My EX3 Picture Profiles first round (applicable to EX1?) at DVinfo.net

Les Nagy · October 26th, 2008, 02:27 AM

I have done extensive "lab" testing and adjusting using a DSC Labs chart and different light sources ranging from 2700K incandescent to 6400K compact FL. There are many settings left untouched as far as those available in the hidden maintenance menus and others that are of use for specific requirements such as skin detail.

My tests so far have been many hours of playing and readjusting of matrix settings while watching the vectorscope in On Location CS4. I am using a large Camette chart from DSC and have it lit with 2700K incandescent and then with 6400K CFL lamps. There is good agreement between the two lighting setups.

First of all, many have already noticed and mentioned the usefulness of the standard picture profile and even none at all at giving a useful image. I would agree that the image is not bad and actually is a good place to start. On my particular camera I have found that using a standard profile and switching matrix to FL (fluorescent light) mode gives a very good colour balance on a vectorscope. So if you want to try this one, just reset your profile and switch to FL. On my camera, this balances out the scope display but squashes the green/magenta axis a bit. With it set to standard I get the green channel phase shifted too much toward red. Anyway, give this a try and see if your EX3 looks better or not with this standard setup.

In my testing, I found a couple of the gamma settings to be a bit too aggressive at compressing the highlights, namely Cinema2 and STD2. I have not tested with these yet. All my tests involved Cinema 1,3,4 and STD 1 and 3. My goal was to maximize colour accuracy, gamut, dynamic range and then double check visually with my calibrated LCD display monitoring the firewire feed from my EX3 while in SQ mode.

My base setting start with Cinema 1 gamma. The gammas Cinema 3 and 4 need some compensation from the Cinema 1 settings to prevent certain colour channels going out of balance near maximum. The settings assume all others not mentioned are at defaults. As a default I always set my white point at 5600K in the profile. The settings are:

Name.......LAMMA1C1.......LAMMA1C3...LAMMA1C4...LAMMA2S1...LAMMA2S3
Matrix......ON
Setting.....HiSat
Level.......+12
Phase......+8
R-G.........+99 (yes +99)
R-B.........-18
G-R.........-28..................+2..............-3..............-33.............-33
G-B.........±0
B-R.........-18..................+7..............+7
B-G.........-8

Gamma
Level.......±0
Select......Cinema1...........Cinema3.......Cinema4......Standard1....Standard3

Black.......-8.......................................................-4...............-4

Additional settings for the Standard gamma curves: Knee.....ON
.......................................................................Auto.....ON
.......................................................................Satlevel 50

The problem I see with my matrix settings is that one of them is pegged at 99. This usually indicates that something is not going to be repeatable between units as it would be unusual that every camera would have to be run at a maximum deviation from a stock setting to produce acceptable results. I think I might have an anomalous camera.

With these settings, all profiles have the same colour appearance within reason. I have also noticed that one needs to pay very careful attention to skin tones blowing out with some of the cinema gammas, especially Cinema 4 if you cannot do a "proper" test like I have. If you are going to use the gammas that compress the highlights quite a bit, then keep the highlights below 95% and it will help. Its should also be noted that the Knee function tends to "break" when exposures push past a certain level and the control gets lost. For all exposures one MUST ensure that 100% is not exceeded and perhaps even keeping things below 97% would be advised. Auto-exposure always gets things wrong but can be used as a starting point.

LAMMA2S3 seems have about 1-1/2 stop advantage of low light capability over LAMMA1C1. The cine gammas use a lot of light up in their compression of highlights. Everything is a trade off. So far it looks like the settings I am going to prefer are LAMMA1C3,4 and LAMMA2S3. Because the cine gammas use up light, they are also useful in helping to keep you aperture wider for less depth of field if needed.

For unknown situations where I don't have time to think, the afore mentioned standard profile with the FL setting will also be useful.

This first post on this just covers the basics, believe it or not. There are other settings that can be used to tweak things further should the need arise.

All the gammas have the "Level" setting that moves the break points in the curves around and can be used to achieve better coverage of a certain area of exposure in a specifically lit shot.

