INVESTIGATION INTO LOUD
INVESTIGATION INTO
LOUD ADVERTISEMENTS ON TELEVISION
Submission to the
Australian Broadcasting Authority
by
Federation of Australian Commercial
Television Stations
5 April 2002
FACTS
Submission to the Australian Broadcasting Authority on the
Investigation into Loud Advertisements on Television
Page 2
TABLE OF CONTENTS
EXECUTIVE SUMMARY .......................................................................................................3
1. INTRODUCTION..............................................................................................................4
2. FACTORS CONTRIBUTING TO LOUDNESS ............................................................4
3. TESTS FOR MEASURING AND MONITORING LOUDNESS .................................8
4. THE FUTURE – FURTHER STUDIES ...................................................................... 10
5. WHAT BROADCASTERS DO TO ADDRESS THE PROBLEM............................ 11
6. THE IMPACT OF DIGITAL TELEVISION ON THE ABILITY TO CONTROL
LOUDNESS.................................................................................................................... 12
7. THE COMMERCIAL TELEVISION INDUSTRY CODE OF PRACTICE .............. 13
8. LOUDNESS OF ADVERTISEMENTS ON OTHER MEDIA ................................... 14
9. LOUDNESS - THE OVERSEAS EXPERIENCE ...................................................... 14
10 CONCLUSION............................................................................................................... 16
Annex A ................................................................................................................................. 17
Annex B ................................................................................................................................. 19
Annex C ................................................................................................................................. 21
FACTS
Submission to the Australian Broadcasting Authority on the
Investigation into Loud Advertisements on Television
Page 3
EXECUTIVE SUMMARY
•
There is a common belief that differences in “loudness” between programs and
advertisements are caused by advertisements being transmitted at a higher volume
level than programs. This is not correct.
•
Loudness is different from volume. Volume is a measure of the level of sound at any
given instant in time. Loudness is a measure of the intensity with which we perceive
a sound. How loud something sounds depends on a variety of technical and
subjective factors, of which volume is only one.
•
When a sound track is being produced certain audio processing techniques such as
compression and equalisation can be used to enhance sound and make it seem
more “close up”.
•
Program sound tracks are not usually processed, whereas most advertisement
sound tracks are processed. When a processed sound track is intercut with an
unprocessed sound track the processed sound track is perceived as being “louder”
by comparison.
•
Broadcasters have committed substantial resources to developing systems and
techniques for the maintenance of consistent sound levels. However, because of the
higher energy content of processed sound, it will appear louder to the human ear,
even though the sound levels may be the same for both the advertisement and the
program. Broadcasters have no way of measuring the energy content, and therefore
loudness, of a sound track.
•
Of all audio processing techniques, compression is the only technique in relation to
which standards have been formulated internationally. Broadcasters attempt to
control compression in advertisements by recommending to advertisers and
production houses that the application of compression be restricted to certain levels
(FACTS Operational Practice 24). This is consistent with a similar recommendation
made in the UK.
•
The effectiveness of this approach is limited because it is not possible for a
broadcaster to measure or alter the compression that has been applied to an
advertisement. Further, compression is only one of several factors contributing to
loudness. In other countries, such as New Zealand and Canada, regulation has not
been implemented due to the subjectivity of loudness.
•
A large number of advertisements produced overseas are broadcast in Australia. For
this reason, FACTS considers that in the absence of international standards, a
regulatory solution to control loudness in advertisements is not feasible.
•
FACTS believes that a working group should be established between advertisers,
production houses, broadcasters and the ABA to work together to increase
awareness of loudness problems, to encourage compliance with FACTS Operational
Practices and to facilitate input into work already being undertaken internationally
address the issue of loudness.
FACTS
Submission to the Australian Broadcasting Authority on the
Investigation into Loud Advertisements on Television
1.
Page 4
INTRODUCTION
There is a common belief that differences in “loudness” between programs and
advertisements are caused by advertisements being transmitted at a higher volume level
than programs. This is not correct.
Loudness is different from volume. Volume is a measure of the level of sound at any given
instant in time. Loudness is a measure of the intensity with which we perceive a sound.
How loud something sounds depends on a variety of factors, of which volume is only one.
A number of other technical and subjective factors contribute to loudness.
Technical factors include the peak amplitude of the sound, the ratio of peak to average
sound energy, spectral density, the dynamic range of the sound track and the audio
processing that has been applied to the sound track, such as compression and equalisation.
Subjective factors include differences in frequencies of sound, the relative loudness of
adjoining passages of sound and the listening environment.
Broadcasters mainly use volume unit or VU meters to monitor the sound levels from the
station. In some instances, peak program meters (PPM) are used to measure the peaks in
sound levels. (Please refer to Section 3.2 for a detailed description of these meters).
