permitted to operate it. In r ecent years there
have been a number of serious accidents (some fatal) where engineers have
failed to operate the winding mechanism corr ectly and either the winding
handle has spun out of control, hitting the operator, or the tower has come
down too fast and struck the operator or other persons close by.
Pan/tilt units
Earlier in this chapter it was pointed out that a P/T dome housing has
many advantages over the mor e traditional pan/tilt unit and camera
housing assembly. Nevertheless, the older style camera housing still has
some advantages over dome technology; for example, dome housings do
not easily accommodate large lenses, cleaning can be problematic, a dome
assembly does not easily accommodate lighting units, and motorized
dome assemblies, at the time of writing, still tend to be mor e expensive
than equivalent P/T and housing assemblies. Ther e are still many occasions when these advantages tip the choice of assembly in favour of a
camera housing which, when the camera needs to be moved, necessitates
the use of a P/T unit.
Where it has been decided to use a traditional P/T unit and camera
housing, before selecting the unit there are a number of points to consider
such as loading, wind drag, supply voltage, maximum drive speed, and
whether the assembly is to carry lighting.
Loading refers to the physical weight that the pan/tilt unit is to carry
and will include the combined weight of the housing, camera, lens, lighting units and anything else that is on the assembly. A pan/tilt unit with an
excessive load will initially fail to meet operational expectations, its movement being sluggish and possibly having a tendency to overrun when the
operator expects it to stop. However, in the long term the excess load will
cause wear of the gears and possibly motors, r esulting in premature failure of the unit. A further problem that may be encounter ed with excess
280
Closed Circuit Television
loading is that the unit may have insufficient power to lift the assembly up
when it has been tilted downwards beyond a certain point.
Each pan/tilt unit has a maximum load rating which is typically between
12 kg and 50 kg, although with the reduction in size and weight of cameras
and lenses during r ecent years the r equirement for heavier duty units is
becoming less. When selecting a unit the specifier should first of all calculate the total weight of the proposed assembly and then, as a rule of thumb,
add another 2 kg allowance for cable drag and wind r esistance. In environments where snow or ice are likely to build up on the housing, a further
3 kg should be added to the total weight calculation. Having arrived at a
final figure for the maximum possible load, a unit with a rating that exceeds
this figure by at least 15% should be selected. Another consideration with
regard to loading is futur e developments. If it is anticipated that further
equipment may be added to the assembly at a later date (for example, a pair
of lamp units), it is advisable to include the weight of this equipment in the
initial calculation to save having to replace the pan/tilt unit with a higher
specification when the system is upgraded. Wher e infrared lighting is a
requirement, one possible method of reducing the load on the pan/tilt unit
is to use LED lamp units rather than tungsten halogen types (see Chapter 3).
In open areas the problem of wind drag can be quite serious. An underrated pan/tilt unit will not be able to function well in strong winds, and it is
not uncommon to have badly specified assemblies that simply spin ar ound
like a weather cock on a church spire in strong winds because the forces acting on the housing overcome the torque in the gear train. In some cases this
action can break the teeth off the drive gears, although better-quality pan/tilt
units may incorporate some from of clutch assembly which will allow them
to slip without damage. Having said this, a corr ectly rated and installed
pan/tilt unit should function perfectly well in str ong winds, although the
units may r equire more frequent servicing in the form of lubrication and
greasing than units that operate in less hostile conditions.
Some pan/tilt units have the housing mounted on top wher eas others
have it mounted to one side (Figur e 12.3). In general, top-mounted units
tend to have a lower load capacity than an equivalently rated side mount
unit because of the difference in the centre of gravity of the load with respect
to the drive mechanism. This point is illustrated in Figure 12.4 where it can
be seen that, because of the leverage effect of the housing on the drive shaft,
the amount of energy required to lift the assembly back from a tilted position is much greater for a top mount pan/tilt than it is for a side mount. This
does not mean to say that side mount units are necessarily better, because
a top mount design should always be able to perform to its design specifications; it is just a fact of physics that a side mount unit will be able to handle its load with a lesser amount of energy than its top mount equivalent.
Top mount units come into their own when the assembly is required to
have one or two lamp units fitted.Although it is possible to hang a lamp unit
from the underside of a side mount assembly, it is far easier to do this with
top mount units – especially when two lamps are required because in this
Ancillary equipment
281
Tilt
Camera housing
fixing holes
Tower head
fixing holes
Pan
Tilt axis
Camera housing
fixing holes
Pan axis
Figure 12.3 Top and side mount pan and tilt units
Drive shaft
Drive shaft
Figure 12.4 A side mount unit does not have the problems of leverage associated
with top mount units
282
Closed Circuit Television
case they may be mounted below and to either side of the camera housing,
producing a balancing effect on both the assembly and the pan/tilt gear box.
Returning once more to the issue of wind drag, the ef fects of wind on
top and side mount units can be somewhat dif ferent. The pan/tilt unit in
a side mount assembly of fers a degree of shielding to the housing fr om
side winds, whereas for a top mount unit the housing is exposed in all
directions. Furthermore, because of the leverage effect of the housing on a
top mount unit, the wind forces acting on the pan/tilt gears are far greater
than for side mount designs.
Pan/tilt motors are usually rated at 24 Va.c. or 230/110 Va.c., although
d.c. units ar e available, including 12 Vd.c. for internal applications. The
decision whether to opt for a low- or high-voltage device will be dependent largely upon the operating voltage of the rest of the system. For example, if everything else in the camera assembly is rated at 24 Va.c., there
seems little point in employing a high-voltage pan/tilt unit as this would
negate one of the main advantages of having a low-voltage system, i.e.,
simplified installation requirements with regard to Electrical Installation
Codes of Practice.
The methods of power connection to pan/tilt units can dif fer between
manufacturers, but the wiring configuration given in Figur e 12.5 is one
that is shar ed between a number of major manufactur ers. Before purchasing a unit for an application, make certain that the connection
arrangement is suited to the site driver with which it will be expected to
function.
Cable management for fully functional camera assemblies is very
important because a poorly installed connection between the moving
housing and the fixed site driver can result in restricted movement, additional loading, water ingression and failure of the assembly due to broken
connections. For a sound installation the cables should be contained
Pin 1 Neutral/common
4
3
Pin 2 Pan right
Pin 3 Pan left
Pin 4 Tilt upwards
7
2
5
Pin 5 Tilt downwards
Pin 6 Autopan – not always employed
Pin 7 Electrical earth
1
6
Figure 12.5 Common pan/tilt connection arrangement using an Amphenol plug
and socket connector
Ancillary Equipment
283
within a flexible conduit and fixed at both ends using appropriate moistureresistant glands. All cable entries should be at the bottom of both the housing and the side driver containment to minimize the chance of moistur e
ingression. The cable should be long enough to permit fr ee movement of
the assembly through all of its r equired field of view and should neither
restrict nor be restricted by the movement of the housing. Having a long
drop of cable between the site driver and the housing is the key to unr estricted movement and limited loading effects.
A particular point to look out for is where the housing assembly is fixed
at the end of a boom arm which is itself secured to a solid structure – usually
a building. In this case the cable should be secur ed by some form of tie
wrap along the boom before entering the housing, allowing enough slack
at the housing end to permit fr ee movement without the flexible conduit
rubbing on the boom. Such a rubbing action will result in wear of the conduit and subsequently the cables inside.
The drive speed for a pan/tilt unit is quoted in degr ees per second, and
this can vary considerably between dif ferent units. Panning r equires less
energy than tilting, so for any given unit the panning speed is usually much
higher than the tilting speed. When considering quoted speeds, bear in mind
that these are generally quoted assuming a near to maximum load condition;
however, this speed performance may alter if the load figure deviates either
side of the maximum stated load condition.
To prevent a pan/tilt unit from overrunning in either the horizontal or
vertical directions, limit switches may be employed. For a d.c. motor circuit this principle was discussed in Chapter 4 with regard to the zoom lens
motor and is the same for the pan and tilt motors in a P/T unit. When the
pan/tilt unit reaches the limit position, the switch opens and power to the
motor is cut. As can be seen fr om Figure 12.6, diodes across the switches
enable the motor to power-up when the polarity of the supply is reversed.
For a.c. motors the principle is somewhat dif ferent because it is not possible to reverse the motor by simply reversing the voltage polarity. In this
case the limit switches ar e connected to the site driver , which will cut
power to the motor. Motor reversal is then usually performed by having
two sets of windings in the motor . Limit switch adjustment generally
involves some method of releasing microswitches and sliding them to the
desired positions befor e securing them once again. A minimum of four
switches is required to set limits for left, right, up and down movement.
Limit switch adjustment is important if the housing is to be prevented, for
example, from hitting a wall or fr om tilting downwards to a point fr om
which the motor has insufficient torque to lift it back up.
Pan/tilt units require routine maintenance if reliable performance is to
be assured. This largely takes the form of lubrication or gr easing of the
gears but in some cases may involve the cleaning away of old gr ease that
has become thick due to ageing effects or contamination from dirt or metal
particles from worn gears. During maintenance, gears should be inspected
for signs of wear and r eplacements fitted as necessary. Where the gears
have worn away, or where the teeth have been stripped by the action of
284
Closed Circuit Television
D1
S1 open in "Pan left"
S2 open in "Pan Right"
S1
A
M
B
Pan motor
S2
Pan left
Pan right
A 24 V
A 0V
B0V
B 24 V
D2
Figure 12.6 Limit switch circuit for a pan motor. The same arrangement would
be used for the tilt motor
Figure 12.7 A P/T unit where the tilt mechanism gears have failed. This is a
fairly common problem with these units, especially if the unit is carrying an
excessive load
high winds or excessive loading, the P/T unit will either spin ar
ound
freely (if the pan gears have failed) or fall down to a vertical position
(where the tilt gears have failed). A typical example of such a failur e is
shown in Figur e 12.7. It should be pointed out that any servicing that
Ancillary Equipment
285
requires the unit to be dismantled would normally be undertaken in a
workshop rather than at the camera head.
Motors are prone to overheating when the movement of the unit becomes
restricted, which in turn may be caused by faulty gears or overloading of
the unit. A motor that has overheated will often have short circuit turns in
its windings which will result in a reduction of torque, causing the unit to
become sluggish. In extr eme cases a motor may cease to function altogether, resulting in the loss of either the pan or tilt function. Wher
e a
motor is found to be defective, it is worth investigating the possible causes
for its failure rather than simply installing a replacement.
Other servicing points would include a visual inspection of all cables,
conduits, mounting brackets, etc. for signs of wear/corrosion. Also check
that limit switches are secure and are still functioning.
Monitor brackets
In many r espects the same r ules that apply to the selection of camera
brackets also apply to monitor bracket selection; however, it must be kept
in mind that a CR T monitor is generally much heavier than an internal
camera installation. There are a number of swivel mount bracket designs
available specifically for monitor use and it is advisable to employ these
rather than adapt something else, if only to avoid litigation in the event of
any mishap. But of course, it is still up to the installer to ensur e that the
bracket chosen is rated to carry a load in excess of the expected monitor
weight.
