Addendum 34 rev1 Sony HVR-Z5E and Z7E

Colorimetric and Resolution requirements of cameras
Alan Roberts
ADDENDUM 34 rev1 : Assessment and settings for Sony HVR-Z5E and
HVR-Z7E
Data for this addendum is taken from a short examination of production model of each of the Sony Z5 (serial
number 00001, clearly an early model) and Z7 HDV (a retailed model) camcorders and from their manuals.
The cameras are nearly identical, therefore it makes sense to report on them together.
They are HDTV camcorders, physically very similar to the Z1, with 3 ⅓” cmos sensors, but the Z7 has an
interchangeable lens with a Sony proprietary lens mount while the Z5 has an integral Sony lens. The
manuals claim that the sensors are approximately 1.12Megapixels, delivering resolution up to1440x810, with
approximately 1,037Megapixels used for 16:9 capture. They both record HDTV using the HDV algorithm
onto standard DV tapes (1080i and 1080psf) or Compact Flash solid-state cards, and SDTV using DVCAM
or DV format. The similarity to the Z1 is only superficial.
The cameras are relatively light (about 2.8kg in including lens and battery) and have an integral viewfinder,
with side lcd panel, and seem aimed at the high-end professional market rather than full broadcast, which
would normally demand an image format larger than ⅓”. The lens mount of the Z7 is specific to the
⅓”format, but adaptors are available to permit the use of ½” lenses. The Z7 was tested using the supplied
Zeiss 12:1 zoom lens.
There are internal menus for setting the performance, not as complex as in a full broadcast camera, but
enough to control some of the important features, albeit only in “on/off” states. They are not well suited to
multi-camera operation since they cannot be remotely controlled. There are analogue-only video outputs
(components and SD-composite via a multi-pin connector and S-video SD) and digits via IEEE1394 Firewire
(known as i.Link by Sony) and 8-bit HDMI. A broadcast camera would normally be expected to have a 10bit HDSDI output.
The same assessment procedure was used as for other HD cameras, partly attempting to get a good “filmlook”, and the settings reflect that. However, because of the lack of internal test signals, it was necessary to
make more complex measurements than normal, through the lens. Since many camera parameters are
undefined in the specifications, more measurements than usual were necessary. In the search for a “filmlook” setting it is normal to think of the camera to be mimicking a film camera and telecine, with “best light”
transfer to tape, with about 10 stops of tonal range. Assuming that a grading operation will be used in postproduction, the settings attempt to give the colourist the same range of options as with film, but without
achieving the full 10-stop dynamic range. The recommended settings allow about 1.7 stops of over-exposure
and one of under-exposure relative to normal operation. This is not as good as can be achieved in ⅔”
cameras, and arises from the difference in pixel size (the pixels here are much smaller, so sensitivity is
maintained at the expense of highlight handling and video noise).
1
Colorimetric and Resolution requirements of cameras
Alan Roberts
ADDENDUM 34 rev1 : Assessment and settings for Sony HVR-Z5E and
Z7E
Data for this is taken from tests on production models of the Sony HDV camcorders, HVR-Z5E and Z7E.
They are near-identical camcorders with three ⅓” cmos sensors (5.9mm diagonal, each approximately
1440x810 with precise half-pixel offset of green from red and blue). They record with HDV coding
(1080i/25, 50Hz interlaced or progressive, and a proprietary 1080p/25 format) onto miniDV tapes, and
standard definition (576i/25) as either miniDV or DVCAM. They can also record onto Compact Flash cards.
The cameras are essentially professional models, having some professional features such as having XLR
connectors at mic or line level. The Z5 has an integral lens while the Z7 has interchangeable lenses, both
have integral viewfinder and side-top lcd panel. Resolution is good enough for HDV, but vertical resolution
could be better. Minimum exposure is claimed to be 1.5 lux.
There are many internal menus for setting the performance and a reasonable selection of external controls.
There are analogue video outputs (components and composite plus S-video at SD, all via multi-pin
connectors) and digits via IEEE1394 Firewire (known as “i.Link” by Sony) and HDMI. All measurements
were made using the 8-bit HDMI output, converted to 10-bit HDSDI for image monitoring and capture.
The normal assessment procedure for full broadcast cameras could not be used, largely because the Z7 does
not have a selectable test signal. Therefore, testing had to be done the hard way, via the lens. Recommended
settings allowing for a “video-look” and a “film-look” have been derived, although there are some significant
compromises that have to be taken into account.
It is useful to think of the camera, when used with “film-look”, to be mimicking a film camera and telecine,
with “best light” transfer to tape. Measurement results are given after the settings tables, in order to explain
the decisions. At best, the camera can deliver about 10 stops of exposure range, similar to other HD
cameras, but it is easy to set the camera such that exposure range drops to 7 stops or less. In the target
market for this camera, a grading operation may well not be used in post-production, so the settings should
be used with care.