However these meters are not capable of measuring loudness due to the number of factors
which contribute to loudness and the subjective factors involved. Nor is it possible for
broadcasters to control the factors which lead to loudness.
The contents of this submission is as follows:
Section 2 provides an overview of the factors contributing to loudness;
Section 3 discusses the tests available for measuring and monitoring some of the factors
that lead to loudness and their limitations;
Section 4 outlines the studies and research into loudness;
Section 5 discusses what broadcasters do to address loudness;
Section 6 discusses the likely impact of digital television on the ability to control loudness;
Section 7 addresses why regulation of loudness under the Commercial Television Industry
Code of Practice is not appropriate;
Section 8 looks at the issue of loudness of advertisements on other media;
Section 9 provides a discussion on the overseas experience; and
Section 10 details the conclusion of this paper.
2.
FACTORS CONTRIBUTING TO LOUDNESS
Following is a discussion of the factors contributing to loudness in more detail. To
understand the impact of these factors, it is instructive to outline the processes involved in
television sound.
Sound tracks for advertisements are created by production houses. The production of the
sound track involves the recording of the voice track and "layering" this with music and
effects to achieve the sound engineer's artistic aims in accordance with their client's
specification. This will often involve compression and equalisation of some or all of the audio
components to achieve the desired outcome.
FACTS
Submission to the Australian Broadcasting Authority on the
Investigation into Loud Advertisements on Television
Page 5
Broadcasters require advertisements to be recorded at a standard reference output level
(refer to para 3 of FACTS Operational Practice 24 at Annex A) This is designed to ensure
that consistent sound levels are maintained. If not at the standard level, the advertisement
may be rejected by the station or corrected before playout.
It is important that sound sources remain close to the station reference level as the transmit
channel to the viewer is bounded; if the sound is too high the sound quality will be degraded
by distortion, if the sound is too low the sound quality will be impaired by noise. Hence
broadcasters use volume unit (VU) meters to monitor the sound levels from the station and
attempt to maintain sound levels which cause the VU meter to peak at or about the 0VU
reference point. To protect against distortion and noise and to keep the station signal within
the technical limits of the ABA's licence, the broadcaster applies very mild compression and
limiting to the complete station output.
The final stage of the process occurs in the viewing environment. The viewer will adjust their
volume to suit their surroundings. Often the viewer will set their volume to understand the
dialogue above the surrounding sounds (e.g. the family talking in a living room while others
watch the television).
Technical factors
2.1
Compression
Most advertisement sound tracks are “processed,” whereas in most cases program
material is not.
Non processed sound is basically that which has originated from a microphone with
some minor adjustments to equalise the quality in certain respects such as matching
between microphones.
On the other hand, material that is processed is passed through devices such as
equalisers and compressors, which are designed to enhance its presence ie to make
it sound “up close” or brighter or more immediate. This is referred to as
“compression”.
Material that has been compressed will sound louder, even though there is no
increase in volume. This is because compression of a sound track raises the energy
content of the sound by reducing the dynamic range (i.e. the difference between the
peaks and the average level of the sound) thereby making it more dense.
2.2
Spectral processing
The human ear is not equally sensitive to all frequencies in the audible range. As an
approximation, the frequencies between 2,000Hz and 4,000Hz are those to which the
ear is most sensitive. If this range of frequencies is boosted at the mixing desk
during post production, then this version of the product will sound louder than a
version where these frequencies were not boosted. This process of varying the
spectral content is known as equalisation.
FACTS
Submission to the Australian Broadcasting Authority on the
Investigation into Loud Advertisements on Television
Page 6
A sound track which has been subject to equalisation processes will sound louder
when compared to the same sound track to which such processes have not been
applied, even though there has been no increase in volume.
2.3
Dynamic range
In most cases, sound tracks used for advertisements have a more limited dynamic
range than program sound tracks.
Dynamic range is the difference between the lowest level audio signal and the
highest level audio signal ie the range between the softest and loudest parts of an
audio signal. This range is expressed in decibels - dB. In general terms, the more
limited the dynamic range is the more “full on” or “concentrated” the experience of the
sound will be.
For example, typical classical music will have a wide dynamic range, the level
varying from the full orchestra at full volume to a solo instrument so "soft" that it can
hardly be heard - and there may be long passages at these low levels.
On the other hand, pop music generally has a dynamic range that remains relatively
"full on" all the time. It could be described as "concentrated" sound. In general, pop
music is subject to significant audio processing such as compression, limiting,
spectral processing and the introduction of synthesized sound.
Program sound tends to be more like classical music in terms of dynamic range and
may contain long passages of low levels. However, sound tracks mixed for
advertisements usually have a limited dynamic range. This makes the advertisement
sound “full on” or "concentrated".