The installer must also verify that the surface to which the bracket is to
be fixed is capable of r etaining the proposed fixings. Where there is any
doubt, additional means of support must be applied – monitors ar e generally sited where people circulate, and one simply cannot take the risk of
a unit falling onto a person.
Power supplies
It is usually a simple thing to connect all of the equipment in the contr ol
room to the 230/120 V mains supply. However, this is not always as easy
for cameras and other r emote equipment because it may involve a lot of
expensive civil work to install the cables. Alternatives to mains-operated
cameras are 12 Vd.c. and 24 Va.c. versions. These use appr opriate lowvoltage power supplies which may be installed either in the control room
or at suitable locations around the site. Extra low voltage (ELV) cables are
installed alongside the co-axial signal cables to carry the power.
A variety of 12 V and 24 V power supply units are available offering different current ratings, typically between 1 A and 4 A. As with any power
supply, the r equired rating is determined by the load, and car e must be
286
Closed Circuit Television
taken not to over-run a power supply as it will very quickly fail. The issues
surrounding power supply rating for cameras were discussed in Chapter 6.
An a.c. power supply unit is primarily a mains to 24V step-down transformer with a rating of at least 1 A. For higher curr ent ratings a lar ger
transformer may be used, but this can lead to problems with voltage regulation. When a transformer is operating with little or no load, the secondary voltage often rises above that stated, the output voltage falling
progressively as the load current is increased. It is asking a lot of a transformer to provide a constant 24 V across a load current range between 0 A
and 4 A, and because of this some larger a.c. power supplies incorporate a
number of separate 1 A rated transformers. Alternatively a transformer
with a number of secondary outputs may be used.
A d.c. power supply is more complex because it requires rectifier, filtering and regulation circuits (see Figure 12.8). The voltage regulator has the
task of maintaining a constant 12Vd.c. output irrespective of load current.
However, it will have a maximum current rating, usually between 1A and
4 A. Overload protection is essential if damage is to be prevented in cases
where the rating is exceeded or an accidental short cir cuit occurs across
the line. A fuse is the simplest form of protection; however, many regulator ICs incorporate an overload protection circuit which switches the output voltage to 0 V until the overload condition is removed.
P
230 V
Step-down
transformer
Rectifier
Filter
Voltage
regulator
12 Vd.c.
N
Overload
protection
Figure 12.8 Block diagram of a d.c. power supply
Power supplies are generally constructed within metal housings to prevent external RFI getting onto the supply lines, and also to contain any
EMI coming from the power supply. This screening only functions effectively if the earth connection to the casing is properly made.
The main advantage of using EL V is the fact that cable installation is
very much simplified and avoids the mor e involved inspection and testing required for 230/120 stallations, which include visual inspection of all
mechanical protection and housings, insulation resistance tests, earth loop
impedance tests, and polarity checks. In essence, mains installations must
be carried out by a ‘competent person’, which does not simply mean
someone who thinks they know what they ar e doing, but someone who
has proven their competence – in other wor ds, a qualified electrician. By
contrast, in the UK, 12/24cables only have to comply with IEE regulations
Ancillary Equipment
287
regarding segregation and mechanical protection, and they must be tested
for earth leakage and insulation r esistance integrity prior to connection
into the system.
Voltage drop
The main problem associated with low-voltage supplies is that of voltage
drop, especially in longer cable runs. For any given conductor, if the crosssectional area is doubled, the area through which electrons pass is doubled,
so the resistance is halved. Conversely, doubling the length of the conductor
will double the resistance. This relationship is expressed in the formula:
R
L
A
where the symbol means ‘is proportional to’. This relationship is illustrated in Figure 12.9.
1m
2m
1cm
Resistance R1
Resistance 2 R1
2 cm
Resistance 2 R1 R1
2
Figure 12.9 Effect of conductor length and area on resistance
All cable contains r esistance, and although this is very small in a short
length, because the r esistance increases with conductor length, for longer
cable runs the resistance can amount to a few ohms. It should also be er membered that a power supply requires two cables for positive and negative or,
in the case of an a.c. supply , send and r eturn. Therefore the resistance is
doubled. The equivalent circuit for a power supply and load showing the
cable resistance is given in Figure 12.10.
For d.c. and 50/60 Hz a.c. supplies the cable can be considered to have
a purely resistive effect, and thus voltage drop can be considered from the
point of view of Ohm’s Law . For the cir cuit in Figure 12.10, the voltage
drop would be found fr om V I (Ri Rii), where the curr ent can be
considered to be the rated current for the equipment constituting the load.
288
Closed Circuit Television
Ri
RL
E
Rii
Figure 12.10 Cable resistance in a circuit, represented by Ri and Rii
For example, if the load is a colour camera for which the manufacturer has
specified a current of 300 mA, then the voltage drop would be taken to be
V 300 103 (Ri Rii).
Cable resistance is usually quoted in /m or /km; however, where
this is the case, the resistance figure usually refers to the total series resistance of the two conductors in the circuit (i.e., Ri Rii), so there is no need
to double this figure when performing cable calculations.
In some cases a manufacturer might quote actual voltage drop figures,
in which case you might see something like: 80mV/A/m at 22°C. Note that
the resistance of copper incr eases with temperature, and this is why the
temperature at which the measur ement was taken must be stated. If this
cable were made to function at 30°C, then the resistance would rise slightly
and there would be a corresponding increase in the voltage drop.
The resistance can vary considerably between different types of cable, and
a manufacturer’s technical information or support line should be consulted
where voltage drop needs to be calculated during the system planning stage.
Example
The camera in the system illustrated in Figure 12.11 is quoted as having a
current rating of 250 mA and a minimum operating voltage of 9 Vd.c. The
resistance of the cable is quoted as being 0.03 /m at 20°C. Calculate the
voltage drop in the cable, and determine whether or not the camera would
be able to operate satisfactorily.
Solution
Total cable length 450 m
Total cable resistance at 20°C 450 0.03 13.5 Voltage drop I Rcable 250 103 13.5 3.375 V
Voltage at camera Vpsu – Vcable 12 – 3.375 8.625 V
Ancillary Equipment
289
P.S.U.
12 Vd.c.
450 m
Figure 12.11
Thus it can be seen that the camera would not function corr ectly, if at all.
One possible method of overcoming this problem would be to use a highergauge cable, remembering that increasing the cross-sectional area reduces
the resistance. Where a multicore cable is being used with spare cores available, doubling up the cor es can cure the problem. In the case of the above
example, doubling the cores would reduce the cable resistance to 6.75 , and
therefore the voltage dr op would be 250 103 6.75 1.7 V. This gives
12 – 1.7 10.3 V at the camera, which is within its operating voltage range.
It is sometimes stated that the voltage drop for a.c. is ‘less than for d.c.’,
and that this is why a.c. power supplies are better. While it is true that a.c.
supplies are better because the voltage drop appears to be less, in fact the
50/60 Hz alternating curr ent is subject to just the same r esistance, and
hence power loss, as a d.c. current along the same cable. The reasons why
voltage drop figures appear to be less for a.c. powered cameras are: (1) the
current consumption of the camera is less because of the higher operating
voltage (24 V); (2) the 24 V is actually the RMS figur e, and the tr ue peak
value is 24 1.414 33.9 V. Summarizing this point: the voltage drop for
a.c. power supplies is less of a pr oblem because the r eduction in current
causes a reduction in voltage drop; and a voltage drop of, say, 3 V equates
to an RMS loss of only 3 0.707 2.12 V.
13 Commissioning and maintenance
Commissioning
System commissioning is vitally important. This is the point in the installation where every part of the system is tested against the original specifications to ensure that they ar e met in every way . It is wher e the system
may be proven to be meeting any operational requirement (OR) that is in
place (in terms of such things as live and r ecorded image resolution, camera fields of view, zoom capability, and ability to perform under dif fering
lighting conditions). It is where the quality of the installation workmanship
is proven. It is where the safety of the system is proven (in particular, does
it conform to the IEE 16th Edition of the Wiring Regulations – BS 7671?). It
is where test results are recorded for future reference and, when commissioning is properly carried out, it is wher e the installation company can
hand the system over to the client with confidence.
To begin with, a visual inspection of all parts of the system should be
carried out. In particular the connections to fully functional cameras
should be checked for fr eedom of movement and, in the case of external
cameras, integrity of weatherproof seals. All cables should be clearly tagged
and identified.
Programming and setting up of equipment should be checked to ensur e
that the system will meet the OR, or wher e this does not exist, that it will
meet the customer’s requirements and needs. The inspection should include
the programming of multiplexers, matrix switchers, operator levels of
authority (restricted access), PTZ limit switches, dome pr esets, DVR/VCR
time-lapse settings, alarm input response(s), and VMD zones and sensitivity.
The correct operation of all equipment must be confirmed, not for getting peripheral items such as lights, video r eplay facilities, alarm detectors, video printers, and bulk tape erasers.
It should go without saying that the results of all of these tests and inspections should be documented and filed for futur e reference. The installing
company should also notify relevant authorities such as the local council
(where planning permission was originally r equired for external equipment installation), police, Inspectorate body, etc.
Measuring resolution
The subject of r esolution has been discussed on a number of occasions
in this book, but how can this actually be measur
ed in a way that is
Commissioning and maintenance
291
meaningful to all concerned? Or to put this another way, how can the commissioning engineer actually pr ove that the system is performing to the
design specifications, or operational r equirement? Resolution is affected by
many factors, and one cannot make assumptions about the quality of the picture, even when the system has been built using high-specification equipment throughout. Changing lighting and weather conditions can have a
dramatic effect on system performance, as can other changes in the environment in which the cameras are operating. For example, in a town centre system when the Christmas illuminations are turned on, the dramatic change in
the colour temperature can cause iris settings to alter , which in turn may
cause changes in the depth of field and focusing, not to mention the fact that
some cameras may be blinded almost completely.
Various methods have been devised for testing the picture performance
of a system, per haps the simplest of these being the use of an appr oved
test card. The card is positioned in front of the camera, which is adjusted
so that the image fills the entire monitor screen. In this condition each set
of markings on the car d corresponds to a particular TVL picture resolution, which is normally specified in the instr uctions that come with the
test card. Note that the monitor should be adjusted for best picture before
any observations are made.
To test the system performance under site conditions the test card should
be held or fixed in front of each camera in turn, and the camera adjusted so
that the image of the card fills the monitor screen. The maximum TVL resolution is determined from the smallest set of markings that can be r esolved
on the displayed card. It is not uncommon for the measur ed resolution figure to be lower than that quoted for the camera or monitor. This may be due
to such factors as lighting conditions, length of cable r uns, or the quality of
other components in the system. For example, a high-quality camera fitted
with a poor-quality lens will have its performance impaired.