The HDV performance is acceptable at sub-broadcast level for HDTV production, but is not up to the
standards expected for full HDTV broadcast. There are a few problems with the picture performance as an
SD camera for professional or broadcast purposes, although both cameras are better than the HVR-Z1 both in
HD and SD modes. Performance with the recommended settings is probably adequate for consumer and
semi-professional use, but better performance can be expected with a professional hardware or software
down-converter. The reasons for this statement are given in the measurements section of this document.
The controls for these cameras are not as flexible as for full “broadcast” cameras, so more effort was
expended in measuring performance than in trying to derive a specific “look” for it. Very small physical lens
apertures (less than F/4) soften the picture and produce visible colour-fringing due to diffraction effects in
the iris, the included neutral density filters are the better alternative to small apertures when shooting in very
bright light.
Many of the menu items have little or no effect on image quality. Those that have significant effect are
highlighted. The full set of menu items is given for completeness. In boxes with a range of numeric settings,
the values indicate the range, and no scales are given. The numbers represent the count of bars in the
thermometer presentation from the left, usually 1 to 16 with 8 being the central (default) value. Default
settings, where known, are underlined. My recommendations are in the last column, labelled “BBC”, where
appropriate. Settings are given for:
v
f
Television production
Film-look television
2
In the tables, items that have an important effect on picture appearance are highlighted with grey
background. Rather than just making assertions about performance, I have included measurement results
that illustrate the reasons for recommending settings. Virtually all picture control is in the Profile menus.
Note that, in each power-switch mode, the menus can be separately customised, adding or removing any
menu item from the entire set of menus.
This is not intended as a replacement for reading the manual.
1
Switches and Menu settings
The test session on the Z5 was very brief, and there was no manual available either on paper or as a
download. The names and positions of controls here are for the Z7, but are expected top be accurate for the
Z5 as well, but I cannot guarantee that. Also, menu items are taken from the Z7 manual; where the Z5
differs from the Z7, items are identified as such.
SWITCHES and BUTTONS
name
Zoom
Expand focus
Iris
Push Auto
Rec Start/Stop
ND filter
Assign 1 to 3
Assign 4
Assign 5
Assign 6
Gain
Audio 1/2
Auto/Manual
Shutter Speed
White Balance
Sel/Push Exec
White balance stores
White Balance
Gain
1394 connector
Analogue component
A/V Out
LANC
Headphones
Zoom
Zoom
Rec Start/Stop
Status Check
Picture Profile
Menu
Shot Transition/Focus Marking
A
B
Zoom
Digital Extender
TC/U-Bit
Tape controls
Volume/Memory
Memory/Delete
Memory/Index
Memory/Play
Reset
place
Right
Right
Right
Right
Right
Left
Left
Left
Left
Left
Left
Left
Left
Left
Left
Left
Left
Left
Left
Back
Back
Back
Back
Back
Handle
Handle
Handle
Left
Left
Left
Top
Top
Top
Left
Left
Handle
Handle
Handle
Handle
Handle
Handle
Handle
feature
Rocker
Push
Switch
Push
Push
Push
Push
Push
Push
Push
Switch/dial
Switch
Push
Push
Dial
Push
Switch
Switch
Socket
Socket
Socket
Socket
Socket
Rocker
Switch
Push
Push
Push
Push
Push
Push
Push
Switch
Push
Push
Push
Push
Push
Push
Push
Push
3
comment
User buttons
Zebra/Assign 4
AE Shift/Assign 5
Rec Review/Assign 6
Manual control/indicator
Audio level controls
Menu controls
Cycles through white balances
Proprietary format mini-connector
Another proprietary mini-connector
Remote control
Enables handle zoom rocker
Most of the important camera controls
Multi-function transition control
Condition A settings
Condition B settings
Auto/Manual
Lens range extender, not a good idea
Toggles between timecode and user bits on lcd
The usual tape controls
Tape play volume, or select still to view
For viewing stills
For viewing stills
For viewing stills
Recessed, factory reset
Basic camera settings
CAMERA SET menu
item
Iris/Exposure
Ring Rotate
Gain setup
Gain setup
Smooth Gain
Hyper Gain
AGC Limit
Minus AGC
WB preset
AGC Limit
Black Balance
WB preset
Z5
Z5
Z5
Z7
Z5
Z5
Z5
Z7
Z7
Z7
WB Outdoor Lvl
WB Temp Set
ATW Sens
Smooth WB
AE Shift
AE Window
AE response
AT Iris Lmt
ECS Freq.