There is also the black gamma setting which adjust the steepness of the curve at the bottom end right near black, This can be adjusted to help crush shadows and darker areas in a scene, or to help bring up that last bit of dark detail if needed.

Low Key Sat can be used to enhance or kill colours in shadows. It should be used very carefully if you don't want some really bizarre effects unless that's what you want!

More to come as I get some comparison footage with the different settings.

Les Nagy · October 28th, 2008, 08:23 AM

Revision time. I hope everyone here don't mind watching my brain working from the inside.

First of all, scrap the FL setting idea for now. It wasn't as good idea as it appeared. It could still be useful for creative effect of course.

I have simplified the matrix settings by not trying to chase things as much and by balancing any errors from ideal a bit better. The result looks mostly the same but I was noticing colours going out of whack at extremes with the old settings. The new settings are renamed to LAMMAA,B

Name.......LAMMAAC1.......LAMMAAC3...LAMMAAC4...LAMMABS1...LAMMABS3
Matrix......ON
Setting.....HiSat
Level.......+12.................+8..............+10.............+10.............+10
Phase......+3 (was+8)
R-G.........+85 (was +99)
R-B.........±0 (was -18)
G-R.........-28.................-28 (was +2).-28 (was -3).-28 (was -33).-28 (was -33)
G-B.........±0
B-R.........-34 (was-18)....-34 (was +7).-34 (was +7) -34 (was ±0)..-34 (was ±0)
B-G.........±0 (was-8)

Gamma
Level.......±0
Select......Cinema1...........Cinema3.......Cinema4......Standard1....Standard3

Black.......-8.......................................................-4...............-4
Blk G.......-4

Additional settings for the Standard gamma curves: Knee.....ON
.......................................................................Auto.....ON
.......................................................................Satlevel 50

If I were to play with settings further, it would be to change between Hi Sat and Standard Sat, and the level for different lighting situations to get the saturation I would want for that situation. The default Picture Profile is always a safe bet to start.

Samples sometime this week.

Barry J. Anwender · October 28th, 2008, 09:58 AM

Quote:

Originally Posted by Les Nagy

My tests so far have been many hours of playing and readjusting of matrix settings while watching the vectorscope in On Location CS4.

Les, I have also played with On Location CS4. Unless I've missed something, it only works when connected to a video camera via Firewire (iLink). In the case of the EX1/3 Firewire/iLink delivers an HDV video stream (1440x1080) and that is with the EX1/3 video format set to SP (25Mbs).

I've also tried to go via the Import Clip route but On Location has very limited QuickTime .mov import capabilities.

Now I'm on Mac with CS4 Master Suite, so your mileage may vary on a PC, yet adobe's manual seems to indicate the same thing for the Window's environment.

Les Nagy · October 28th, 2008, 12:26 PM

You are correct Barry. I am monitoring the waveforms in HDV mode, or SP mode as the EX camera states it.

Jon Sands · October 29th, 2008, 04:55 AM

Sounds like some great work! Sadly I don't think it's applicable to the EX1, I loaded it up on mine and had some odd color shifts, mainly light blues turned greeny.

Les Nagy · October 29th, 2008, 05:54 AM

Quote:

Originally Posted by Jon Sands

Sounds like some great work! Sadly I don't think it's applicable to the EX1, I loaded it up on mine and had some odd color shifts, mainly light blues turned greeny.

I am beginning to believe more and more that Picture Profile colour matrices do not translate from one camera to the next. At the very least it appears that they don't carry from EX1 to EX3 or vice versa. If the former is true it means that we are all on are own for setting colours on our cameras.

This does not mean that the way to set these things up and the other settings in PPs are not "shareable". To this end I will be going on to work with the different settings in this thread to help illustrate some of the things missing from the manual about PP settings.

What do people think about doing an experiment? I send out a standard colour chart printed out all at once by me on one printer to anyone who requests it. The camera chart would be imaged using whatever EX camera the participant has and they send me a short clip via email or ftp or file sharing so that I can test the footage and see how cameras differ. Perhaps then we could figure out how best to share settings and how to set a camera up to a standard. I am willing to start this out of my pocket for the first three EX3 and the first EX1 participants.