An important point to note from the test procedure described above is
that it applies to a situation wher e a stationary image is filling the height
of the screen. But how does this relate to an image of a person filling only,
say, 25% of the screen height, and moving at speed? This is a much mor e
subjective measurement and has been the cause of much debate.
In the UK, the Home Office Scientific Development Branch (HOSDB, formerly PSDB) has carried out research into the issue of image size and picture
resolution, and from this research a set of guidelines has been devised. These
have been adopted by inspectorate bodies such as NACOSS and the SSAIB.
The HOSDB have formulated the following classifications of CCTV
system use monitoring, detection, recognition and identification.
Monitoring is defined as an image that allows the observer to see the
location, speed and direction of a person in the field of view. This will usually be a wide-angle view. For monitoring purposes the image of a person
will be not less than 5% of the screen height.
Detection should allow the observer to locate a person with a high degree
of certainty having been prompted to do so by a guard, police, alarm system
292
Closed Circuit Television
or other means. For detection purposes the image of a person will be not less
than 10% of the screen height.
For recognition, the pictur e quality must be adequate to permit an
observer to say with a high degree of certainty that the person on the monitor is the same one they have seen before. In this case the image of a person
will be not less than 50% of the screen height.
Identification is the highest resolution image of a person and must contain sufficient detail to enable the operator to see the person clearly
enough to be able to describe them, or to identify them again. Such an
image is only possible from a close-up or zoom shot, the disadvantage of
this being that you cannot see or record any activity other than that by the
person being monitored. For identification purposes the image of a person
will be not less than 120% of the screen height.
These image sizes are illustrated in Figure 13.1. Note that for each classification it has been assumed that all parts of the system are adjusted and are
functioning correctly, and that the height of the person is 1.6m. It might also
be necessary to test the performance under different lighting conditions.
From these specified image sizes it is possible to draw up system specifications which include a TVL figure for a given image size. For example,
Monitoring
Image 5% of the screen height
Detection
Image 10% of the screen height
Recognition
Image 50% of the screen height
Identification
Image 120% of the screen height
Figure 13.1 HOSDB classifications of CCTV images, with recommended
minimum image size for each classification
Commissioning and maintenance
293
the specification for a certain camera location could be that it will be used
for recognition purposes, and that the image must have a r esolution of at
least 250 TVL. In other words, the image of a person filling 50% of the screen
height must have a resolution of 250 TVL. For an engineer to test this during final commissioning using a test car d, the camera will be adjusted so
that the card fills 50% of the screen height. Under these viewing conditions,
the 250 TVL markings on the card should be discernible.
An alternative to using a test car d is to employ a test tar get known as a
Rotakin, illustrated in Figure 13.2. This target stands 1.6 m tall when fixed to
its frame and, in addition to the human head outlines, has a range of markings relating to TVL. The TVLfigures for each marking are given in the table
Marking
TVL figure for 100% R
A
B
C
D
E
F
G
H
J
K
L
M
500
450
400
350
300
250
200
150
100
80
40
20
Figure 13.2 Rotakin standard target. The higher resolution bars are highlighted
in the close-up picture
294
Closed Circuit Television
accompanying Figure 13.2. Note that these figur es apply when the image
height of the Rotakin fills the monitor screen, a condition defined as 100% R.
Up to this point we are still looking at a stationary, clear target, and we
have yet to take into account a camouflaged moving target.
To meet the problem of camouflage the Rotakin comes with a combatstyle jacket which can be fitted when r equired. This is particularly useful
when testing the system performance for picking out people other than
black and white striped ones! In other wor ds, in this mode the Rotakin is
no longer being used to determine TVLresolution, but rather the ability of
the system to display a useful image under dif ficult conditions. With the
jacket fitted it is also possible to estimate the response time of the operator.
To simulate a moving target the stand on which the Rotakin is mounted
has a 12 V motor which rotates the target at a rate of 25rpm. This provides
a representation of a person ‘moving quickly but stealthily’, to quote the
Rotakin manual. The original function of the rotating target was to measure the amount of lag in the image, which is an ef fect that is produced by
tube cameras; however , because tube cameras ar e no longer used, the
Rotakin is generally only ever used as a stationary target.
System handover
Handover is one of the most important stages in a CCTV installation, and
yet it is too fr equently glossed over as something that is only of limited
importance. It has been proven that where a well-presented handover has
taken place there is much less chance of multiple-site r e-visits during the
first few months of the system going live. Such visits are far more likely to
occur when the owner/operator has little or no idea of how to operate and
maintain the system.
Handover is the point where the engineer becomes a training officer, and
as soon as he/she realizes this, the whole issue of instr ucting the customer
takes on a new importance. Have you ever attended a manufacturer’s training course where the person delivering the training appears covered in dirt,
wearing overalls that have been thr ough a few filthy loft spaces, his hair
completely unkempt, smelling of sweat and other unsavoury odours, and
looking like he really doesn’t want to stay to deliver the training because he
is just too tired? One would hope not! Yet too often this is what the pr oud
owner of a new, expensive CCTV system (or other security system) is pr esented with when the engineer presents the handover.
A well-presented handover begins like any other well-presented training
session, with planning and pr eparation. The installer should know in
advance when the system will be ready for handover, and should therefore
plan ahead. It is important to know how many people will be involved in
the use of the system and, wher e relevant, the level of authority of each of
these people. The owner should be made awar e of the need to have these
people available to attend the handover training. Also let the owner know
that the area must be ‘controlled’ during the handover; that is, interruptions
Commissioning and maintenance
295
should not occur, and it should be possible to demonstrate the operation of
all equipment. For example, there is no point in trying to demonstrate VMD
in a department stor e when it is full of customers. A mutually convenient
day and time should be agr eed, which may be done verbally , but it is
much better if the arrangements are made in writing, therefore covering any
‘misunderstandings’. Finally, when the engineer turns up for the handover
session, he/she should appear well presented and looking prepared. Some
installing companies insist that befor e beginning a handover the installer
washes, puts on a shirt and tie and generally makes themselves presentable.
In the case of larger installations it is more than likely that handover will not
take place directly following the fixing of the last cable clip and sweeping of
the control room floor area, so the engineer will be travelling to the site,
either directly from home or the office, and should do so suitably dressed.
The amount of preparation for handover will of course be governed by
the size of the system. The handover of a town-centre-sized system cannot
be compared to that for a small system comprising two cameras, one monitor and a DVR in a local petrol station. Nevertheless, the same principles can
still be applied to both situations.
Before beginning the handover, make sure that all documents are available and corr ect. These documents should include such things as user
manuals, handover agreement, maintenance agreements, contact numbers
and addresses, and copies of commissioning checklists.
Upon meeting the owner and any other persons involved in the handover, establish by means of questioning how much they know or do not
know about CCTV. In the majority of cases people know very little about
the operation of CCTV equipment; however , where a system has been
upgraded, it might be that the installer is only required to explain the differences the upgrade may have made to the system operation and capability.
At the start of the handover, walk the owner around the system, showing them the location of every item of equipment, making them awar e of
the field(s) of view for each camera. Demonstrate the operation of the
equipment, making sure that the customer regularly handles it for themselves during the handover. In fact, the owner should handle the equipment more than the engineer during this period.
Tape management, as outlined in Chapter 8, should be explained, as
should the implications of the Data Protection Act 1998 (applicable in the
UK). That is, that the owner of a CCTV system is legally r esponsible for
the security of r ecorded material, and should any member of the public
bring a formal complaint about recorded material being released into the
public domain, either intentionally or otherwise, the CCTV system owner
might well be br ought to court. In the UK, the days of passing CCTV
footage to TV companies for the purposes of entertainment are at an end,
because any member of the public upon seeing themselves ‘starring’ in
such a programme, or even appearing incidentally, could complain that it
is a misuse of the system, and that it is in breach of their human rights.
In relation to maintenance, instruct the owner in the use of wash/wipe,
de-misters, etc., as well as VCR head cleaning cassettes (Chapter 8). Inform
296
Closed Circuit Television
them of the need to stor e tapes away fr om the magnetic fields which
emanate from many items of electrical/electronic equipment. Also, where
relevant, point out the health hazar d associated with unshielded tape
degaussers (bulk erasers); there is the possibility of these units upsetting
the operation of heart pacemakers, and people who have one of these
fitted should never operate this equipment.
Preventative maintenance
The BSIA Code of Practice for the Planning, Installation and Maintenance
of Closed Cir cuit Television Systems; 109: Issue 2: October 1991 br eaks
preventative maintenance down into two levels: Level 1, which is basically a full system test, and Level 2, which is a full system test plus an inspection of all equipment associated with the installation.
A Level 1 visit must be made at least every twelve months, the first visit
taking place within twelve months of the handover. The engineer should
test the operation of every piece of equipment, including each wash/wipe
facility, lights, PTZ, etc. The pictur e quality fr om each input should be
examined, and careful attention paid to loss of quality due to such things
as dirt on the housing windows or cobwebs. Also check the picture contrast, colour quality and noise level, noting any possible deterioration and
logging it for further investigation.
Compare the system operation against the original OR. Make sur e that
fields of view ar e unaltered, housing stop positions ar e correct, etc. Also
make sure that the field of view for each camera has not become obstructed
by objects which wer e not pr esent when the system was installed – for
instance, the large tree which was a mere sapling a few years previously, or
the new lamp post which the council thought to site directly in front of the
camera. Although it is not the job of the engineer to cut down the tr ee (or
dismantle the lamp post!), it is his/her r esponsibility to inform the owner
of the system of environmental changes which mean that the system is no
longer able to perform to the OR.
A Level 2 visit, as defined by the BSIA, does not have a fixed frequency,
but is to be agreed between the owner and the installing company. On the
other hand, the sort of checks that this visit involves should be made at
least every two years if not every year.
During a Level 2 visit, all of the system tests for a Level 1 visit ar e carried out. In addition, visual inspection of all equipment must be carried
out. Camera housings, receiver units and P/T units must be opened and
inspected for corroding of electrical terminals, water ingr ession and any
other signs of wear or damage. P/T units should be greased, wiper blades
replaced as necessary, heater performance verified, flexible cables checked
for damage or wear, etc.
Camera towers or columns should be inspected for signs of damage or
corrosion, as should wall-mounting brackets. In particular , the fixing
Commissioning and maintenance
297
points of towers and brackets should be inspected for signs of corr osion,
fatigue and loose or missing bolts.
All lighting should be tested, and wher e the lights ar e not easily accessible it is suggested that the lamps are replaced as a matter of course whilst
access equipment is in place. This is pr obably more cost-effective than
having to return in the not too distant future to replace a defective lamp.
Any repair or replacement work that cannot be completed at the time of
the maintenance visit should be recorded, and arrangements made for the
work to be completed at the earliest opportunity . All maintenance work
should be recorded and the document signed by both the engineer and the
owner or a representative of the owner. A copy of the record should be left
for the owner, and the original r eturned to the company of fice where it
should be stored in a secure place for future reference.