The real problems arise when these two types of sounds are intercut ie when
advertisements are inserted into programs. In any normally adjusted system the
programs and the advertisements are transmitted at the same reference volume
setting as described in the outline of operation above - this is fundamental. A
number of objective studies have been done and they show that if anything, the
advertisement peaks “are quieter than the program”.1 However, often a more limited
dynamic range can make the advertisements sound louder.
Subjective factors
2.4
Contrast of sound tracks
Loudness is complicated by the relative loudness of adjoining passages of sound. A
noticeable difference in sound often occurs because of the considerable contrast
between two or more sound tracks.
For example, the same advertisement inserted into a pop music program and then
inserted into a softly spoken drama program will potentially produce quite different
reactions from the audiences of each program. The listeners’ perceptions will be
1
Television Advertisement Noise Levels: Assessment of Television Advertisement Noise Levels, ARUP
Acoustics, February 2000.
3
Quantifying loudness for complex sounds, Carini, J.P., Department of Physics, University of Indiana, 1998.
FACTS
Submission to the Australian Broadcasting Authority on the
Investigation into Loud Advertisements on Television
Page 7
coloured by what they were listening to before and the degree of contrast between
the two juxtaposed elements.
Advertisements in movies are probably the most problematic area to address. The
audio track of a movie is normally balanced for play-out in a wide band, high level
acoustic environment, and will have a wide dynamic range (from very soft to very
loud, depending on the plot element). Because of the “cinema” balance, much of the
audio energy may be at low or high frequencies, to which the ear is much less
sensitive at typical domestic listening levels and which are filtered out by a typical TV
set.
The movie’s sound track can therefore be perceived as quiet in the domestic
environment even when the sound peaks as read on a VU meter are higher than the
advertisements.
2.5
The listening environment
How loud an audio signal sounds depends both on the listener and the listening
environment.
When compared with the control room/edit suite at a television station or the typical
cinema situation, conditions at home are very different and listener perceptions
change accordingly.
The home situation will vary with some viewers listening on a portable TV receiver
with a built in 2 inch speaker, at a low level, while others will have a home cinema hi
fi and be listening at cinema levels. Typically however, the home environment is one
of listening at a low sound pressure level with relatively high ambient background
sound levels as opposed to environments where high sound pressure levels are
experienced by a listener e.g. concerts. This environment can lead to signals with a
wide dynamic range being masked or "lost" at the lower levels and this introduces a
strain factor to the viewing.
The modern cinema sound track has a wide dynamic range. The dialogue may well
be 15 to 20 dB lower than the music and effects. In the cinema the sound pressure
level is set for very comfortable reproduction of the dialogue - the music and effects
then appear free of distortion, at a dramatically higher level. This is suitable in the
cinema but not for the home.
If the broadcaster is showing a movie, the broadcast sound level would be adjusted
(referenced) to the music and effects (with the speech levels often 15 - 20 dB below
this level). In the case of disaster movies, for example "Twister", this difference may
even be up to 31dB. The viewer will adjust their volume in order to understand the
dialogue, so the music/effects will sound loud but will form part of an acceptable
viewing experience. Then, if the station break contains a advertisement with
compressed audio, the dialogue will be near the reference level. Hence the
advertisement will sound louder because the viewer has adjusted (increased) their
volume for the lower movie dialogue levels, not because the broadcaster has
transmitter a higher sound level.
As loudness is a measure of the intensity with which we perceive a sound, it
necessarily varies with each listener. It is subjective, depending as much on the
FACTS
Submission to the Australian Broadcasting Authority on the
Investigation into Loud Advertisements on Television
Page 8
characteristic of the individual’s ear as on the environment in which the sounds are
heard. The listening conditions and the way in which a person is listening will also
have a bearing on the subjective loudness of the sound. For instance, age is a
determinant of hearing ability with older people’s spectral appreciation often
compromised at the higher frequencies.
Recent studies disclosed that the male ear and female ear have different responses
to loud sounds, with male hearing most sensitive at 3000Hz and female hearing at
4500Hz.
The difficulty is that there can be no accurate description or definition of what is loud
in the mind of each individual listener. Industrial standards written for exposure levels
to intense sound identify where the sound level is beyond the threshold of pain.
However, the task of measuring acceptable variations in loudness is quite different,
and extremely difficult to achieve with any degree of success.
2.6
Differences in frequencies
Finally, it is a well documented fact that if the “tonal quality” of a voice corresponds to
shouting, we will hear it as loud, whereas if the “tonal quality” is that of a whisper, we
will perceive it as quiet. This will be the case even if the “measured volume levels”
are nearly identical. Studies have shown that it is difficult to come up with a single
number that quantifies the loudness we perceive when sounds of many frequencies
are present3.
3.
TESTS FOR MEASURING AND MONITORING LOUDNESS
3.1
General
Broadcasters have committed substantial resources in processing equipment to
maintain consistent audio levels to the extent possible.