Flikr reduce
Cntrst Enhcr
Back Light
Spotlight
Steadyshot
AF assist
Focus Macro
Handle zoom
Speed Zoom
D.Extender
Fader
Smth Slw Rec
Interval Rec
DV Frame Rec
S.Trans
Trans Time
Trans Curve
Start Timer
Rec Link
S.Trans/F.Mark
x.v.Color
Color Bar
Flange Back
Z7
Z5
Z7
Z5
Z5
Z5
Z5
Z5
Z7
Z7
Z7
range
Iris, Exposure
Normal, Opposite
-6, -3, 0, 3, 6, 9, 12, 15, 18, 21dB
0, 3, 6, 9, 12, 15, 18, 21dB
Fast, Middle, Slow, Off
On, Off
21dB, 18, 15, 12, 9, 6, 3, 0, -3, -6,
Off
On, Off
Outdoor, Indoor, Manu WB Temp
21dB, 18, 15, 12, 9, 6, 3, 0, Off
Exec
A, B, Outdoor, Indoor, Manu WB
Temp
-7~0~+7
2300~15000K
Intelligent, High, Middle, Low
Fast, Middle, Slow, Off
-7~0~+7
Standard, Type1~5
Fast, Middle, Slow
F11, F9.6, F8, F6.8,F5.6, F4.8, F4
50~200Hz
On, Off
On, Off
On, Off
On, Off
On, Off
Hard, Standard, Soft
On, Off
On, Off
1~3~8
On, Off
On, Off
White Fader, Black Fader
comments
Allows direct control of iris
BBC
Set gain for each position of the gain switch
Speed of gain change, Off=instant
On disables Backlight/Sptolight compensations
Upper limit for video auto gain
Allow AGC to go to -6dB
Upper limit for video auto gain
Temporary black balance, lost at power-off
Select A/B preset then push to balance
Colour offset for Outdoor
Colour temperature for Manual
Auto white balance, High reduces colour errors,
Low increases them
Speed of white balance response when changed
Drive auto-exposure, -7=dark, +7=light
Various window mask shapes for auto-exposure
Speed of auto response
Maximum aperture limit for auto exposure
For ECS shutter, slowest is 25 for Progressive
Supposed to reduce lighting flicker
Claims to improve high contrast scenes, e.g. backlit
Compensate for back-lighting
Compensate for spot-lighting
Set Off when on a tripod
On, Off
Use focus ring for fine focus adjust in auto-focus
Focus within 80cm
Zoom speed for FIX position, 1=slow, 8=fast
Allows crash zoom
Digital zoom to 150%
Shot fades to/from black/white
About 4x cranked, see manual for details
Stop-frame recording, see manual for details
Records about 5 frames/push of Rec Start/Stop
3.5 ~ 15 sec
Soft, Stop
Off, 5, 10, 20
Off, Shot A, Shot B
Shot Transition, Focus Marking
On, Off
On, Off
Type 1~4
Auto Adjust, Manu Adjust
Assign Shot Trans/Focus Mark button
Wide-gamut colour, only for compatible displays
Type 1=SMPTE, Type 2=ARIB,
Type 3=100% full, Type 4=75% full
Back focus setting3
F5.61
Off
Off
1 or 22
AUDIO SET menu
item
DV Au.Mode
Audio limit
Int Mic Set
Mic NR
range
FS32k, FS48k
Off, On
Z5
Z5
comments
Audio sample rate, SD
Only when in Manual audio level
Internal mic controls
Noise reduction
On, Off
BBC
________________________________________________________________________________________________
1
In 1”/3 sensors, iris diffraction starts to be visible at about F/5.6, i.e. the picture gets softer and has chromatic
aberration effects.
2
SMPTE bars are ubiquitous, but ARIB bars have some advantages, agree the type of bars to be used with the postproduction people before shooting.
3
Backfocus setting should always be done when changing lenses, or when the camera temperature change significantly.
The Zeiss lens has a calibrated back setting, so it may be possible to set the lens off the camera.