Where maintenance or servicing involves the temporary disconnection
of equipment which will leave a part of the system inoperative, a temporary disconnection notice should be completed and signed by both the
engineer and the owner, or other responsible person.
With particular reference to VCRs, as discussed in Chapter 8, because of
the wear and tear on the mechanism in a time-lapse machine, a full service
should be carried out every twelve months. This will involve the removal
of the machine from the site for a period of time, and it may be necessary to
install a machine on temporary loan. In this case the engineer must compare the features on the loan machine with the one that it is replacing ensuring that either (a) the machine is capable of offering all of the features needed
to maintain the system performance or (b) the customer is made aware of
any VCR operations that are not available.
At all times the engineer must be car eful not to degrade the performance of any part of the system by r eplacing defective components with
parts of a lower specification or parts that are incompatible. For example,
if a camera is found to be defective, the r eplacement should be a model
that is of equal if not better specification. Similarly , if the har d disk in a
digital recording system fails, replacing it with a lower-capacity unit will
have a serious impact on the recording and archive period.
Corrective maintenance
Just as it is necessary to cr eate accurate and legible pr eventative maintenance records, the same applies to any corrective work that is carried out on
the system, either during routine visits, or in response to a breakdown call.
Copies of all corr ective maintenance documents must be stor ed in a
secure place in the system file at the company office. On the one hand this
protects the company in the case of any redress from the owner who may
in the future find reason to accuse the company of failing to carry out certain work. On the other hand the information in these r ecords can prove
helpful in the event of future problems with the system.
298
Closed Circuit Television
Fault location
This subject has to a lar ge extent alr eady been dealt with, having considered the fault diagnosis and repair techniques pertinent to the equipment
under discussion in each chapter. However, we will now look at fault diagnosis from a system point of view.
A logical approach to fault diagnosis is essential if servicing is to be both
sound and cost-effective. But what is logical fault diagnosis? Well, perhaps
it is easier to begin by stating what it isn’t: the engineer who, when faced
with a problem, falls into a state of panic and thr ows all theory knowledge
aside and instead resorts to replacing items of equipment at random until at
last the problem goes away, is certainly not using a logical approach to fault
diagnosis. Having personally observed both CCTV and intr uder alarm
trainees in this ‘state’, it has been interesting to note that in many cases not
only is the theory thrown aside, but the test equipment remains safely stored
in (unopened) cases. In some instances trainees have been seen to pull out
and replace a complete cable r un just to see if it is the cause of a pr oblem,
before having applied any form of test meter to the cable. This is certainlynot
a logical approach to fault diagnosis!
To locate a fault efficiently, an engineer would usually begin by giving
careful consideration to the fault symptom(s). T elevision can be a very
informative piece of equipment in this respect because each symptom displayed on the screen very often points to the fault area, as long as the engineer can recognize the symptoms. Fr om the fault symptom, the engineer
should be able to deduce which parts of the system may be responsible, and
which parts can be ruled out.
The next step in logical fault diagnosis very often is to fall back onto
experience. Here the engineer ’s mind is working somewhat like a computer database, sear ching though ar chive files for any pr evious experiences of the same symptoms on similar equipment. This does not mean to
say that the cause of the trouble will definitely turn out to be the same, but
where the engineer is awar e of a common pr oblem in a certain item, or
combination of equipment, it makes sense to begin by testing in this area.
Where experience alone does not produce a rapid diagnosis, the engineer must change gear and begin to sear ch his/her mental ‘database’ for
the appropriate CCTV and/or electr onics theory. The skill and ability of
an engineer can usually be determined fr om how well they can do this
because, as much as it is unsatisfactory to throw the theory aside, it can be
equally unhelpful to pull down so much theory that the engineer becomes
totally confused. Having the ability to apply the corr ect theory at just the
right moment is a skill in itself.
It is one thing to have the knowledge, yet this may not be a lot of use
unless one has access to the test equipment to enable voltage and waveform measurements to be taken. A calibrated, quality multimeter is the
bare minimum that an engineer should carry . With this the engineer can
verify the presence and value of a.c. and d.c. voltages and currents, as well
Commissioning and maintenance
299
as test for open circuit, short circuit and resistance faults. Remember that
when measuring mains voltage potentials, only appr oved shrouded and
fused test leads must be used.
For testing co-axial cables the multimeter resistance ranges are of limited
use, and therefore a time-domain reflectometer is another very useful piece of
test equipment for the service engineer, which can very quickly pay for itself
by saving a lot of time and effort in trying to locate a fault along a length of
co-axial cable. The TDR was discussed in Chapter 2 (Figure 2.32).
Because we are dealing with video signals it has to go without saying
that the best way of testing the condition and level of these signals is to display them on an oscilloscope. And yet the CCTV industry is pr oving very
reluctant to move in this direction for fault location. There are a number of
reasons for this. First of all it is fair to say that lugging a scope around a site
(possibly in the rain) is not the best thing for either the engineer or the scope.
Secondly, a scope suitable for CCTV use will cost at least £400 brand new ,
and thus it is not viable to provide every engineer in a company of any size
with such an item. But there is a third reason for not using oscilloscopes, and
this is that many engineers are afraid of them and simply don’t know how
to operate one. This third reason is not a valid one for ignoring the value of
an oscilloscope in locating certain difficult types of fault.
The truth is that the CCTV engineer does not have to be an expert at operating an oscilloscope, because he/she is not trying to display a wide range
of signals. In practice the engineer usually wishes to display the same type
of signal every time, this being one or two lines of a 1 Vpp CCIR television
signal. Therefore, once the scope has been set to display this signal, it will
never require anything more than possibly minor fine adjustment each time
it is used for this purpose. Later in this chapter we shall look at some default
settings for using an oscilloscope for CCTV applications.
Hand-held oscilloscopes are available and ar e much more convenient
to use for CCTV system testing. However, the less expensive units tend to
offer only a limited bandwidth and can have a slow scr een refresh rate,
making it difficult, if not impossible, to notice any rapid fluctuations in
signal level or condition. This is not to say that all hand-held scopes suffer
this problem, and devices such as that illustrated in Figur e 13.3 are very
effective and are ideally suited to CCTV applications. Being menu driven,
scopes such as this very often include a ‘V ideo’ menu option which sets
the vertical and horizontal scales such that the scope will display one or
two lines of video information. The main drawback with units such as this
is their relatively high cost compared with comparable bench models.
An alternative to using an oscilloscope to determine the condition of a
video signal is to use any one of a number of hand-held video signal level
indicators that are available. Compared to an oscilloscope these are far less
expensive, cumbersome and delicate and are a worthwhile addition to any
service engineer’s test kit. An example of such an indicator is illustrated in
Chapter 2; Figure 2.31. This version comes with its own signal generator ,
which brings us to our next point.
300
Closed Circuit Television
Figure 13.3 A hand-held oscilloscope is far more suited to CCTV testing than a
full-sized bench model. Units such as this incorporate rechargeable batteries,
and have default settings to enable rapid display of analogue video waveforms
Scoping or measuring video signals from cameras is not the best way of
testing for some faults, and the engineer sometimes needs to know for
sure that the signal being injected into a cable or item of equipment is a
CCIR standard video signal. T o this end, a small portable video signal
generator is very useful for fault-finding in CCTV systems. The simplest
of these puts out a black and white bar signal; more sophisticated versions
produce a selection of colour ed and black and white patterns, including
perhaps the most useful which is the standard colour bar display, or staircase when viewed on an oscilloscope.
Oscilloscope default settings
One TV line has a time duration of 64 s, and a peak-to-peak voltage of
approximately 1 V (when terminated into 75 ). To display this waveform
on an oscilloscope the volts/div contr ol must be set to ar ound 0.2 V, and
the timebase speed adjusted for around 10 s/div. If the signal level is correct, then a stable trace should be obtained with the trigger contr ol set to
the ‘Auto’ or ‘Normal’ position. These are the default settings, and once a
Commissioning and maintenance
301
scope has been adjusted to display a stable video signal waveform, if the
controls are left in this position, then a stable trace should be displayed
immediately a video signal is applied to the input.
The expected display for these settings when a grey scale video input is
applied is illustrated in Figure 13.4.
Oscilloscopes differ between models, and some of fer features that the
CCTV engineer will never use; however, these features result in a lot more
controls on the front, making them more difficult to set up. (They are actually
no more difficult to set, but the additional controls give more opportunity
for the unwary operator to become confused.) When choosing an oscilloscope for CCTV use, a basic dual-beam model with a minimum bandwidth
of 20 MHz is an ideal choice. The bandwidth of an oscilloscope is important as, among other things, it determines the triggering ability. Anything
below 20 MHz will struggle to display a stable video waveform, and models
with a bandwidth of 30–40MHz are generally a good compromise between
quality and price (for CCTV applications).
For those who do not regularly use an oscilloscope, the following pr ocedure should produce a stable, single beam display on the majority of
models (refer to Figure 13.4).
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
Switch on
Adjust INTENSITY (Brightness) control to mid position
Set TRIGGER LEVEL control to AUTO (NORMAL) position
Adjust X and Y POSITION controls for centre position
Set channel 1 Y GAIN/AMPLITUDE control to the 0.2 V position (or
nearest value on your scope)
Set Timebase control to 10 s/div
Set TRIGGER SELECT switch to CH1 position
Set trigger switches to AC and positions
Set INPUT COUPLING switches on channels 1 and 2 to AC position
Connect scope to the equipment under test.
Adjust FOCUS and INTENSITY controls for best definition.
Provided that the signal under test is good, then a stable display should be
obtained. If the display is jittering or moving horizontally across the screen,
try adjusting the trigger contr ol, perhaps switching it to the MANUAL
position. If in doubt as to whether or not the scope is set correctly, try applying a known good video signal rather than the one under test. A correct
display will confirm both that the scope is set up corr ectly, and that the
signal under test is indeed incorrect.
Y Position
Time/Division
(Timebase)
Volts/Division
CRO 30 MHz
Channel 2
Channel 1
1
.5
.2
1
.05
2
5
.02
.01
100
20
5
10
5
msec
.1
500
2
10
µsec
Pull Auto
ac|dc|gnd
ac
.5
.2
1
.05
.02
.01
ac|dc|gnd
tv
Focus
Intensity
Probe
Test
0.5 Vpp
1 kHz
Input 1
Input coupling
Figure 13.4 Typical basic dual beam oscilloscope
Trigger slope
Input 2
ext Trig
Trigger select
.1
2
5
ch1|ch2|ext
Power Off
X Position
Trigger level
Glossary of CCTV terms
Some of the terms listed in this glossary may have mor e than one meaning,
depending on the context in which each is taken. Wher e this is the case, the
definitions given here relate to closed circuit television.