There are a range of instruments used for measuring sound levels such as the
volume unit (VU) meter, the peak program meter (PPM) and a range of instruments
recently developed for digital audio.
All these instruments provide a means of technical measurement representing
precise values of test signals and a measure (however inaccurate) of the complex
sound track signal.
However, while the metering devices used to monitor sound levels can indicate the
same level for both program and advertisement material, because of the higher
energy content of the processed sound, it will appear subjectively louder to the
human ear. Broadcasters have no way of measuring the energy content, and
therefore loudness, of a sound track.
3.2
VU meters
The problem of measuring sound levels has been an issue for the broadcasting
industry from the very beginning. In the US in the late thirties and early forties an
attempt at standardising a sound level measuring instrument for broadcasting led to
FACTS
Submission to the Australian Broadcasting Authority on the
Investigation into Loud Advertisements on Television
Page 9
the adoption of the VU meter. (This was a device originally developed by the
telephone industry to measure telephone speech levels).
Today, most broadcasters balance program sound manually using VU meters.
This conventional metering system is intended to indicate the average level of a
signal. Its ballistics are designed to approximate the response of the human ear.
However it is only a very poor approximation of the actual perceived loudness and
as such cannot accurately measure loudness. It is nothing more than a blunt
instrument for this task.
The meter is an average reading voltmeter with a mechanical integration time in the
order of hundreds of milliseconds, the stepped response reaching 100% after some
300 milliseconds. If the audio signal exists for a shorter period of time than this, then
the VU meter will indicate a lesser value depending on the waveform, peak amplitude
and the duration of the audio signal. Also, this meter can continue to indicate
although the audio signal has decreased to zero. The meter will decay to zero, but
only after a finite interval of time, approximately 300ms after the audio signal has
ended. Although these characteristics are approximately the same as the ear’s
performance in the relatively restricted telephone environment, the VU meter is far
from being a precision "loudness" reading instrument for the wide range and types of
broadcast programmed material. (The only signal which can be measured accurately
on this type of meter is a steady state sine wave signal - which is what is used to
calibrate the instrument ).
It has however been in use for measuring sound levels in broadcasting for many
years because it is internationally standardised, and so allows repeatable and
standardised measurements and comparisons to be made of the electrical level of an
audio signal. This explanation tends to over-simplify the behavior of the meter but it
does give an approximate idea of its characteristics.
A programmed audio signal is a vastly more complex waveform than the hypothetical
signal discussed above. In the case of audio signals having sharp (short duration)
bursts of high energy (high amplitude), then the quasi-peak amplitude is some 6 - 8
dB above the indication that would be shown on a VU meter. The true peak value is
more likely to be 10-12 dB above that shown on a VU meter.
3.3
Peak program meters
Another type of meter is the peak program meter (PPM) which measures close to the
actual peak of the audio signal. Its ballistics are designed to have a rapid attack time
(approximately 3 microseconds) and a very slow decay time in the order of one
second. There are a variety of different scales used on PPMs (from different
manufacturers and their country of origin) and therefore it is not as closely
standardised as the VU meter.
A PPM gives no indication of perceived loudness. It can be used in conjunction with
readings from other meters to assessing whether compression has been applied to a
signal, however it can not measure the compression applied or allow a broadcaster
to alter the compression.
FACTS
Submission to the Australian Broadcasting Authority on the
Investigation into Loud Advertisements on Television
Page 10
The main use of a PPM is to indicate the amount of headroom (i.e before distortion
will be experienced) remaining in a broadcast system and is particularly relevant to a
digital broadcast network.
3.4
Loudness meters
So called “loudness meters” (as may be used for example in industrial areas) are
equally unsuitable for the purpose of broadcast level control. These are generally
very slow response devices, that are calibrated only for wideband “non-coherent”
noise (like machinery, air conditioning or traffic). They are not designed for, and
cannot be used to compare the loudness of wide dynamic range complex program
sounds.
Any instruments used for testing and measurement purposes in
broadcasting would need to be designed specifically for program audio.
3.5
Other factors
At the stations, broadcasters use a reference test signal to align the optimum
operating point of the device in question which may be a audio recorder, videotape
recorder or an entire transmission system (refer Annex C). When the advertisement
is received by the broadcaster, this reference signal is used to align the replay level
to a standard level which matches the test signal recorded point of origination of the
sound track. This does not mean that there is any direct relationship between
reference levels and dialogue or music levels.
The introduction of a limiter/compressor at the broadcast station output may help to
“smooth out” large level variations, however, this in itself is not the answer. Limiter
compressors can introduce artefacts into the audio (noise pumping for example) and
can even make things worse, by pulling down the occasional high levels of a wide
dynamic range program (like a movie) thus reducing its overall perceived loudness,
while letting the compressed narrow range advertisement go through relatively
unaffected. Another aspect of excessive compression is a problem of listener fatigue
when presented with long duration highly compressed program material. Such
mechanisms also usually result in producers engineering sound tracks for programs
or advertisements in a way designed to defeat the operation of the limiter.