Piotr Wozniacki · October 29th, 2008, 05:57 AM

Quote:

Originally Posted by Les Nagy

What do people think about doing an experiment? I send out a standard colour chart printed out all at once by me on one printer to anyone who requests it. The camera chart would be imaged using whatever EX camera the participant has and they send me a short clip via email or ftp or file sharing so that I can test the footage and see how cameras differ. Perhaps then we could figure out how best to share settings and how to set a camera up to a standard. I am willing to start this out of my pocket for the first three EX3 and the first EX1 participants.

Great idea, Les; I'm in !!!

Trying and abandoning Bill's PP's, I have come to the same conclusion - each of the EX cameras is different, and certainly not calibrated perfectly for default matrices values (otherwise, why would Sony allow as much room for adjustment, as -99 through +99)?

Jay Gladwell · October 29th, 2008, 06:02 AM

Quote:

Originally Posted by Piotr Wozniacki

... each of the EX cameras is different, and certainly not calibrated perfectly for default matrices values (otherwise, why would Sony allow as much room for adjustment, as -99 through +99)?

Hasn't this always been the case? Isn't this why the video engineers in the truck or in the studio on multi-camera shoots have to tweak the cameras so they all "match"?

Les Nagy · October 29th, 2008, 06:38 AM

Quote:

Originally Posted by Jay Gladwell

Hasn't this always been the case? Isn't this why the video engineers in the truck or in the studio on multi-camera shoots have to tweak the cameras so they all "match"?

I agree Jay. It's just that as you know with you and I both owning XL2s that the presets shared on those resulted in people getting relatively close to each other compared to what we are seeing with the EX series. I don't expect things to be perfect but I expected there to be a bit closer agreement that would allow people to just tweak their own camera to the picture profile and get a close result just like we were able to do with presets on the XL2s.

So with that in mind I am taking up your earlier challenge in another thread to help figure out some of the finer points of picture profiles and share the information. :)

Bill Ravens · October 29th, 2008, 06:39 AM

Les...

There's a reason DSC color charts are so expensive. The reason is that color printers are not that repeatable that you can make chromatically correct prints twice in a row from the same negative. This is especially true for the CMYK printing process. By doing what you're suggesting, you will add more confusion to the mix unless you can verify each and every color print you make with a spectrophotometer. If you proceed with this idea, you may want everyone who gets one of your prints to sign a waiver that they understand the print they are getting does not comply with SMPTE color standards. Someone will try to set up their camera with one of these prints.

Furthermore, spectral highlights must be carefully suppressed during camera calibration. Normal resin coated print paper is way too reflective for this use. Expecting everyone who reports back to you to carefully manage their lighting is...well, overestimating the crowd and their skills.

Les Nagy · October 29th, 2008, 07:00 AM

I know Bill. It is not my intent to make a standard chart that is usable for anything anywhere. The chart I will print out will be usable only under one lighting condition of my choosing and will be checked out on my camera to ensure they all read the same with mine. They will also probably only be valid for a month or two as they change in characteristic. The idea here is just to get an idea of the baseline differences in cameras to work out what is feasible to try and share as far as picture profiles go.

The chart will probably be of similar colours as the 6 main colours on a DSC Labs chart. Lighting will be from two 60W plain frosted incandescent bulbs placed on each side. White balance would be on the chart itself. The camera will need to be on for at least 15 minutes and black balanced first. The test footage would be in 1080i SP mode shot with the same picture profile settings, probably off. I think its worth a try.

Piotr Wozniacki · October 29th, 2008, 07:14 AM

Quote:

Originally Posted by Jay Gladwell

Hasn't this always been the case? Isn't this why the video engineers in the truck or in the studio on multi-camera shoots have to tweak the cameras so they all "match"?

Absolutely Jay. I shot a 3-EX1 live music performance project lately; at rehearsals I spent couple of hours trying and tweaking PP's on my own EX1. And when I thought I found the best settings combination, just before the first concert began, the 2 other operators arrived; I made them dial-in my settings - only to see we were getting very different pictures (thanks to God, I had my Vaio laptop with me, whose 17" screen was enough to spot it).