Alarm activation A piece of equipment that is made to change its mode
of operation by the activation of an alarm input. T ypically a VCR, multiplexer or matrix switcher.
Algorithm A mathematical term used to describe a procedure for resolving
a problem.
Amplitude modulation (AM) A transmission method where the signal to
be sent is added to a much higher frequency carrier signal in such a way that
the carrier amplitude is made to change in sympathy with the wanted signal.
Used to transmit audio and video signals. Also used in magnetic video tape
recording.
Analogue A signal that is represented by changes in voltage level.
Aperture The size of the hole produced by the iris. The size of the aperture determines the amount of light that will fall onto the pick-up device.
ARP Address resolution protocol. A protocol which enables communication with a host using its MAC address in order to assign an IP address.
Artefact In relation to digitized video signals, an artefact is any visible component in the picture display which is not a part of the original. This is rather
like ‘noise’ in analogue displays; however, where large amounts of compression of the video signal has taken place, many artefacts may appear as a ersult
of both losses and incorrect attempts by the decompression circuits to ‘guess’
at the picture content.
Aspect ratio The ratio of the horizontal and vertical monitor scr
dimensions. For CCTV this is 4:3.
een
Aspherical lens A non-spherical lens that produces much less optical distortion than a conventional lens and, because of the way that the light travels
through the glass, offers a higher light output for the same aperture setting.
Attenuation Term used to describe a reduction in signal amplitude as it
passes through a system or medium.
Automatic gain control (AGC)
Amplifier circuit that automatically
changes its gain as the input signal level varies to maintain a constant output
signal level.
304
Glossary
Automatic iris (AI) Iris that is controlled by the signal level in the camera.
Changes in light input level r esult in changes in signal amplitude which
are in turn used to adjust the iris to compensate.
Automatic level control (ALC) A circuit that maintains a constant output signal level despite large changes in input signal level (see also AGC).
In CCTV such a circuit may be used to control the lens iris.
Back focus Mechanical adjustment of the position of the image device in a
camera. Used to set the distance between the back of the lens and the pick-up
to 12.5 mm for a CS mount lens, or 17.5 mm for a C mount lens.
Back light compensation (BLC) Feature in a camera that causes the iris circuit to ignore the bright ar eas of the image and to open up suf ficiently to
allow the darker areas to be visible – albeit at the expense of the light areas.
Back porch Period lasting 5.9 s following each horizontal sync pulse. It
is at black level to ensur e that the electron beam in the CR T is cut off for
the duration of the line flyback period.
Backbone Main bus onto which all network devices connect. Each individual connection to the backbone is known as a segment. The backbone
usually has a much broader bandwidth than the segments as it has to carry
all of the traffic on the network.
Balanced cable A two-wire method of signal transmission wher e the
signal phases are such that unwanted interference signal are cancelled out
at the receiving end.
Bandwidth A term that is used to define the range of frequencies between
the upper and lower cut-off points of a transmission system (measured at the
–3 dB points). In CCTV there is a direct relationship between bandwidth and
picture resolution.
Baud Unit of measuring the data rate in a digital system, usually in bits
per second (bps).
Bit Derived from the term binary digit. This is a single piece of data at
either logic zero or logic 1.
Black level In a video waveform, this refers to the voltage level that produces black level on the CRT or other display device. For a CCIR standard
signal this is 0.3 V.
Blanking period Times during both the horizontal and vertical scanning
periods when the electron beam in the CRT is returning (‘flying back’) to the
position for the next start of scan. During these periods the video signal
waveform is held at the black level (0.3 V) to ensur e that nothing is displayed on the scr een. For a UK P AL transmission, the line or horizontal
blanking period is 12 s and the field or vertical blanking period is approximately 1 ms.
BNC Co-axial cable connector commonly employed in CCTV installations. The term ‘BNC’ is derived from ‘Bayonet–Neil–Concelman’.
Glossary
305
Bridge A network device which connects two segments (similar to a switch,
except that a switch can connect a number of pairs of segments). It is used to
keep traffic on the two segments separate except when it is necessary to pass
it over.
Bus Electrical conductor carrying one or more data signals. Many items
of electronic equipment have internal bus lines in the form of copper
tracks on the PCB, but in CCTV applications we ar e usually referring to
external data bus lines such as RS 485, RS 422, RS 232, etc.
Byte Term used to define a gr oup of 8 bits (binary digits) in a digital system. The term was derived because the earliest computers used only 8-bit
words for communication, whereas modern processors commonly communicate using 32-bit words.
C mount Common screw thread developed for lenses used by the cine
industry but adopted by the CCTV industry . A C mount lens has a distance of 17.5 mm between the image device and the back of the lens.
Cable equalizing Equipment added to a co-axial cable transmission
system for the sole purpose of corr ecting for high-frequency losses in the
cable.
Candela The unit of luminous intensity. One candela is the amount of light
that is radiated in all dir ections from a black body that has been heated to a
temperature equal to that at which platinum changes from a liquid to a solid
state.
CCD (charge coupled device) A solid state device that is capable of storing
small electrical charges. Originally intended for use as a digital store for computing applications, it has been adapted to store the output voltages from the
photo diodes in a camera pick-up prior to them being passed on to the signal
processing stages.
CCIR (Committée Consultatif Internationale des Radiocommunicationale)
The English translation means ‘Consultative Committee for International
Radio’. This is the European body that has been responsible for the setting
of Standards for television systems in Europe.
CCTV (closed circuit television) A television system that is not broadcast
to air and therefore its images can only be accessed by persons with a connection to that system.
Characteristic impedance The impedance of a cable. This is a function of
the inductive, capacitive and d.c. resistive properties of a cable and is usually
quoted assuming a (theoretical) cable of infinite length. For CCTV applications, co-axial cable must have a characteristic impedance of 75 .
Charge An accumulation of electrons in a device or area. Electrical charge is
measured in coulombs and is equal to an accumulation of 6.289
1018
electrons.
306
Glossary
Chroma burst A sample of the colour subcarrier signal. It is transmitted
during the back por ch period and comprises 10 cycles of the 4.43 MHz
subcarrier for a PAL transmission, or up to 10 cycles of the 3.58 MHz subcarrier for NTSC transmissions. The burst signal serves as a form of sync
signal for the colour decoding circuits in a television, monitor or VCR.
Chrominance Term used to define the colour signal components in a television transmission.
CIF Common interchange format. A set of standards which specify picture formats in pixel sizes.
Co-axial cable An unbalanced cable comprising a core surrounded by a
braided or solid screen. The two conductors are separated by an insulating
material that is designed to act as a capacitive dielectric. The total inductive,
capacitive and d.c. resistive properties of the cable produce an opposition to
a.c. currents that is known as the characteristic impedance. For CCTV applications, co-axial cable must have a characteristic impedance of 75 .
Colour temperature The type of light produced by a source. It is a scientific
measure of the wavelengths of light, stated in degrees Kelvin (K).
Composite sync A signal that contains both the horizontal and vertical
sync pulses but no video signal components.
Composite video Signal containing the luminance, chr ominance, sync
and colour burst components.
Compression With respect to digitized video signals, this r efers to the
removal of binary data containing information r elating to the signal that is
considered, for one r eason or another , to be r edundant. Compression is
required to overcome the need for the large amounts of digital storage capacity associated with digitized video signals.
Cross-modulation An effect whereby two high-fr equency (r.f.) signals
interact, resulting in an unwanted interference.
CS mount The most common lens mounting type used in CCTV. It has a
distance of 12.5 mm between the image device and the back of the lens. A
CS mount lens cannot be used with a C mount camera.
CSMA/CD Carrier sense multiple access/collision detection. Method of
data transmission used in Ethernet whereby hosts access the data bus as and
when required. When a data collision occurs, hosts r e-send their data, and
will continue to do so until the transmission is successful.
CVBS
See Composite video.
Dark current Thermally induced curr ents in a camera pick-up device
which produce an apparent signal output even in a total absence of incident
light.
Data cable Cable used for the transmission of digital signals. T ypically
twisted pair but may be co-axial.
Glossary
307
Datagram A data packet with a TCP header. Ethernet devices transmit
data across a network by breaking it down into these smaller packets and
re-assembling them at the receiving end.
dB (decibel) Logarithmic unit used to define the ratio between two signal
amplitudes. A change of 3 dB represents a doubling or halving of electrical
power; a change of 6 dB represents a halving or doubling of electrical
voltage.
Depth of field The range (in distance) in fr ont of the lens wher e objects
remain in focus. This decr eases when either the focal length or the apertur e
size is increased (F-number decreased).
DHCP Dynamic host configuration protocol. A server application that is
used to automatically assign IP addresses to hosts on a network.
Digital signal
zero and one.
A voltage signal used to r epresent the binary values of
Direct drive (DD) Term used to describe a CCTV camera lens having an
electro-mechanical iris that is controlled via a changing d.c. voltage derived
in the camera.
Distribution amplifier A device containing a single input and a number of
amplifiers which provide separate (isolated) outputs. Used for sending a
single video signal to a number of pieces of equipment without affecting the
1 Vpp level.
DNS Domain name service. A service which enables people to work
with user-friendly names rather than IP addresses. The DNS server
resolves (converts) the domain names into the IP addresses of the hosts to
which users of the network wish to connect.
DOS (disk operating system) Software used by the majority of computers
to enable communications between the CPU and har dware devices such as
hard disk and printer.
Duplex When used with r espect to CCTV multiplexers (MUX), this term
refers to a unit that is able to of fer simultaneous recording and replay (note
that for analogue systems, two VCRs must be available).
DVR Digital video r ecorder. Video recording equipment that employs a
hard disk medium for the storage of video information.
EI (electronic iris) A circuit in a CCTV camera that varies the amount of
time that the pick-up device is active during any field period. The ‘exposure’
time is reduced as the light input incr eases, thus emulating the action of a
mechanical iris. This feature enhances the performance of a camera that has a
fixed iris lens fitted.
EIA (Electronics Industry Association) The body that has been r esponsible for the setting of Standar ds for television systems in the USA, Canada
and Japan; namely, the NTSC 525 line system.
308
Glossary
Electromagnetic radiation Electric field with an associated magnetic field
travelling through free space at the speed of light in the form of waves. A
wide range of fr equencies, known as the electr omagnetic spectrum, exist.
Some of these are used for the purposes of radio signal transmission; others
produce the visible light spectrum, whilst others are harmful to humans (i.e.,
X rays, Y rays[**G.1] and gamma rays).
Electron beam Stream of electrons emitted from the cathode of a cathode ray tube (CRT).
EMI (electromagnetic interference) Electromagnetic signals that enter
a system or item of equipment and corr upt the wanted signals. EMI may
be naturally occurring (i.e., lighting, solar flar es, static electric char ge), or
man-made (i.e., radio transmissions, electric motors, electrical switch gear).
F-number Figure used to denote the size of aperture in a lens. It is determined by the ratio of the focal length to the ef
fective diameter of the
lens/iris aperture. A small F-number indicates a large aperture and a high
light throughput.