4.
THE FUTURE – FURTHER STUDIES
Television broadcasters have been tracking developments in loudness monitoring and
measurement, including the development of loudness meters which are based on more
recent subjective measurement scaling techniques i.e. using audience research to test
perceived loudness.
This work is currently a subject of investigation at the ITU (refer section 9).
Any study however that attempts to compare perceived loudness is only valid for that
particular set of listening conditions. The methodology would need to account for how the
audio was presented – ie was it through a TV set or a home-theatre system, in a typical
room environment or an acoustic chamber and at what level? Ideally, a methodology would
need to address not only the perceived difference in loudness between programs and
advertisements and but also eliminate the variables of the listening environment.
FACTS
Submission to the Australian Broadcasting Authority on the
Investigation into Loud Advertisements on Television
Page 11
The methodology would need to take into account how the comparison is to be made
between program and advertisement material. Some programs and advertisements may
measure the same in level while others may exhibit a much greater difference. Would the
comparison be made to the average program level or would a comparison be made to the
last 15 or 30 seconds of program proceeding the advertisement? What duration of program
material would be appropriate and with what frequency would measurements be made? Any
process would need to consider the levels of the various items of program content and their
relationship to various types of advertisements.
It needs to be accepted that any measurement process will only provide a guide to average
or median values. Level measurements do not necessarily correlate with annoyance values
which are predominantly subjective in nature and are thus not capable of measurement.
It must also be borne in mind that any methodology for measurement of loudness would
need to reflect international standards as a number of television advertisements broadcast in
Australia are produced overseas.
The Australian Content in Advertising standard provides for up to 20 per cent of total
advertising transmission time to be available for foreign produced advertisements. In 2000,
foreign advertising averaged 7.4 per cent of all advertising on the Nine network, 10.6 per
cent on Seven and 10.5 per cent on Ten. The majority of these advertisements are
transmitted using a sound track produced overseas 4.
Without international consistency there would be a potential problem of contrasting loudness
between advertisements depending on whether they are locally or overseas produced.
5.
WHAT BROADCASTERS DO TO ADDRESS THE PROBLEM
As discussed above, broadcasters’ ability to control loudness in advertisements is limited.
Of the technical factors contributing to loudness (described in section 2), compression is the
only factor in relation to which standards have been formulated internationally. No standards
exist for the application of other sound processing techniques such as equalisation.
Broadcasters attempt to control compression in advertisements by recommending (via an
Operational Practice applying to producers of sound tracks for advertisements ) that the
application of compression be restricted to certain levels 5 (refer Annex A). This approach is
consistent with approaches in place overseas, such as the Technical Performance Code
implemented in the UK by the ITC6.
4
May 2001 ABA Update - Australian content in TV advertising: 2000 compliance report, Australian
Broadcasting Authority, 2002. www.aba.gov.au/tv/content/advertising/compliance.htm
5
FACTS Operational Practice OP 24 recommends limiting excessive compression when producing sound
tracks for advertisements by requiring that, where audio compression is used, it is restricted to 6dB, and the
onset should not occur below 0dBr (0VU). This means that audio which if unprocessed would exceed zero VU
by twice the range of the red scale above zero on a standard VU meter will, after processing, read +3 on that
meter. That represents a compression ratio of 2:1. Setting the onset of compression above zero VU ensures that
audio material reading below zero on the meter will not be compressed.
The same principle in the digital domain is contained in FACTS OP 36
6
ITC Technical Performance Code: Interface Standards and Signal Levels, Clause 22 – www.itc.co.uk
FACTS
Submission to the Australian Broadcasting Authority on the
Investigation into Loud Advertisements on Television
Page 12
In this way, broadcasters try to limit excessive loudness in advertisements. Unfortunately, its
effectiveness is limited because:
•
it is not possible for a broadcaster to measure the compression applied to a
advertisement as compression is only measurable at the time it is applied – ie at the time
of production;
•
the ability to remove compression is limited. Once the compression has been applied it
is not practical for the broadcaster to detect the compression characteristic of the sound
track and alter it;
•
compression is only one of the factors contributing to loudness.
Australia and the UK are two countries where recommendations are made in relation to the
application of compression in the production of sound tracks for television advertisements.
Elsewhere, such as New Zealand and Canada regulation has not been implemented due to
the subjectivity of loudness. Refer to Annex B for a Fact Sheet on Loudness of Commercials
available on the website of the Canadian Radio, Television and Telecommunications
Commission website at http://www.crtc.gc.ca/ENG/INFO_SHT/G3.HTM
6.
