Monito -- Alan wrote:

Is that sufficient, or is it not useful to profile each lens-body combination?  It is well known that lenses have color casts.  Is it feasible to have a profile of each lens and marry it to the profile of the body used for the photo to create an input profile for that combination?  I suppose it would be more accurate and possibly actually easier to measure a profile for each body-lens combination directly.

The reason I ask is because I'm seeing very noticeable differences between at least three combos:  1) 100-400 / 5D, 2) 50 / 20D, 3) 10-22 / 20D.  I'm sure I'd find differences between the other combos I could try.

Of course this could be a bit of a nightmare for a photographer with five bodies and ten lenses: 50 profiles.  And then one would have a workflow problem applying the profiles.

Ultimately I want to master and completely control color in my workflow, even if I have to take the long way around to learn it and even though I might not be able to afford all the equipment immediately.

I love threads like this one!

I "think" there is a better way to correct for differences be color on different cameras. Notice I did say "think" I have not tested or worked with this.

Adobe has a free editor for their Adobe Camera Raw (ACR) camera profiles. ACR's camera profiles are not color management profiles and I wish they had not called them profiles.

Using the ACR profile editor you can make a profile each camera. Then set a default for that camera using your custom profile. You can then set ACR to make the default specific to the camera's serial number.

So when you open an image in ACR it reads the camera's serial number from the EXIF data and automatically uses the correct color "profile". It would not automatically take into consideration the lens being used but would account for the differences in sensor.

If anyone has tested this work flow I would really like to read their comments on it!

photoboykane wrote:

Monito -- Alan wrote:
Is that sufficient, or is it not useful to profile each lens-body combination?  It is well known that lenses have color casts.  Is it feasible to have a profile of each lens and marry it to the profile of the body used for the photo to create an input profile for that combination?  I suppose it would be more accurate and possibly actually easier to measure a profile for each body-lens combination directly.

The reason I ask is because I'm seeing very noticeable differences between at least three combos:  1) 100-400 / 5D, 2) 50 / 20D, 3) 10-22 / 20D.  I'm sure I'd find differences between the other combos I could try.

Of course this could be a bit of a nightmare for a photographer with five bodies and ten lenses: 50 profiles.  And then one would have a workflow problem applying the profiles.

I use the single camera profile to correct the intrinsic problems with the way a camera's chip renders the colors and levels of light that strike it.  A lens, filter, light modifier, or light source will have a broad effect that is EASILY corrected with white balance in camera or in post.  For that I have one of those cheap collapsible "gray" targets that goes into a setup shot for a quick fix.  It's a known value and I've tested it to know what it should read in ACR to give me a baseline of neutral color response.


I have the DNG Editor and have worked some with it, but it's easier to get lost in it than it is to get the desired results out of it. 

I set up each camera with Chromaholics AcrCalibrator script. However, it is relying on a very small number of patches on a ColorChecker and is only able to do some very broad based adjustments.  In looking at a ColorChecker that has been "corrected" in ACR according to the Calibrator script, you often end up with artifacts in the blue and green patches, and overall density issues with magenta and cyan.  It also only partially corrects the differences in "exposure" and "contrast". 

It's useful to include a scaled back result of the AcrCalibrator as part of your Camera Raw Default, then finish the job with a well applied corrective profile.  If anyone is interested, I'll dig out and email an ICC "corrective" profile and ACR Default settings for the cameras I've had to tweak and still have on file.  PM me if you have one of the following and are interested in seeing if it has any value to you...
EOS 1DS Mk II
EOS 20D
Nikon D300
Fuji S2

It's a pretty broad mix of results from photographers I know who have gone this route.  Some (like me) find it to be crucial to "controlling" color, others find it worthless, and others still find that it totally screws up what they want to get from their images.  I'm not so naive to think that what works for me should be a "recommended" way to work for anyone else!

Thanks for the feed back. I will look in to the ACR profile editor when I have more time.

I have simply been using the ACR D2X Mode profile since they came out with the beta of it. I think it make a great improvement in the look of my portrait work! Before that profile was available I was checking out the profile editor. I simply did not like the look of the ACR profiles before then. After ACR 5.2 came out and with the new profiles the look was good enough for my tastes.

I normally shoot everything with one camera so I have not had the need to make my other cameras match. I do think it would be interesting to see if I could make a Canon and Nikon look more similar with the right profiles.

Studio 144 wrote:
I "think" there is a better way to correct for differences be color on different cameras. Notice I did say "think" I have not tested or worked with this.

Adobe has a free editor for their Adobe Camera Raw (ACR) camera profiles. ACR's camera profiles are not color management profiles and I wish they had not called them profiles.

Using the ACR profile editor you can make a profile each camera. Then set a default for that camera using your custom profile. You can then set ACR to make the default specific to the camera's serial number.

So when you open an image in ACR it reads the camera's serial number from the EXIF data and automatically uses the correct color "profile". It would not automatically take into consideration the lens being used but would account for the differences in sensor.

If anyone has tested this work flow I would really like to read their comments on it!

I use the profile editor in combination with a Gretag card to create custom profiles off of two exposure (1 tungsten, 1 daylight) and then let it autorun the profile.  I find it gets color far more neutral than any included script (including the camera-specific ones), no artifacting, and a very easy process once you get it down.  I have not, OTOH, profiled for each of my lenses, though it wouldn't take that long to do - if LR included support for applying presets on the basis of what lens is attached I would in a heartbeat.

Studio 144 wrote:
.......With a large enough profile the difference between the Delta E of an RGB image and one converted to CMYK is very low. Issues do develop where the RGB image is out of CMYK gamma.....

This is quite possibly a trivial question, but how do you measure Delta-E of an RGB image vs a cmyk image, and what software is used? Is this using Lab values of the RGB and the Lab values of separation?  I don't have any experience with measuring them digitally, but do check the Delta-E of our printed jobs for quality control, so i'm fairly familiar with it. I was just wondering out of curiosity's sake (I Love seeing how things tick

Studio 144 wrote:
.....I check to make sure when I am editing in RGB where it falls in the CMYK color space. I do think that the algorithms for moving the colors into gamma are getting better but not perfect.....

I'm very interested in how this is done. Right now I have been viewing by eye, comparing the two. This might get me close when using my Eizo Coloredge monitor, since it displays almost all of adobeRGB. It would be very cool if you could see exactly what is out of gamut and
exactly where at, so you can tell how much is being compressed.

**Edit I forgot about the Gamut warning in PS, but is there a way to see how far off you are?