F-stop Aperture setting on a camera lens. Each stop position r epresents
a doubling or halving of light throughput.
Fast scan A technique for transmitting CCTV signals digitally over the
ISDN telephone network.
fc (foot candle) Unit of light measur ement, used primarily in North
America. 1 foot candle (1 fc) is approximately equal to 10 lux.
Fibre-optic transmission Signal transmission method that uses light
pulses passing through a thin length of optically clear material. Capable of
low loss transmission over long distances and is impervious to the effects
of RFI and EMI.
Field This contains one half of the video information in a television
frame. The CCIR Standard has 50 fields per second, each field containing
3121⁄2 lines. The EIA Standard has 60 fields per second, each field containing
262.5 lines.
Field of view The area that may be viewed through a lens. It is determined
by the relationship between the angle of view and the distance between
the object and the lens.
FIT (field interline transfer) chip CCD image device that combines the
technologies of the interline transfer and frame transfer chips, of fering
improved S/N ratio, low smear and good low-light sensitivity.
FM (frequency modulation) A transmission method where the signal to
be sent is added to a much higher frequency carrier signal in such a way that
the carrier frequency is made to change in sympathy with the wanted signal
amplitude. Used to transmit audio signals. Also used in magnetic video tape
recording.
Glossary
309
Focal length The distance, measured in millimetres, from the secondary
principal point in the lens to the final focal point (at the image device). A
short focal length produces a wide angle of view.
Format With respect to CCTV lenses and c ameras, this describes the size
of the active area of the pick-up device; typically 2⁄3, 1⁄2, 1⁄3, or 1⁄4.
Frame One complete television pictur e made up fr om 625 (525 NTSC)
line of information over two interlaced fields. The frame rate is 25 (30
NTSC) per second.
Frame store Commonly a digital memory device capable of storing the
data relating to one complete digitized television frame.
Frequency Measurement of the number of times in one second that a
cycle repeats itself. The unit of measurement is hertz (Hz) or, in the USA,
cycles[**G.2].
Front porch Period lasting 1.4 s prior to each horizontal sync pulse. It is
at black level to ensure that the electron beam in the CRT is cut off for the
duration of the line flyback period.
FTP File transfer protocol. A part of the TCP protocol suite that is used
as a means of transferring files between two hosts.
Galvanometer A device that can convert electrical energy into mechanical energy without the mechanical complexity of a d.c. motor. Used to control the iris veins in a DD lens.
Gamma correction Modification to the shape of a luminance signal in a
camera made in or der to corr ect for non-linearity in a CR T phosphor
response. In many CCTV cameras this is adjustable to compensate for differing monitors or lighting conditions.
Gateway A network device through which all data must be r outed in
order for it to be passed on to another network, either local or wide ar ea.
A gateway is normally a router.
Gen lock Means of maintaining sync between cameras in a CCTV system.
A master sync signal is taken from either one camera or from a master generator and is fed to the ‘Gen Lock’ input on each camera.
Ground loop Circulating earth curr ent formed by the scr een on a coaxial cable. Such curr ents can cause r olling shadows on the pictur e
(known as a hum bar), poor synchronization, horizontal tearing of the picture, or loss of telemetry functions.
Ground loop transformer A 1:1 transformer capable of passing video signal
frequencies. Placed in series with a co-axial cable to break the direct earth
connection between camera and monitor formed by the cable screen.
Hard disk Data storage device employed in all computers. Uses a solid
metal disk onto which data is stored using magnetic recording.
310
Glossary
Hardware address Also known as the MAC (media access contr
ol)
address or the Ethernet addr ess. A unique address comprising a twelve-bit
hexadecimal code, separated by hyphens, that is assigned to every network
device. Used by applications such as DHCP to identify a host so that, for
example, an IP address may be assigned.
Hertz (Hz) Cycles per second. The unit by which frequency is measured.
Hop count Normally set at 32, this determines the maximum number of
network routers that the datagram may pass through before it is discarded as
undeliverable. If there were no hop count, undelivered data would simply
continue to pass around the Internet indefinitely, causing the Web to become
congested with such data.
Host An Ethernet network device that uses an IP address.
HTTP Hypertext transfer protocol. A protocol that manages communications between the web br owser and the web server and ensur es that a
requested folder, link, etc., is opened.
Hub A simple network device used to construct a star network topology. A
typical hub will have four or eight ports. Data on any one port is immediately
connected to all of the other ports.
Hue This term refers to the fr equency of a light sour ce: red, green blue,
etc.
Hum bar See Ground loop.
Illuminance Measurement of light in lumens per square metre. The unit
of measurement is the lux.
Illumination Measurement of light coming from a secondary surface or
source. The unit of measurement is the lux.
Imaging device Device that is able to convert light ener gy into electrical
energy. In modern CCTV cameras this will be a charge coupled device (CCD)
chip.
Impedance (Z) The opposition to curr ent flow in an a.c. cir cuit, measured in ohms. It is the combined effect of the d.c. resistance and the inductive and/or capacitive reactances.
Infrared cut filter A filter that blocks the passage of infrared light frequencies. Such filters are used in colour CCTV cameras to pr event the ingress of
IR light that would otherwise result in incorrect colour signal production.
Infrared (IR) light Frequencies of light that are just below those of visible
light (wavelengths between 700 nm and 10 mm). All CCD image chips ar e
sensitive to these fr equencies and in many cases this can be used to an
advantage.
Infrared pass filter A filter that only allows infrared light frequencies to
pass through. Such filters are placed in front of white light sources in the
Glossary
311
manufacture of IR lighting units. Typical wavelengths for CCTV lighting
units are 715 nm and 850 nm.
Infrared spectrum See Infrared (IR) light.
Intranet A private network comprised entir ely of private lines connecting
the sites.
IP conflict A situation that occurs on a network when two hosts ar e
assigned the same IP address. When such a condition exists, neither host
will be seen on the network until one of them is removed, or one of the IP
addresses is changed.
IRE (Institute of Radio Engineers (of America)) A unit of measurement
for a 1 Vpp video signal. It divides the signal between sync tip and peak
white into 140 equal levels. For example, when 140 IRE 1 Vpp video signal, the 0.3 V sync pulse level would be 42 IRE, etc.
Interlaced scanning Method of producing a television picture by scanning each frame in two parts, first the odd television lines and then the
even lines. Used to reduce picture flicker in CRT display units.
Interline transfer chip Type of CCD image chip where the storage areas
are adjacent to the photodiodes. This technique eliminates the vertical
smearing problems associated with frame transfer chips, but the chip is
less sensitive in terms of light input/signal output.
Internal sync Horizontal and vertical synchr onizing pulses ar e produced by the camera internally.
IP rating The index of protection rating for an enclosure, e.g., IP 65. The
first digit indicates the level of protection against ingression by solids; the
second digit indicates the level of protection against ingression by liquids.
Iris mechanism in a camera that governs the amount of light input.
ISDN (integrated services digital network) A high-speed telephone transmission system for digital signals. 64 kB and 128 kB speeds are available.
Kell factor Figure used when deriving the horizontal resolution for a television picture.
LAN (local area network) Term used to describe a data communications
network within a defined area, usually a single site.
LCD (liquid crystal display) A device which uses the twisted nematic
property of liquid crystal to contr ol the light passing thr ough in order to
produce a visual display.
LED (light emitting diode) Gallium arsenide diode that emits light when
a current passes through its PN junction.
Lens In CCTV, this term usually r efers to a lens assembly which is an
array of lenses with an iris mechanism. Its function is to gather light and
focus it onto the pick-up device.
312
Glossary
Light meter Hand-held measuring device comprising of a photo pickup and a meter. The meter is calibrated to display the light level in units
of lux.
Line fed The camera power is supplied via the co-axial video signal cable.
Line locked Synchronization pulses pr oduced by each camera ar e referenced to the a.c. mains frequency.
Luminance (Y) Monochrome or black and white content of a video signal.
Lux
Unit of measurement of light.
MAC address
See Hardware address.
Matrix switcher Equipment used to switch a number of cameras
between one or more monitors.
Microwave transmission For CCTV, this is a video signal transmission
method using microwave radio signalling, usually in the 3GHz to 10 GHz
band. Greater range than infrared signal transmission.
Modal distortion Signal distortion in fibre-optic cable caused by a multiplicity of light paths through the cable.
Modem Term derived from its function, which is a modulator/demodulator. Used to interface equipment having a digital output, and a conventional analogue telephone line.
Monochrome A black and white television signal or system.
MPEG (Moving Picture Experts Group) A body set up by the ISO in 1988
to devise standards for audio and video compression.
Multimode Fibre-optic cable where the light is made to ‘bounce’ from side
to side as it travels along. This cable type introduces modal distortion, but it
is much less expensive than mono mode cable, wher e the light travels in a
direct line along the cable length.
Multiplexing Signal transmission method where two signals are transmitted on the same cable or carrier in such a way that they can be separated at
the receiving end. The term also r efers to equipment capable of pr ocessing
video signals such that more than one image can be displayed on a screen, or
recorded simultaneously (see MUX).
MUX Abbreviation for a video multiplexer. A unit that processes video signals such that more than one image can be displayed on a screen, or recorded
simultaneously.
NAT (network address translation) A method that allows efficient use
of IP addresses when transmitting data fr om one network to another
across the Internet. A router or proxy server takes the datagrams on a private network and removes the local IP address, replacing it with another
address before passing it over the Internet.
Glossary
313
ND (neutral density) filter A lens filter that affects all frequencies of light
in the visible spectr um by the same amount, ther efore causing an overall
reduction in the light level entering the lens. It is used to simulate low-light
conditions, forcing the iris to open and giving the best conditions for lens
focus adjustment.
ND spot filter A graduated filter at the centr e of a lens. It has minimal
effect when the iris is wide open; however , its effect increases as the iris
closes. This type of filter prevents the aperture from becoming too small to
control effectively under bright light conditions.
NetBEUI (netBIOS extended user interface) See NetBIOS.
NetBIOS (network basic input output system) The standard interface to
networks for IBM PCs. Although relatively simple to set up and administer, it
cannot be routed, which makes it inef ficient in terms of data traf fic, and is
only viable for smaller networks.
NIC (network interface card) Also known as the Ethernet car d. Every
host on an Ethernet network r equires one of these to connect to the network (although devices such as CCTV cameras usually do not employ an
actual card). The NIC performs the functions of or ganizing the data into
frames, managing the transfer of these frames between hosts, and err or
correction.
Node General term used to describe any device that is connected to a
network (see also Host).
Noise Electrical interference on the video signal. Usually manifest as
grain (speckles) over the picture.
NTSC (National Television Standards Committee) Committee that set
the colour television system standards for North America and Japan.
NVR (network video recorder) A digital video r ecorder that incorporates an Ethernet connection to permit both IP camera inputs and IP monitoring facility.