THE IMPACT OF DIGITAL TELEVISION ON THE ABILITY TO
CONTROL LOUDNESS
The transition to digital television has seen broadcasters encode sound as MPEG-1 Layer II
and Dolby Digital (AC-3) bitstreams.
In the immediate future, the introduction of digital television will have little effect on the ability
of broadcasters to control or regulate the loudness of advertisements. Digital television is a
change of the carriage medium and the issues associated with loudness still need to be
addressed prior to transmission.
Indeed, the conversion to a digital environment could exacerbate the loudness variations
due to the increased system dynamic range of sound tracks. Taking advantage of the
available dynamic range has the potential for introducing even greater level control
problems. Therefore, the regulation of loudness may become more difficult, not easier.
Metadata appears to be the method by which the level can most conveniently be controlled.
The metadata features of Dolby Digital (AC-3) may afford the broadcaster and the listener
the ability to address the relative difference in dialogue levels between programs and
advertisements. However, the application of metadata may not be the total solution either.
Metadata may assist, but the basic problem of determining what metadata value can be
assigned will continue to exist. Refer below regarding the discussion of Dolby parameters.
However, metadata is only a limited solution as it is a feature of Dolby Digital but not of
MPEG-1 Layer II and broadcasters are required by legislation to provide an MPEG audio
stream.
“Standardised" values for various types of sound material could be overcome by those intent
on getting their message across louder than others, and there will always be the perennial
FACTS
Submission to the Australian Broadcasting Authority on the
Investigation into Loud Advertisements on Television
Page 13
problem of the low level movie scene followed by the psychological interruption factor of the
advertisement break. The application of metadata is not likely to be a solution to this
problem.
The Dolby Digital technology has incorporated techniques that address perceived loudness
of dialogue. Audio coding techniques have been developed to compress the audio's
dynamic range by a predetermined amount when appropriate, such as late at night.
This technique permits the listener to automatically adjust the receiver’s output level to
produce consistent playback loudness of dialogue on all audio content, including
advertisements.
Australian television broadcasters are currently investigating the approach that can be taken
to use Dolby dialnorm, dynrng and compr parameters to best suit the overall task of
conveying and broadcasting audio material from the contribution stage through emission
stage to consumer option stage.
Issues being assessed are:
(a)
Can an ideal environment be achieved where there is no audio compression and the
listener sets that parameter?
(b)
Can Dolby metadata parameters such as dialnorm be generated at the postproduction stage and the outside broadcast in the case of live sport?
(c)
Will broadcasters encounter some sound tracks from post-production and live audio
that have no Dolby metadata parameters?
(d)
Can metadata parameters be handed across the networks in program-distribution
arrangements ?
(e)
What is the effect of metadata parameters constantly changing from segment to
segment ?
Broadcasters may be able to develop operating practices in a digital environment for the use
of metadata that will allow the playout level of the advertisement to be adjusted to reflect the
program content it abuts.
This would amount to an industry-wide recommended operational practice for the application
of Dolby parameters to material at the contribution point and set parameters for
transmission. This approach is not far removed from the prescription of FACT’S OP-24/36 –
that “Volume compression where used, should be restricted to no more than 6dB of
compression, and the onset of compression should not occur below zero VU”.
7.
THE COMMERCIAL
PRACTICE
TELEVISION
INDUSTRY
CODE
OF
In its Issues Paper, the ABA has asked; “Does the current Commercial Television Industry
Code of Practice operate to provide appropriate community safeguards in relation to this
issue?”
FACTS
Submission to the Australian Broadcasting Authority on the
Investigation into Loud Advertisements on Television
Page 14
The Commercial Television Industry Code of Practice regulates the content of commercial
television (in accordance with s123 of the Broadcasting Services Act). It does not and
should not address technical and transmission issues. Moreover, as outlined in this
submission, the issue of loudness of advertisements is largely outside the control of the
broadcasters.
Furthermore, as stated earlier, foreign advertising averaged 7.4 percent of all advertising in
2000 on the Nine Network, 10.6 percent on Seven and 10.5 percent on Ten. These
advertisement sound tracks are produced by overseas producers and advertising agencies.
Any standards put in place in Australia would need to consider how they may be imposed
upon overseas program makers and advertising agencies.
Any variance by an Australian regulation or code of practice to international standards will
need to take into account the effect this would have on the cost of importing the sound
track.
FACTS believes consideration of this issue needs to take account of international studies.
8.
LOUDNESS OF ADVERTISEMENTS ON OTHER MEDIA
The ABA has asked if the community has similar concerns about the loudness of
advertisements relative to surrounding programming on other media, particularly commercial
radio.
FACTS contends that the listening experience for the audience of commercial radio and
commercial television are quite different.
Commercial radio programming does not often resemble the characteristic of television
programming where there are long passages of program material broken up by shorter
duration periods of advertisements. While the effect on the listener is to believe advertising
sound tracks on radio are loud, they may believe that the program music (often popular
music), is also loud.