We ended up shooting with PP off, and I'm now almost finished cutting the 3 cameras in Vegas...

Jay Gladwell · October 29th, 2008, 08:04 AM

Piotr, exactly my point.

Isn't that what the RM-B750 (Sony Product Detail Page - RMB750) is for, matching
cameras?

Ted OMalley · October 29th, 2008, 08:22 AM

Quote:

Originally Posted by Les Nagy

What do people think about doing an experiment? I send out a standard colour chart printed out all at once by me on one printer to anyone who requests it. The camera chart would be imaged using whatever EX camera the participant has and they send me a short clip via email or ftp or file sharing so that I can test the footage and see how cameras differ. Perhaps then we could figure out how best to share settings and how to set a camera up to a standard. I am willing to start this out of my pocket for the first three EX3 and the first EX1 participants.

Count me in too, Les, I'm extremely curious about this and would be happy to help.

Jon Braeley · October 30th, 2008, 06:54 AM

My own take on this is may be different.

Looking around the forum's much of the chat is about what PP's everyone is using. It's just not feasible for users to think that someone else's PP's are suitable for your situation. I think many early novice users, see what PP's the pro's are using, and emulate these - and this for them is the holy grail ... now all is well. Switch on, take great images. Earn money.
Yes, I've dialed in PP's myself - but I find myself stripping them back out fairly soon after. I havent found one that was useable for me - I think once on an extremely sunny day, I left a PP to deal with it - I think Mr Bloom's. I think all my PP's are now very close to the defaults and I adjust in camera during the shoot mostly. I just look after white balance carefully, adjust the temp after, check histogram always while shooting, set for shadows - highlights, then check overall exposure and done! When I shoot on location I will save to a PP for the duration of the days I am there - as a template.
I have a PP for indoors (STD), one for outside (sunny and overcast-CINE) - but no holy grails. I may be wrong, but I feel many users rely on PP's from wherever they can get them, without really understanding why. This just did not work for me like they seem to work for other users.