Octet In relation to IP, an octet is an 8-bit word having 256 possible combinations. An IP version 4 (IPv4) address contains four octets, each separated by a dot.
Optical fibre A transparent material along which light can be transmitted.
Oscilloscope Displays electronic signals in a graphical form, enabling
them to be measured and analysed.
Overscan The monitor display is adjusted such that the electr on beam
scans farther than the edges of the scr een, resulting in a loss of some picture information around the edges.
PAL (phase alternate line) The most common system for the transmission of analogue colour television signals. The system maintains corr ect
314
Glossary
colour reproduction by cancelling out the ef fects of signal phase err ors
that occur during transmission.
Patchcord A short length of flexible fibre-optic cable with a connector fitted
to each end. Used to reconfigure a route between two pieces of equipment.
PCM (pulse code modulation) Signal modulation method whereby the
width of a continuous stream of square wave pulses is made to change in
sympathy with the modulating signal.
Pelco D A set of telemetry protocols devised by Pelco for CCTV applications. Pelco permits other manufacturers of CCTV equipment to use these
protocols, which adds a much welcome uniformity to the industry.
Photo cell A device, the electrical resistance of which is determined by
the light falling upon it.
Photodiode A diode in which the forwar d bias curr ent is determined
not only by the applied voltage but also by the light falling upon the PN
junction.
Ping command Computer command executed in DOS. Used to confirm
that an IP addressable item of equipment is connected to, and communicating with, the IT network.
Pixel Term derived from picture elements. A pixel is a single element of
picture information – the greater the number of pixels, the greater the picture
resolution.
Port In relation to TCP/IP, a port is an address which defines the association between the data being transmitted and the applications on both the
source and recipient PCs for which the data is intended. The sour ce port
identifies the application that sent the data, whilst the recipient port identifies the application for which it is intended. Each port has a specific number; for example, HTTP is assigned port 80.
Primary colours The three colours (frequencies) of light – red, green and
blue – that are perceived by the human eye and are integrated by the brain
to derive all the colours of the visible light spectrum.
Progressive scan A method of producing a picture display in some flat
panel display devices whereby all TV scanning lines are produced simultaneously, rather than using traditional interlaced scanning.
PSTN (public switched telephone network) In the UK, the original low
bandwidth telephone network provided by the Post Office and later British
Telecom. Intended only for speech transmission, it is now used for the transmission of digital signals; however, its low bandwidth makes it of little use
for CCTV applications.
Pulse and bar generator An item of test equipment that pr oduces a
continuous black, white and grey video test signal.
Glossary
315
Quad Abbreviation for a unit that enables four camera signals to be displayed simultaneously on a monitor.
RAID (redundant arrays of independent disks) A method of creating a
large disk capacity by grouping a number of hard disk drives into an array,
with the array acting as a single drive. The method by which data is distributed across the disks provides a high degree of protection against data loss in
the event of a disk drive failure.
Raster The blank white scr een that results from the scanning action of
the electron beam in a CRT before video information is applied.
Reflectance A figure which represents the ratio of light falling onto and
returning from a surface, expressed as a percentage.
Reflected light Area illumination multiplied by reflectance.
Refraction Effect on a ray of light whereby as it passes through different
medium its velocity alters, causing it to bend.
Reserved IP address An IP address that is set in a DHCP scope (by the
IT administrator) to be assigned to the same host (and only that host)
every time that host connects to the network.
Resolution In relation to the definition of a television pictur e, this is a
measure of the smallest detail that may be discerned. The most common
unit of measurement is TVL (television lines).
Router An intelligent network device that is on a par with a PC in terms
of complexity and technology. It is used to intelligently forward data from
one network to another by finding the quickest path, per haps through a
number of other routers.
Routing table A list of IP addresses of network hosts built up inside a
router. When a host attempts to send data to a host on another network,
the router will use its r outing table to determine the shortest path by
which to send the data.
RS port RS recommended standard. Standard input/output connectors
used for data communications. Common standar ds are RS 232, RS 422 and
RS 485.
Scan coils Inductive coils placed ar ound the neck of a CR T. Currents
passing through these coils produce electromagnetic fields which interact
with the electron beam, causing it to deflect both vertically and horizontally.
Scanning Rapid horizontal and vertical deflection of the electr on beam
in a CRT to produce a light output from the entire screen area (see raster).
SECAM (sequential couleur avec memoire – sequential colour with
memory) Colour television br oadcast system developed and used in
France. It is not directly compatible with either PAL or NTSC systems.
316
Glossary
Segment A term used in networking to describe a data bus. A typical network will have a number of segments separated by routers or switches. Each
segment may have a large number of hosts connected to it.
Server A network administration device that provides services to the client
PCs and other devices connected to the network. Such services include
DHCP and DNS, mail services, web connection, fir ewall, application software, print services, file management, etc.
Silicon intensified target (SIT) Type of camera pick-up device developed for use in very low light conditions.
Simplex When used with r espect to CCTV multiplexers (MUX), this
term refers to a unit that can record or replay images, but not at the same
time (unlike duplex units).
Slow scan An early video signal transmission system that used the conventional PSTN telephone network. It employed analogue transmission
and had a very slow refresh rate.
S/N ratio (signal-to-noise ratio) Measurement of the amount of noise in a
signal, expressed in decibels (dB). For video signals, any figur e less than
40 dB will result in unacceptable amounts of noise (grain) in the picture.
Static IP address An IP address that has been assigned to the host manually, rather than automatically via DHCP. One assigned, such IP addresses
will remain unchanged, unless a person makes a manual adjustment.
STP cable (shielded (screened) twisted pair cable) A twisted pair cable
that has an outer screen to provide added protection against the effects of
RFI/EMI.
Subnet mask Data included in the datagram that is used to identify the
network and host parts of the IP address. Subnetting makes it possible to
independently route networks, so increasing the efficiency of the network
by reducing network traffic, reducing the size of the r outing tables, providing a simple way of isolating one network fr om another, and enhancing the ability to provide network security.
S-VHS (super video home system) Analogue video r ecording format
based on VHS, but of fering much higher pictur e resolution (400 TVL as
opposed to the 240 TVL for VHS). S-VHS r ecordings will not r eplay on
conventional VHS machines.
Switch An Ethernet device used to connect hosts for the duration of
their communication. At first glance a switch appears similar to a hub;
however, a switch cr eates a dedicated path between the hosts. This
improves network performance because it means that, for much of the
time, more than one pair of hosts can communicate at the same time.
Synchronizing pulses Pulses added (at the camera) to the video signal.
Used by the monitor to maintain correct scanning correlation between the
camera and the monitor.
Glossary
317
TCP/IP (transmission control protocol/Internet protocol) Two protocols
(sets of rules) used together for the transmission of data over Ethernet.
Telemetry Signalling system used to contr ol functions at a camera head
such as pan, tilt, zoom and wash. May be analogue, but is now mor e commonly a digital system that employs decoders (site drivers) at each camera
installation.
Telephoto lens Correct term for a zoom lens. These lenses have a long focal
length, giving them a high magnification but a narrow angle (field) of view.
Telnet A virtual terminal service which permits a user to connect two
network hosts and take full contr ol over the r emote host. All mouse and
keyboard activity is passed dir ectly to the r emote host so that applications
may be opened and run remotely.
Termination With respect to CCTV , this term is fr equently used to
describe the 75 termination impedance required at each end of a co-axial
cable. Modern equipment employs automatic termination methods that
detect the presence of a BNC connector at the socket; however , for many
years a termination switch would be found close to the video output
socket on monitors or other signal processing equipment.
Timebase corrector (TBC) An item of equipment that is used to align the
timing of the sync pulses fr om all cameras in a system, thus pr eventing
picture roll during camera switching. In most cases the TBC is incorporated inside the MUX.
Time-lapse VCR A video recorder (commonly S-VHS) that is capable of
recording for very long periods on a standar d three-hour cassette by operating a continuous record/pause action.
Topology A term frequently used in networking to describe the wiring
configuration for a particular network; e.g., Bus, Star.
Transducer Any device that converts one form of ener gy into another. For
example, a CCD chip converts light ener gy into electrical energy, whereas a
CRT will do the opposite.
Triplex When used with r espect to CCTV multiplexers (MUX), this term
refers to a unit that is able to of fer simultaneous live monitoring, r ecording
and replay (note that for analogue systems, two VCRs must be available).
Twisted nematic A property of liquid crystal wher eby the molecules
form into helices, causing light to change polarity by 90°. When a voltage
is applied across the crystal structure, the helices break down and light is
able to pass through unaffected (see also LCD).
Twisted pair Type of cable having two wir es that are twisted together,
producing a balanced ef fect whereby electrical interfer ence signals ar e
cancelled out. Capable of much gr eater transmission distances that
co-axial cable types. In some cases the cable may have an outer screen.
318
Glossary
UDP User datagram protocol. A connectionless protocol that simply transmits datagrams acr oss the network to the r ecipient host wher e they may
arrive out of order, and are not subject to any error checking or correction.
Underscan Switchable feature on a monitor that enables the entir e picture to be seen.
Unshielded twisted pair (UTP) Twisted pair cable type having no outer
screen (see Twisted pair).
URL (universal resource locator) Used with HTTP to identify the location
of a network device. A typical URL looks something like http:/ /www.
sitename.com.
Vacuum tube Device such as a CRT, fluorescent display or camera pickup tube. Uses thermionic electr on emission to pr oduce a curr ent flow
between a heated cathode and an anode.
Varifocal Lens having a manually adjustable focal length, giving a
degree of choice over the field of view.
Vertical streaking Bright vertical lines produced in some types of CCD
chip under certain conditions.
VHS (video home system) Analogue video r ecording format developed
for the domestic market and later adapted for CCTV applications in the form
of time-lapse recorders. It is a low-resolution format – 240TVL for colour and
300 TVL for monochrome recordings.
Video launch amplifier Amplifier used to corr ect for signal losses in
exceptionally long lengths of cable. Can be tuned to give greater compensation at higher frequencies where the cable losses are most acute.
Video line corrector Equipment used to corr ect for uneven fr equency
losses in long cable runs. Similar in action to a video launch amplifier, but
it may not have such a high gain.
Video motion detection (VMD) An alarm detection facility whereby the
picture content from a camera is analysed to look for evidence of change
(movement). Originally analogue, modern equipment employs digital
analysis techniques.
Video signal See composite video.
Wavelength Measurement (in metr es) of the pr opagation distance of
one cycle of an electromagnetic wave. It is taken between any two adjacent
points in a waveshape. Relates directly to the signal frequency.
Wavelet (compression) Method of compressing a digitized video signal in
order to reduce the amount of data (file size) needed to r estore each picture
frame. Uses a mathematical algorithm known as wavelet transform.
White level Voltage level in a video signal that produces peak white. For
a CCIR signal this is 1 V (0.7 V above the black level).