Similar problems are experienced in the cinema industry with the relative levels of film
trailers and feature material8.
9.
LOUDNESS - THE OVERSEAS EXPERIENCE
Internationally, television broadcasters have undertaken considerable investigation into
perceived loudness and proposed measurement techniques. However, in the measurement
of loudness, true standardisation is not possible due to its subjectivity and it has been
outpaced by the developments of audio signal processing and mixing technology.
This is illustrated by the US study in 1978, by the Federal Communications Commission
“Evaluating Loud Commercials: An Experimental Approach. This study dealt with partially
effective loudness monitoring and control technology. Commenting on this report in 1982,
the Commission reported “this has been a difficult issue to resolve because there are no
commonly accepted standards or techniques by which to judge loudness”.
8
“Are Movies Too Loud”, Ioan Allen, SMPTE Conference, March 1997.
FACTS
Submission to the Australian Broadcasting Authority on the
Investigation into Loud Advertisements on Television
Page 15
Australian television broadcasters instituted similar techniques.
However, while
standardised calibration of the technology is yet be achieved, these techniques can only deal
with some of the technical factors of loudness, and not the all important subjective factors
contributing to perceptions of loudness.
This was followed soon after in 1984 by the Independent Television Companies Association
releasing Technical Standards and Transmission Requirements, which included a standard
for Volume Compression.
In 1985 the then CCIR (now ITU-R) addressed the issue of “Measurement of Programme
Level in Sound Broadcasting”. This work addressed the conditions to be observed using VU
and PPM meters to ensure optimum performance in programme exchange. Again,
Australian television broadcasters developed and adopted an Operational Practice based on
this work. This was introduced in 1986.
In 1994, Dr John Emmett at Thames Television in the UK published a paper that promoted
the development of a meter for perceived loudness. Following the paper several prototype
meters were produced and trials undertaken. The designers of the loudness meters sought
acceptance of the meter in the UK and by a European Broadcasting Union group in 1994.
The EBU has not adopted this technology for standardization.
In 2000, with the establishment of Study Group 6 (Broadcasting Services) at the ITU-R the
following topics were proposed for further study up to the World Radio Conference in 2003:
•
•
•
•
•
•
Subjective assessment of sound quality
Method for objective measurements of perceived audio quality
Test signals and metering to be used on international sound programme
connections
Standards for digital audio techniques
Subjective assessment of sound quality in broadcasting using digital techniques
Subjective assessment of small, medium and large impairments in sound quality
In September 2001, the European Broadcasting Union set up a new Project “ Advisory
Group on Audio”. This group is to be chaired by Dr. John Emmett of the UKIB. The tasks to
be addressed by the group include:
•
•
Current trends used for audio production, and
Answers to inquiries on audio production technology.
FACTS is a member of the EBU and also a member of this new group. FACTS was
represented at the first meeting of the group on 20 March 2002 in Geneva. In its
presentation to the group FACTS proposed inclusion of “loudness” within the group’s project
tasks. This proposal has received the support of Dolby Laboratories. In his remarks to the
meeting, Project Group Chair, Dr John Emmett, stated the group should provide from its
investigations input to the work in the ITU-R.
FACTS
Submission to the Australian Broadcasting Authority on the
Investigation into Loud Advertisements on Television
10
Page 16
CONCLUSION
FACTS does not believe that it is possible to effectively address this issue by regulation.
The absence of internationally recognised measurement techniques means that a standard
and objective approach does not yet exist.
However, this does not mean that the issue should be ignored. FACTS believes that a
working group should be established between advertisers, production houses, broadcasters
and the ABA to work together to increase awareness of loudness problems, to encourage
compliance with FACTS Operational Practices and to facilitate input into work already being
undertaken internationally address the issue of loudness.
FACTS OPERATIONAL PRACTICE OP-24
VTR COMMERCIALS
Issue 5
February 2000
Page 17
Annex A
FACTS Operational Practice 24 – VTR Commercials
FACTS Operational Practice 24 is a summary guide to the basic requirements for television
commercials necessary to fit the contemporary operational requirements for television
presentation.
1.
TRANSMISSION STANDARD
VTR commercials shall be supplied on 25.4 mm (1 inch) C Format tape (other tape
formats may be used only by agreement with the station/s concerned.)
2.
RUNNING TIME
2.1
The duration of commercial or promotional material recorded on videotape should be
measured from the start of active program video to the end of active program video.
2.2
Sound should commence 0.5 seconds after the commencement of active program
video. This provides time for the sound channel to be opened after the start of video
without risk of sound clipping and provides aural separation between adjacent
commercials.
2.3
The end of program sound including any sound tag shall occur 0.5 seconds before
the end of active program video.