October 26th, 2008, 02:27 AM	#1
Les Nagy Major Player Join Date: May 2006 Location: A Canadian in Canada, and sometimes Chile Posts: 265	My EX3 Picture Profiles first round (applicable to EX1?) I have done extensive "lab" testing and adjusting using a DSC Labs chart and different light sources ranging from 2700K incandescent to 6400K compact FL. There are many settings left untouched as far as those available in the hidden maintenance menus and others that are of use for specific requirements such as skin detail. My tests so far have been many hours of playing and readjusting of matrix settings while watching the vectorscope in On Location CS4. I am using a large Camette chart from DSC and have it lit with 2700K incandescent and then with 6400K CFL lamps. There is good agreement between the two lighting setups. First of all, many have already noticed and mentioned the usefulness of the standard picture profile and even none at all at giving a useful image. I would agree that the image is not bad and actually is a good place to start. On my particular camera I have found that using a standard profile and switching matrix to FL (fluorescent light) mode gives a very good colour balance on a vectorscope. So if you want to try this one, just reset your profile and switch to FL. On my camera, this balances out the scope display but squashes the green/magenta axis a bit. With it set to standard I get the green channel phase shifted too much toward red. Anyway, give this a try and see if your EX3 looks better or not with this standard setup. In my testing, I found a couple of the gamma settings to be a bit too aggressive at compressing the highlights, namely Cinema2 and STD2. I have not tested with these yet. All my tests involved Cinema 1,3,4 and STD 1 and 3. My goal was to maximize colour accuracy, gamut, dynamic range and then double check visually with my calibrated LCD display monitoring the firewire feed from my EX3 while in SQ mode. My base setting start with Cinema 1 gamma. The gammas Cinema 3 and 4 need some compensation from the Cinema 1 settings to prevent certain colour channels going out of balance near maximum. The settings assume all others not mentioned are at defaults. As a default I always set my white point at 5600K in the profile. The settings are: Name.......LAMMA1C1.......LAMMA1C3...LAMMA1C4...LAMMA2S1...LAMMA2S3 Matrix......ON Setting.....HiSat Level.......+12 Phase......+8 R-G.........+99 (yes +99) R-B.........-18 G-R.........-28..................+2..............-3..............-33.............-33 G-B.........±0 B-R.........-18..................+7..............+7 B-G.........-8 Gamma Level.......±0 Select......Cinema1...........Cinema3.......Cinema4......Standard1....Standard3 Black.......-8.......................................................-4...............-4 Additional settings for the Standard gamma curves: Knee.....ON .......................................................................Auto.....ON .......................................................................Satlevel 50 The problem I see with my matrix settings is that one of them is pegged at 99. This usually indicates that something is not going to be repeatable between units as it would be unusual that every camera would have to be run at a maximum deviation from a stock setting to produce acceptable results. I think I might have an anomalous camera. With these settings, all profiles have the same colour appearance within reason. I have also noticed that one needs to pay very careful attention to skin tones blowing out with some of the cinema gammas, especially Cinema 4 if you cannot do a "proper" test like I have. If you are going to use the gammas that compress the highlights quite a bit, then keep the highlights below 95% and it will help. Its should also be noted that the Knee function tends to "break" when exposures push past a certain level and the control gets lost. For all exposures one MUST ensure that 100% is not exceeded and perhaps even keeping things below 97% would be advised. Auto-exposure always gets things wrong but can be used as a starting point. LAMMA2S3 seems have about 1-1/2 stop advantage of low light capability over LAMMA1C1. The cine gammas use a lot of light up in their compression of highlights. Everything is a trade off. So far it looks like the settings I am going to prefer are LAMMA1C3,4 and LAMMA2S3. Because the cine gammas use up light, they are also useful in helping to keep you aperture wider for less depth of field if needed. For unknown situations where I don't have time to think, the afore mentioned standard profile with the FL setting will also be useful. This first post on this just covers the basics, believe it or not. There are other settings that can be used to tweak things further should the need arise. All the gammas have the "Level" setting that moves the break points in the curves around and can be used to achieve better coverage of a certain area of exposure in a specifically lit shot. There is also the black gamma setting which adjust the steepness of the curve at the bottom end right near black, This can be adjusted to help crush shadows and darker areas in a scene, or to help bring up that last bit of dark detail if needed. Low Key Sat can be used to enhance or kill colours in shadows. It should be used very carefully if you don't want some really bizarre effects unless that's what you want! More to come as I get some comparison footage with the different settings. Last edited by Les Nagy; October 26th, 2008 at 05:47 AM. Reason: spelling corrections

October 28th, 2008, 08:23 AM	#2
Les Nagy Major Player Join Date: May 2006 Location: A Canadian in Canada, and sometimes Chile Posts: 265	Revision time. I hope everyone here don't mind watching my brain working from the inside. First of all, scrap the FL setting idea for now. It wasn't as good idea as it appeared. It could still be useful for creative effect of course. I have simplified the matrix settings by not trying to chase things as much and by balancing any errors from ideal a bit better. The result looks mostly the same but I was noticing colours going out of whack at extremes with the old settings. The new settings are renamed to LAMMAA,B Name.......LAMMAAC1.......LAMMAAC3...LAMMAAC4...LAMMABS1...LAMMABS3 Matrix......ON Setting.....HiSat Level.......+12.................+8..............+10.............+10.............+10 Phase......+3 (was+8) R-G.........+85 (was +99) R-B.........±0 (was -18) G-R.........-28.................-28 (was +2).-28 (was -3).-28 (was -33).-28 (was -33) G-B.........±0 B-R.........-34 (was-18)....-34 (was +7).-34 (was +7) -34 (was ±0)..-34 (was ±0) B-G.........±0 (was-8) Gamma Level.......±0 Select......Cinema1...........Cinema3.......Cinema4......Standard1....Standard3 Black.......-8.......................................................-4...............-4 Blk G.......-4 Additional settings for the Standard gamma curves: Knee.....ON .......................................................................Auto.....ON .......................................................................Satlevel 50 If I were to play with settings further, it would be to change between Hi Sat and Standard Sat, and the level for different lighting situations to get the saturation I would want for that situation. The default Picture Profile is always a safe bet to start. Samples sometime this week.