4
Mic Sens
Mic Wind
XLR set
XLR AGC link
AU. Man Gain
Input 1 Mic NR
Input 1 trim
Input 1 wind
Input 2 Mic NR
Input 2 trim
Input 2 wind
Separate/Linked
Audio Ch sel
DV Audio Mix
Z5
Z5
Normal, High
On, Off
Z7
Z7
Separate, Linked
Separate, Linked
On, Off
-18, -12, -6, 0dB, +6, +12dB
Off, On
On, Off
-18, -12, -6, 0dB, +6, +12dB
Off, On
Separate, Linked
Ch1Ch2, Ch1, Ch2
Ch1Ch2, Mix, Ch3Ch4
Z5
Sensitivity
Wind noise filtering
Settings for external sources
Link for stereo (valid for Auto level only)
Link for stereo (valid for Manual level only)
Noise reduction for Mic input, not Line level
Mic level
Wind noise reduction
Noise reduction for Mic input, not Line level
Mic level
Wind noise reduction
Links channels for AGC
Ch1Ch2=stereo, Ch 1 /2 =input Ch to both
Monitoring
Viewfinder and LCD
DISPLAY SET. menu
item
Zebra
Level
Histogram
Peaking
Colour
Level
Marker
Centre
Aspect
Safety Zone
Guideframe
Cam Leveling
Exp. Focus Type
Cam Data Disp
Au. Lvl Disp
Z5
Zoom Display
Focus Disp
Shutter Disp
LCD Bright
LCD color
LCD BL level
VF B.Light
VF color
VF Power Mode
Letter Size
Remaining
Disp Output
range
On, Off
70~100,100+
Off, Normal, Advance
On, Off
White, Red, Yellow
High, Middle, Low
On, Off
On, Off
On, Off
On, Off
On, Off
On, Off
Type 1, Type 2
On, Off
On, Off
Bar, Number
comments
Exposure metering
Signal level, %
Advance adds a marker for average video level
Help with focus, does not get recorded
Colour of enhanced edges
Sensitivity
All sorts, not available when Date Rec is on
Small centred cross
Boundaries at 4:3, 13:9 or 14:9
Edge at 80% or 90%
Grid
75 {f}, 100{v}
14:9
4
Expanded focus, Type 2 is in mono
Shows camera settings (gain, shutter etc)
Audio level meters, stereo
Lens focal length, Number is 0~99, not much
help
Focus distance
360 degrees=1/frame rate
Side lcd brightness
Saturation
Brightness, always Bright on external power
Brightness, always Bright on external power
Set v/f to monochrome
Auto switches vf off when lcd is open
Screen text size
Auto shows length of tape left on Play or
Batt/Info
All Output sends everywhere (VF, LCD,
Analogue, HDMI, the lot)
Meter, Feet
Second, Degree
Normal, bright
Normal, bright
On, Off
Auto, On
Normal, 2x
Auto, On
LCD Panel, V-Out/Panel, All
Output
VTR matters
IN/OUT REC menu
item
Rec format
range
HDV1080i, DV
VCR HDV/DV
Auto, HDV, DV
HDV Progre.
Rec Type
Scan Type
DV Progre.
Scan Type
DV Rec Mode
DV wide rec
Video Out
Component
BBC
comments
Recoding format
iLink/Frewire connection format, disconnect to
force system to set itself correctly
Progressive scan format in HD
5
Interlace, Progressive
50, 25
BBC
Interlace
25 {f}, 50{v}
Progressive scan format in DV
50, 25
DVCAM, DV SP
On, Off
25 {f}, 50{v}
SD recording format
16:9/4:3 recording (DV only)
Control of component analogue output
Set according to TV set type
576, 1080i/576i
________________________________________________________________________________________________
4
If this is what I think it is (i.e. I didn’t check it) its hugely useful, a 2-d spirit level. I wonder…..
5
Rec Type=Interlaced is the normal HDV recording format, whether progressive (psf) or interlaced (i). Rec
Type=Progressive is a proprietary format for progressive recording, and will not play on all HDV players.
5
Downconvert
Squeeze, Letter Box, Edge Crop
DV Wide Conv
i.Link Set
HDV-DV Conv
Down Convert
Ext Rec Ctrl
Squeeze, Letter Box, Edge Crop
For SD from HD: component, S-video,
composite
For SD playout
Control of IEEE1394 Firewire output
Set downconversion via 1394
On, Off
Squeeze, Edge Crop
Rec Ctl Mode
Control of external 1394 recorder
Synch records internal/external, Relay goes
external when internal tape is full
Stop mode for external recorder
Off, Synchronous, Relay, Ext Only
Stby Command
Rec Pause, Stop
Timecode etc
TC/UB SET menu
item
TC Preset
range
Preset
Reset
TC Countup
UB Preset
Preset
Reset
TC Run
TC Make
TC Link
UB time rec
UB-Date/TCTime
comments
Set TC, see manual
Enter time code, use Sel/Push Exec dial
Reset to zero
Add 1 hour, minutes=0, when TC Make=Preset
Set User Bits, see manual
Set user bits with Sel/Push Exec dial
Reset to zeroes
Free run is real time
Regen sets Rec Run
Sync on multiple cameras, see manual
On sets real time in User Bits
Sets date and time into User Bits
Exec
Exec
Exec
Rec run, Free run
Regnerate, Preset
Off, On
Memory Stick
MEMORY SET menu
item
All Erase
Format
File No.
New Folder
Rec Folder
PB Folder
BBC
range
All Files, Current Fldr
Exec
Series, Reset
Exec
Exec
Exec
comments
Pretty obvious
Wipes the lot
Series increments file numbers, Reset starts again
Each folder can hold 9,999 images
Select recording folder
Select folder to play back
BBC
comments
Profiles. Up to 99 on a stick, 2 in the camera
Sel/Push Exec to load a profile
Sel/Push Exec to save to stick or camera
Sel/Push Exec, select profile and edit the name
Sel/Push Exec to select, then delete it
BBC
OTHERS menu
item
Camera Prof.
Load
Save
Change
Delete
Copy
range
Z5
Assign Buttons
Assign Buttons
Clock Set
World Time
Z7
Exec
Exec
Exec
Exec
Exec
Focus, Exp Focus, Focus Macro, D
Extender, Ring Rotate, Hyper Gain,
AE Shift Push AT Iris, Index Mark,
Steadyshot, Back Light, Spotlight,
Fader, Color Bar, Smooth Slow
Rec, Last Scn Rvw, Rec Review,
End Search, Zebra, Marker,
Peaking, Display, TC Reset, TC
Countup, Photo, Picture Profile 1~6
Focus, Exp Focus, One Push, Focus
Infinity, Focus Macro, D Extender,
Hyper Gain, Blk Balance, AE Shift,
Index Mark, Steadyshot, Back
Light, Spotlight, Fader, Color Bar,
Last Scn Rvw, End Search, Zebra,
Marker, Peaking, Display, TC
Reset, TC Countup, Photo, Rec
Lamp(F), Rec Lamp(R), Picture
Profile
Assign any to buttons 1~6
Factory defaults are:
L1 Button=D.Extender
Button 4=Zebra
Button 5=AE Shift
Button 6=Rec Review
Assign any to buttons 1~6
Factory defaults are:
L1 Button=D.Extender
Button 4=Zebra
Button 5=AE Shift
Button 6=Rec Review
This comes up every time the camera powers up
until you set the time/date
Select local time relative to original setting
6
Language
PB Zoom
Quick Rec HDV
Date Rec
Beep
Rec Lamp (F)
Rec Lamp (R)
Remote Ctrl
Hours meter
How do you get back if you select a language you
can’t read? ☺
Z7
On, Off
Off, On
Use zoom lever to zoom on playback, up 1.5
On is quicker, but breaks the MPEG GoP
structure, may not work with some NLEs
Burns time/date onto recording
Warning sound on start/stop recording
Front recording lamp
Rear recording lamp
Enables remote control
VTR hours meters display
Off, On
Off, On
On, Off
On, Off
On, Off
7
PICTURE PROFILES menus, default settings
item
PP1
PP2
PP3
PP4
PP5
PP6
range
comments
Camera control
BBC
User=Normal
User=Normal
Pro Color=Professional camcorder, ITU709 gamma
PD Color=Handy cam (I guess PD170)
Film Look 1=Colour Negative
Film Look 2=Colour Print
PICTURE PROFILES menus, manual settings
item
Black Level
Master Black
Black R
Black G
Black B
Gamma
range
-15~+15
-15~+15
-15~+15
-15~+15
Standard, Cinematone1,
Cinematone2, ITU709, G5.0,
PD, x.v.
comments
Camera control
BBC
No calibration, cap the camera and use waveform
monitor or Histogram to set black levels.
RGB values are added to Master level
G5.0 is BBC 0,4 law
Black Gamma
ITU7096
Black stretch
Range
Level
High, Middle, Low
-7+7
- is black compression, + is stretch7
Compress overexposure
Knee
Mode
Auto Set
Max Point
Sensitivity
Manual Set
Point
Slope
Color Mode
Type
Level
Color level
Color phase
Color Depth
Color Correct
Type
Memory Sel
Mem1 Color
Mem1 Revisn
Mem2 Color
Mem3 Revisn
WB shift
Filter Type
LB Col Temp
CC MG/GR
R Gain
B Gain
Detail
Level
Manual Set
V/H Balance
0
Auto, Manual
Manual
90%~100%
High, Middle, Low
90%8
Middle
75%~105%
-5~+5
87.5%
-29
Standard, Cinematone1,
Cinematone2, ITU709 Mtx
1~8
-7~+7, -8
-7~+7
-7~+7
Off, Color Revisn, Color
Extract
1, 2, 1&2
LB-CC, R-B
-9~+9
-9~+9
-9~+9
-9~+9
ITU709
Cross-fade, 1=Standard, 8=what you selected
Saturation, -8=monochrome
Greenish to reddish
6-axis multimatrix, set saturation for RGBCMY
See manual, rather complicated
3
0
0
010
Off
2 colour stores
Define colour in Memory 1
RB gain modification to colour in Memory 1
Shift aim point, - for blue, + for red
LB-CC works on secondaries, RG on primaries
Blue to red
Green to magenta
-7~+7
On, Off
Set balance of horizontal to vertical detail
0 {v}, -2{f}
On
0
________________________________________________________________________________________________
6
709 gives the most accurate colour rendition. The Cinematone curves give a more film-like appearance, but actually
capture significantly less contrast.
7
If Standard gamma is used, some black stretch may be a good idea, to improve colour fidelity.
8
Although performance is usually best with manual knee set, there are clearly advantages to using automatic for
simplicity. These settings should be acceptable.
9
These settings slope the knee point to reach 250% at peak white. Thus captures the biggest contrast range the camera
can do.
10
Colour controls are best though of as a ‘special effect’, which is usually best done in post production where more
control and better monitoring will deliver best results.
8
B/W Balance
Black Limit
White Limit
Crispening
Hi-Light Dtl
Skintone Dtl
Level
Color Sel
Phase
Range
Saturation
Reverse
Y Level
Y Range
One Push Set
Profile Name
Copy
Reset
Type1~5
0~+7
0~+7
0~+7
-2~+2
On, Off
1~8
0~32~64~96~127
0~31
0~31
0~31
1~32
Control positive-/negative-going detail
Limit –ve going detail, 0=limit, 7=not
Limit +ve going detail, 0=limit, 7=not
Detail enhancement in high signal levels
See manual, rather complicated
1=less detail, 8=more
Define the colour to be tweaked
0=purple, 32=red, 64=yellow, 96=green, 127=blue
0=off, 1=narrow, 31=wide
0=mono, 31=max saturation
Reverses the colour selection, i.e. do all the others
0=dark, 31=bright
Brightness range, 1=narrow, 32=wide
Auto adjust colour at centre marker
Set a profile name
Copy one profile into another
Factory reset this profile
Type 4
3
7
0
+2
Off11
________________________________________________________________________________________________
11
Skin tone detail may be useful on occasions, but can be tricky to set and deliver unexpected results as lighting levels
change. Use with care.
9
2
Measurements
All measurements were made on frames captured via the HDMI output via a 10-bit HDSDI converter box
(Focus Enhancements MCSDI-1: this does not filter the signals, and delivers a 10-bit HDSDI signal), to
avoid the known limitations of the recording formats. The HDMI signal from the camera delivers only 8-bit
data, but this was not considered to have affected the measurements in any significant way. Although not the
usual practice for camera tests, this image capture process was far easier to manage, and did not lead to any
confusion or misrepresentation of the camera performance.
2.1 Colour performance
The gamma curves were not explored in detail, because there is little that can be done about them. Instead, a
calibrated Macbeth chart was used, correctly exposed with studio illumination. The pictures were displayed
on a HD monitor and compared with another chart illuminated at D65. Frames were grabbed for analysis as
well, but not used because the visual examination provided the expected results.
With the ITU709 gamma-correction curve, which should be correct for HDTV, and Color Mode (i.e.
matrix) set to ITU709 the colours were all somewhat oversaturated, too colourful. The cure was to set the
Color Level to 3, apparently taking 3/8 of the ITU709 matrix and 5/8 of the Standard matrix. Colour
performance was then judged to be quite accurate, and can be expected to be very similar to the G5.0
gamma-correction curve (approximately the BBC 0.4 law). This combination delivered about 1.7 stops of
overexposure, making a total exposure range of about 10.5 stops with the recommended Knee settings.
The Cinematone curves were not investigated here, but are almost certainly those of the Z1. Cinematone 1
should deliver about 8.9 stops, while Cinematone 2 is more gentle and delivers about 8 stops. Both these
curves are appropriate for shooting where grading is not expected, but scenic contrast will effectively be
abandoned at the time of shooting. The recommended settings will always capture the maximum contrast,
although grading will be needed to achieve any specific “look”. The user is strongly advised to make tests
before using these or any other settings
Black Gamma raises the gain near black by a significant amount, extending the exposure range by between
0.5 and 0.8 stops while increasing video noise levels.
The Standard curve (Normal in the Z1) will deliver good colour rendering using only about 45% of the
sensors’ exposure range (the normal television mode), reserving the remainder of the range for highlights,
significantly compressed.
If the Standard curve is to be used, some Black Stretch would be a good idea if the intention is to capture a
large dynamic range and use post-processing to achieve a film look, while Cinetone curves should be used if
the intention is to produce a film look without further processing, but greater attention will have to be paid to
getting exposure levels exactly right.
None of the colour adjustment/correction controls were investigated, they are best thought of as “special
effects”, trickery.
2.2 Resolution, 1080-line
The camera specification says that it has sensors of 1.20 Megapixels, and claims dimensions of 1440x810.
At this point, it makes sense to speak of photo-sites, rather than pixels, since pixels belong in the picture and
not in the camera. A photo-site is equivalent to a picture element in the sensor. The relationship between
camera photo-sites and image pixels is rather more complex than in most cameras.
The sensors are rotated 45 degrees, and images interpolated from the now diamond-shaped photo-sites. Thus
the resolution delivered is only loosely connected with the actual ‘pixel’ count, and there is probably a
boundary of blanked-out cells to provide information for black level control as well. The advantage of this
rotated arrangement is that the coloured aliasing that results from the normal ‘precision offset’ arrangement
(whereby the green sensor is placed exactly half a photo-site spacing from the red and blue to achieve extra
resolution) is avoided, at the expense of diagonal aliasing in the luma signal.
10
The specification also says that there are 1,037,000 effective photo-sites for 16:9 video capture. If they are all
square (placed on a 3.333µm grid), then the effective image dimensions for 16:9 capture would have to be
about 1358x764 (1358x764=1,037,352). To confirm this, the specification also says that the effective
number for 4:3 capture is 778,000, and 1358x3/4=1018, so the dimensions for 4:3 become
1018x764=778,134.
A circular zone plate test chart was used to
explore resolutions up to exactly
1920x1080. Each pattern limits at exactly
1920 horizontally and 1080 vertically,
frequencies being proportional to distance
from the centre of the pattern.
Measurements were made on the Z7, with
confirmation tests done on the Z5.
Figure 1 Factory
settings, interlaced
1080
Figure 1 is one quadrant of one pattern,
exploring the frequency response from dc
in the centre spot to 1920 horizontally, and
1080 vertically, at the outer rim, frequency
being proportional to distance from the
centre. Unfortunately, during the test
session the chart was not quite correctly
framed at the time of image capture, and a
correction must be made to allow for this,
a zoom factor of 1.035, i.e the maximum
frequencies explored are 1987x1118 in this
case.
The capture was made with factory
settings, interlaced. It shows null zone at 71.6% of the horizontal maximum, 1424. This suggests that the
effective active area of the sensors is actually 1424 photo-sites wide, and therefore 801 high (since
1424/801=16/9), confirming the calculations in the paragraphs above.
Horizontal frequencies beyond the null
show some coloured aliasing, which
indicates that the green sensor is offset
from red and blue by half a photo-site even
in this rotated arrangement. Vertically,
there is no clear null zone, as is to be
expected for an interlaced image.
However, frequencies above 800 are
clearly aliased, but not excessively so.
But, there are strong diagonal aliases,
which are common in cameras using a
single Bayer-patterned sensor, and in 3sensor cameras where the precision offset
of the green from red and blue is done both
vertically as well as horizontally, a
quincunx arrangement. The presence of
these diagonal aliases confirms that there
is no optical filter in the camera between
the lens and the sensors. In broadcast
cameras, there are normally two birefringent filters, to suppress the higher
horizontal and vertical frequencies that
would cause aliasing in the camera. In this
camera, the filters would have to be
Figure 2 Detail set
for Video,
interlaced 1080
11
arranged to limit the diagonal frequencies rather than horizontal and vertical. The absence of this filtering
places the camera out of the broadcast HDTV field.
Figure 2 shows the recommended Detail settings for Video, the aliasing is more pronounced, but is not
excessive. However, any further increase in the Detail settings produces little effective sharpening of the
image, but worsens the appearance of the aliasing. Detail should be used with care, since aliased spatial
frequencies move in the opposite direction to image motion and cause confusion in motion-sensitive
compressors, such as MPEG, effectively consuming bit-rate unnecessarily.
The Detail settings for Film soften the
pictures a little, and reduce the visibility
of the aliases. The result is not shown
here.
Figure 3 psf
Figure 3 shows the performance when the
camera was set to Rec Type Interlaced,
Scan Type 25 frames/second (the standard
nomenclature for this mode is psf,
progressive with segmented frames).
Detail was set for Film. Clearly, there is
more vertical detail, mixed with coloured
aliasing. This is the normal shooting
mode for a film look, and is recorded as
interlaced field-pairs although they both
belong to the same temporal exposure. In
this mode, the signal will travel through
recording, editing, and transmission as
though it were interlaced. It is only in the
display that the field-pairs (frame
segments) come back together to form a
displayed frame. However, there can be
some confusion in editing, as to which
field-pair belongs together, the setting for
this being known as field-dominance.
Figure 4 Progressive
Another problem with psf recording is
that each field or frame-segment is
separately compressed as an entity. For
interlaced images, this is normal, but for
progressive images it should be more
efficient to compress the entire frame in
one go, as is done for images at 720p.