2.4
Accurate identification of first frame of active video on commercials should be
provided by the inclusion on the colour black in the 2 second interval, of a white
marker on the frame, top right of picture, outside picture safe, before first frame of
active video. The white marker should be a minimum picture height of 12 lines and a
width of 18 pixels. Timing of the commercial is referenced from the first frame of
active video.
2.5
Where commercials contain closed captions, the caption information should
commence no earlier than 1 second after the start of program video and a caption
erase signal not later than 1 second before the end of active program video.
3.
STANDARDS
In a replay machine, replay levels shall be adjusted such that a reference flux level of
100 nWb/m from a calibrated tape will produce a reference output level of 0dBr.
(This reference output level will be either + 4dbu or +8dbu and read as 0VU) [0 dbu =
0.775v] The record levels should be adjusted such that a reference level of 0dBr will
produce a fluxivity of 100 nWb/m on tape. This practice should ensure that program
peaks will not exceed a level of +8dbr equivalent to peak deviation of a television
aural transmitter. Failure to restrict this maximum level may result in audio distortion
of the recorded and or transmitted signal. Dolby type encoding shall not be used. In
stereo audio track 1 shall carry the left channel and audio track 2 shall carry the right
channel. In the case of mono the audio signal shall be on both tracks.
FACTS OPERATIONAL PRACTICE OP-24
VTR COMMERCIALS
Issue 5
February 2000
Page 18
4.
SOUND QUALITY - VOLUME COMPRESSION
Volume compression should, where used, be restricted to 6 dB of compression, and
the onset of compression should not occur below 0dBr (0VU).
Failure to adhere to this practice will lead to degradation of quality.
The use of compression purely in order to obtain a louder sound track is deprecated
and with the advent of remote controls now available on television sets may cause
viewers to reduce the volume during commercials.
Television stations reserve the right to reduce the level on transmission to ensure
that the loudness of commercial sound will match that of programs.
Other FACTS Operational Practices (OP's) relating to tape recordings which should
be observed are as follows:
OP10 2
OP13 3
OP15 1
OP18 1
OP19 1
OP28 1
25.4mm video tape interim leader, end and timing
Identification and labelling of recordings of programs and commercials
on format C video tape.
Handling of 25.4mm open reel tape.
Commercials containing closed caption subtitles. The requirement of a
"caption erase signal" requires special attention.
Alignment procedures for C format VTR's.
Digital Betacam
FACTS
Submission to the Australian Broadcasting Authority on the
Investigation into Loud Advertisements on Television
Page 19
Annex B
Fact sheet
It's a Blast!
Loudness of Commercials
You're watching your favourite television program. All of a sudden, you're
startled half out of your seat with a blast of commercial message. Believe it
or not, that commercial probably wasn't as loud as you think!
The CRTC often receives complaints about the loudness of television or
radio commercials. The issue is a difficult one to resolve, since loudness is
a subjective matter. As a result, there are no commonly accepted standards
by which to judge loudness. However, a number of Canadian and American
studies have found a few of the technical and psychological factors which
seem to be involved.
Technically speaking
"Audio compression" appears to be the most frequent cause of listener
annoyance. It's a technical method by which a commercial's sound track is
recorded at a constant and maximum loudness level. When played within a
program featuring a normal range of sounds – some loud, some soft – the
contrast can be startling. Television and radio stations are in a bind. If they
reduce their volume level to try to minimize the effect of commercials, the
range between the lowest and the highest sound levels in their normal
programming would be too restricted.
Buzzsaw barrage
"Re-shaping the sound" is another commercial production method. The
frequencies your ears are most sensitive to are raised in level, while other
frequencies (those which can make sounds more pleasant to listen to) are
reduced. The resulting commercial now features a sound track that grabs
your attention, but may also sound harsh – especially when combined with
the audio compression technique.
Faulty level
A faulty audio level setting at the local station is a possibility, but most radio
and television stations take care to ensure their programs and commercials
don't exceed the maximum level set by Industry Canada's Broadcasting
Regulation Branch. If you think a faulty level setting is to blame, give the
station the specific date, time of occurrence, and commercial involved. Ask
them to check the audio levels being fed to their transmitter. A Regional or
FACTS
Submission to the Australian Broadcasting Authority on the
Investigation into Loud Advertisements on Television
Page 20
District office of Industry Canada can also be asked to investigate.
Rude awakening
Even if a commercial doesn't use the audio compression or frequency
reshaping techniques, it could still give the impression of being louder than
the program. Sometimes a commercial intrudes on an otherwise tranquil
scene in a program. The contrast can be startling – even though the
commercial is not really excessively loud.
This document is available in alternative format upon request.
G3-07-97
FACTS
Submission to the Australian Broadcasting Authority on the
Investigation into Loud Advertisements on Television
Page 21
Annex C
Listening
environment
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

advertising