October 28th, 2008, 12:26 PM	#4
Les Nagy Major Player Join Date: May 2006 Location: A Canadian in Canada, and sometimes Chile Posts: 265	You are correct Barry. I am monitoring the waveforms in HDV mode, or SP mode as the EX camera states it.

October 29th, 2008, 04:55 AM	#5
Jon Sands Regular Crew Join Date: Sep 2008 Location: Indianapolis, Indiana Posts: 107	Sounds like some great work! Sadly I don't think it's applicable to the EX1, I loaded it up on mine and had some odd color shifts, mainly light blues turned greeny.

October 29th, 2008, 06:39 AM	#10
Bill Ravens Join Date: Nov 2001 Location: santa fe, nm Posts: 3,264 Images: 10	Les... There's a reason DSC color charts are so expensive. The reason is that color printers are not that repeatable that you can make chromatically correct prints twice in a row from the same negative. This is especially true for the CMYK printing process. By doing what you're suggesting, you will add more confusion to the mix unless you can verify each and every color print you make with a spectrophotometer. If you proceed with this idea, you may want everyone who gets one of your prints to sign a waiver that they understand the print they are getting does not comply with SMPTE color standards. Someone will try to set up their camera with one of these prints. Furthermore, spectral highlights must be carefully suppressed during camera calibration. Normal resin coated print paper is way too reflective for this use. Expecting everyone who reports back to you to carefully manage their lighting is...well, overestimating the crowd and their skills.

October 29th, 2008, 07:00 AM	#11
Les Nagy Major Player Join Date: May 2006 Location: A Canadian in Canada, and sometimes Chile Posts: 265	I know Bill. It is not my intent to make a standard chart that is usable for anything anywhere. The chart I will print out will be usable only under one lighting condition of my choosing and will be checked out on my camera to ensure they all read the same with mine. They will also probably only be valid for a month or two as they change in characteristic. The idea here is just to get an idea of the baseline differences in cameras to work out what is feasible to try and share as far as picture profiles go. The chart will probably be of similar colours as the 6 main colours on a DSC Labs chart. Lighting will be from two 60W plain frosted incandescent bulbs placed on each side. White balance would be on the chart itself. The camera will need to be on for at least 15 minutes and black balanced first. The test footage would be in 1080i SP mode shot with the same picture profile settings, probably off. I think its worth a try.

October 29th, 2008, 08:04 AM	#13
Jay Gladwell Join Date: Jul 2002 Posts: 1,244	Piotr, exactly my point. Isn't that what the RM-B750 (Sony Product Detail Page - RMB750) is for, matching cameras?

October 30th, 2008, 06:54 AM	#15
Jon Braeley Major Player Join Date: Oct 2007 Location: Miami, FL Posts: 528	My own take on this is may be different. Looking around the forum's much of the chat is about what PP's everyone is using. It's just not feasible for users to think that someone else's PP's are suitable for your situation. I think many early novice users, see what PP's the pro's are using, and emulate these - and this for them is the holy grail ... now all is well. Switch on, take great images. Earn money. Yes, I've dialed in PP's myself - but I find myself stripping them back out fairly soon after. I havent found one that was useable for me - I think once on an extremely sunny day, I left a PP to deal with it - I think Mr Bloom's. I think all my PP's are now very close to the defaults and I adjust in camera during the shoot mostly. I just look after white balance carefully, adjust the temp after, check histogram always while shooting, set for shadows - highlights, then check overall exposure and done! When I shoot on location I will save to a PP for the duration of the days I am there - as a template. I have a PP for indoors (STD), one for outside (sunny and overcast-CINE) - but no holy grails. I may be wrong, but I feel many users rely on PP's from wherever they can get them, without really understanding why. This just did not work for me like they seem to work for other users.