This camera has another mode, Rec Type
Progressive, in which the entire 1080-line
image is compressed as one.
Figure 4 shows the result. There is no
visible
difference
in
resolution,
confirming that there is no image quality
advantage in using this mode, apart from
the slight advantage in recording, which
should result in a slightly lower level of
compression artefacts.
Note that
recordings made in this mode can not be
played back in a standard HDV player,
only in one compatible with this
12
recording mode.
2.3 Resolution, 576-line
Figure 5 576-line SD
The cameras can be operated in SD mode,
therefore down-converting the HD signal
from 1080-line to 576. Figure 5 shows the
result on the same quadrant of zone plate.
Clearly, some of the high-frequency image
content, which should have been suppressed
in the down-conversion, remains in the
output, albeit aliased. There is a null centre
at 1440 horizontally, from the sensor pixel
dimensions, and diagonal aliasing at nearhorizontal frequencies. Interestingly, there
is no symmetrical aliasing near the vertical
axis, presumably due to the interlaced
nature of the output signal.
Vertically, there is a second-order alias,
where frequencies have been folded twice,
the zone centre shifted to 576.
Overall, the performance at SD is not
particularly good, significantly better
pictures should be obtained by using a
proper, external, down-converter. I cannot recommend using this camera in SD mode, because of the level
of these spatial aliases. However, it is significantly better than the Z1 in this respect.
2.3.1 Lens aberrations
The Zeiss lens supplied with the Z7
showed some chromatic aberration in the
corners.
Figure 6 Lens
aberration, Z7
The spatial offset is about 4 pixels, not a
very good performance for an HDTV
lens. The camera’s performance should
be considerably better with a ½” format
lens, not through increased sharpness
(because lenses for larger formats do not
have to be able to pass frequencies up to
200lp/mm to the sensor) but through
better aberration performance. At the
time of writing this document, there was
only one ⅓” lens available for the
camera, the one supplied with it.
The integral lens in the Z5 showed a
similar degree of chromatic aberration,
albeit in different colours and directions.
13
2.4 Video noise
The camera manual gives no specification for video noise level, but video noise plays a crucial part in prosproduction operations such as matte-ing,
keying and colour manipulation.
-39.0
For noise level measurements, a plain
white card was evenly illuminated, and
captured
at
1080-line
resolution,
interlaced, at four exposures to explore the
range of signal levels. The cameras were
set to +6dB gain. A software high-pass
spatial filter was used to reject low
frequencies and statistical analysis done to
calculate the video noise levels. Figures 7
and 8 show the results, plotted in noise
levels (dB vertically) against luma
amplitude
(percentage)
horizontally.
Values were obtained mathematically and
are un-weighted. The differences between
the cameras are trivial.
0
10
20
30
40
50
60
70
80
90
-39.5
-40.0
-40.5
-41.0
-41.5
Figure 7 Noise levels, Z5
Normally, the noise level would be
-42.0
expected to show a strong correlation with
the slope of the gamma curve since the
sensor and head-amplifier noise is amplified by the differential gain (slope) of the gamma-correction, and so
there should be about 10dB difference between the value around 10% and that around 90%. This cameras do
not show any such correlation.
One possible explanation for this could be
the use of analogue head amplifiers with
limited gain-bandwidth product (i.e.
cheaper). Thus, as the gain increases, the
bandwidth reduces and output noise level
goes down. Measurement of captured
resolution is too difficult to do at low video
levels (because of noise), so this
explanation
remains
unconfirmed.
However, the net result is that the noise
performance appears to be better than it
actually is, a desirable thing for such a
camera. Perhaps this was a design decision,
if so, it has worked.
0.0
10.0
20.0
30.0
40.0
50.0
60.0
70.0
80.0
90.0
-35.0
-37.0
-39.0
-41.0
-43.0
Alternatively, the sensor and head-amplifier
noise could be masked by digital noise. In
Figure 8 Noise levels, Z7
a professional camera, the adcs should be at
-45.0
least 10-bit (14-bit is common in high-end
cameras), and internal processing at least
18-bit, with 10-bit recording of HDSDI output. In lower-cost cameras, reduced bit-depths can be expected,
but these measurements were made at +6dB gain, so head noise should dominate, the matter is therefore
unresolved.
Even so, noise performance is not up to the standards of “proper” HDTV cameras (where 54dB is expected),
but the performance is not particularly bad. Compensating for the +6dB gain, the noise levels at 0dB gain
should be between -43.5 and 45.2dB, about 2.5dB better than the Z1. Such noise levels are not generally a
problem for simple programme production, but would be regarded as poor if the production involved any
significant amount of colour-keying or matteing, for whatever reason.
14