Glossary
319
WINS (windows Internet naming service) A service designed to work on
simpler networks where there is no DHCP or DNS server. Such networks
generally use NetBIOS, which is simple to set up and administer,but it is only
suitable for small networks because it cannot be routed and, in terms of bandwidth, is very inefficient.
Y/C Term used to describe the separate luminance (Y) and chrominance (C)
signals. An S-VHS signal connector uses Y/C signal transmission where the Y
and C are passed along separate screened cores within the cable. This results
in improved picture quality because the two signals ar e unable to interfer e
with each other.
Zoom lens See Telephoto lens.
This page intentionally left blank
Index
A/D converter 108, 129, 220, 226
Address Resolution Protocol (ARP)
252, 268, 303
Active line period 91, 95, 110, 114, 128,
145, 198, 213
Additive mixing 48
Alarm input 187, 196, 199, 204, 210,
221, 227, 290
Angle of view 61, 64, 76, 87
ANSI/EIA 24
Aperture 61, 68, 78, 303
Artefact 112, 303
Artificial lighting 53, 297
Aspect ratio 61, 94, 145, 303
Audio/control head 200
Auto re-record 199
Automated patrol 227
Automatic gain control (AGC) 138,
144, 303
Backbone 237, 304
Back focus 86, 272, 304
Back light compensation (BLC) 142, 304
Back tension 200
Balanced cable 21, 304
Baud rate 231, 304
Blanking 98, 128, 225, 304
Bridge 240, 265, 305
Brightness control 159, 162
British Security Industry Association
(BSIA) 4
Bus 234, 305
Cable segregation 16, 287
Camera:
covert 148, 271
format 149
line fed 147
sensitivity 144
Candela
Cable test equipment 43
Category 5 (CAT 5) 24, 29, 236, 237, 268
Category 6 (CAT 6) 24, 236, 237,
266, 268
Cathode ray tube 50, 51, 156
CCIR 96, 101, 104, 299, 305
Characteristic impedance 11, 12, 41, 305
Charge coupled device (CCD) 2, 61,
69, 83, 96, 125
Chromatic aberration 60
Chroma burst (see colour burst)
Chrominance 101, 306
City & Guilds 5
CMOS 2
Co-axial cable 9, 17, 21, 32, 42, 103,
140, 146, 149, 162, 179, 208, 224,
228, 269, 285, 299, 306
Colour burst 102, 141, 306
Colour cameras 54, 99, 134, 137, 142,
147, 288
Colour difference 101, 123, 137
Colour subcarrier 102
Colour temperature 54, 142
Column (camera) 273, 276, 296
Composite sync 306
Composite video 40, 42, 103, 141, 179,
215, 306
Compression ratio 112, 120, 264, 306
Connector:
BNC 10, 17, 22, 41, 106, 140, 160,
179, 205, 304
Phono 41, 162
P plug 79
RJ45 24, 239, 260, 269
SCART 41, 162
UHF 41
Contrast control 162
Control (CTL) pulse 196
D connector 230
D/A converter 215
DAT recorder 183
Data communications/transmission
24, 223, 225, 230, 234, 239, 256
322
Index
Datagram 241, 257, 307
Data Protection Act 1998 3, 295
Depth of field 61, 68, 76, 85, 138,
291, 307
Dichroic mirror 134
Digital timelapse VCR (D-TL) 183, 203
Digital Video (DV) 203
DIL switches 228
Discrete Cosine Transform (DCT) 115
Domain Name Service (DNS) 241, 250,
255, 257, 261, 307
Dome camera 152, 229, 273, 279, 290
Duplex 217
Dwell time 204, 215
Dynamic Host Configuration Protocol
(DHCP) 241, 245, 248, 255, 261,
267, 307
Edexcel 5
EIA/TIA 307
Electromagnetic interference (EMI) 16,
32, 231, 308
Electron gun 158, 160
Ergonomic 181, 272
Ethernet 25, 29, 41, 237, 238, 243, 269
Expander card 211
Extra high tension (EHT) 157
Fast scan 40, 308
Fibre-optic cable 19, 32, 36, 46,
278, 308
Field:
blanking 98, 128, 225
flyback period 90, 98, 126
odd/even 93, 196, 206
of view 59, 61, 65, 76, 110, 141, 149,
220, 308
output stage 164
recording 196
scan 90, 97, 115, 160
sync 97, 137, 196, 206, 214
File Transfer Protocol (FTP) 256, 258,
309
Filter:
IR cut 59, 85, 133, 143, 153, 310
IR pass 85, 310
mosaic 135
ND spot 70, 85, 313
polarising 84
striped 135, 170
F-number 61, 69, 76, 144, 149, 308
Focal length 61, 63, 67, 76, 81, 87,
153, 309
Focal point 59, 83
Focus 60, 65, 68, 81, 272
Foot candle 51, 308
Frame 92, 93, 96, 110, 171, 173, 177,
183, 189, 196, 214, 219, 309
Frame interline transfer (FIT) 128
Frame store 309
Frame transfer 126
Frequency division multiplex 224
Frequency modulation 308
F-stop 69, 85, 144, 308
Fully functional camera 142, 273, 278,
282, 290
Galvanometric drive 73, 80, 309
Gamma correction 138, 309
Gateway 240, 248, 253, 261, 309
Genlocking 140, 207
Ghosting 12, 41
Graded index cable 33
Grey scale 171, 177
Ground loop 17, 22, 232, 309
Hardware (MAC) address 238, 249,
252, 267, 310
HOSDB (Home Office Scientific
Development Branch) 5, 6
Hypertext Transfer Protocol (HTTP)
256, 258, 310
Hue 310
Hum bar 17, 310
Human Rights Act 4, 276, 295
Hunting 80
Illumination 50, 51, 54, 310
Image intensifier 147
Index of protection (IP) 274, 311
Infra-red 36, 47, 54, 280, 310
Interlaced scanning 92, 123,
174, 311
Interline transfer 126, 133, 311
Internet protocol (IP) 241
Intranet 237, 311
Inverse square law 50, 68
Iris 68, 126, 311
Auto (AI) 70, 72, 73, 77, 79, 85,
133, 304
Index
Iris (Continued)
DD 73, 79, 307
Electronic (EI) 76, 133, 137, 307
manual 69, 71, 74, 138, 144
video 73, 138
ISDN 41, 311
Kell factor 96, 145, 311
Launch amplifier 14, 44
Lens:
adjustment 85
aspherical 71, 303
assembly 60, 68, 71, 49
auto 72
calculator 66, 88
concave 59
electrical connections 79
fibre-optic 149
finding 67, 87
format 61, 309
magnification 64
micro 128
mounts 83
pinhole 149
telephoto 77, 317
theory 59
varifocal 68, 318
wide angle 63, 77
zoom 68, 71, 76, 77, 80, 222,
229, 319
Level control 74
Light, characteristics 52
Light emitting diode (LED) 34, 36, 44,
56, 311
Light level meter 58, 312
Light, nature 47
Limit switch 226, 283, 290
Line driver 231
Line flyback period 90, 97
Line lock 140, 146, 312
Line scan 90, 160
Line sync 97, 102, 141
Linear slow speed recording 199
Lumen 50
Luminance 99, 104, 123, 134, 171, 197,
312
Luminous flux 50
Luminous intensity 50
Lux 50, 58, 69, 137, 144, 147, 312
323
Mains isolation 165, 167
Matrix switching 209, 312
Microwave link 37, 312
Mini DV 203
Modal distortion 32, 312
MODEM 40, 312
MPEG 113, 116, 184, 195, 215, 312
Multiplexer 215, 312
NetBIOS 251, 257, 313
Network Address Translation (NAT)
246, 268, 312
Nodal point 61
NTSC 90, 101, 123, 145, 196, 313
NVQ 5
Operational Requirement (OR)
145, 290
Optical speed 69, 78, 149
Oscilloscope 97, 106, 206, 225, 299,
300, 313
PAL 93, 101, 141, 145, 313
PAT test 168
Peak/average control 73, 142
Pinch roller 200
Pincushion distortion 164
PING command 260, 314
Pixel 95, 110, 118, 125, 129, 135, 145,
168, 172, 176, 184, 314
Port 241, 255, 256, 258, 263, 268, 314
Power over Ethernet (POE) 29
Pre-set (PT) positions 223, 226, 290
Primary colours 48, 101, 136, 314
PSTN 40, 314
PTZ 3, 212, 223, 226, 279, 290, 296
Quad splitter 213
Quantization 115
Radio frequency interference (RFI)
10, 22, 32
RAID drive 187, 191, 194, 315
Raster 90, 160, 164, 315
Refraction 52, 59, 315
Refractive index 53, 60
Resolution 8, 15, 61, 91, 93, 106, 112,
120, 125, 134, 140, 144, 145, 150,
156, 170, 172, 181, 183, 187, 194,
197, 203, 290, 315
324
Index
Ribbon cable 31
Rotakin 6, 293
Router 237, 240, 243, 248, 253, 262,
265, 267, 315
RS 232 port 230
RS 422 port 230
RS 485 port 230
Scan coils 160, 315
SCSI drive 191
SECAM 101, 315
Secondary colour 48, 136
Secondary principal point 61
Security lamps 57, 280
Sequential switching 204, 209, 216
Signal transmission 13, 19, 32, 90,
106, 117
Simplex 217, 316
Site driver 223, 226, 232, 276, 278, 282
Skills For Security 5
Slow scan 40, 316
Step index cable 33
STP (shielded twisted pair) 25, 316
Subnet mask 246, 253, 258, 261,
265, 316
S-VHS 43, 162, 183, 190, 197, 203, 316
Switch (network) 239, 253, 265, 267,
269, 316
Switched mode power supply 164
Synchronisation 96, 119, 140, 146
Tape:
management 183, 201, 295
oxide 197, 200
storage 200, 202, 222
TCP/IP 241, 251, 255, 258, 262, 267, 317
Telemetry 12, 20, 23, 41, 213, 219, 231,
273, 317
Telnet 258, 263, 317
Termination switch 12, 178, 205, 239,
317
Time division multiplex 224
Time domain reflectometer 44, 299
Time-lapse recording 183, 188,
197, 317
Tower 146, 276, 296
Tracking 196, 200
Triplex 217, 317
TVL 96, 114, 140, 145, 181, 197, 203,
291, 293
Twisted pair cable 19, 21, 24, 25, 223,
228, 231, 317
User Datagram Protocol (UDP) 257,
318
UHF RF transmission 40
Universal Resource Locater (URL)
259, 318
UTP cable (see twisted pair cable)
Vertical phase adjustment 207
VHS 183, 190, 194, 196, 199, 203, 318
Video compression 107, 110, 123, 184,
187
Video head cleaning 200
Video Motion Detection (VMD) 219,
290, 295, 318
Voltage drop 287
Wash/wipe 222, 228, 271, 295
Wavelet 113, 117, 120, 188, 215, 318
White balance 141
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement