I have some questions related to the use DCP profiles and white balance interaction in Photolab 7.
I have Spyder Checkr Photo and I know how to create a DCP profile from a RAW file in PL7.
Here is the experiment:
First step: I create 4 DCP profiles:
1/ WB set to “As shot” and “set white balance” on during DCP profile generation
2/ WB set to “As shot” and “set white balance” off during DCP profile generation
3/ WB set to “Daylight” and “set white balance” on during DCP profile generation
4/ WB set to “Daylight” and “set white balance” off during DCP profile generation
Second step: I apply this 4 profiles to another picture taken in the same lighting situation. The observations are:
Experiment 1: If I set the WB of the picture to “as shot”, all 4 DCP color profiles result in the same output colors, and the look is as per expectation
Experiment 2: If I set the WB of the picture to “Daylight”, all 4 DCP color profiles result in the same output colors, but the colors are clearly not as per expectation
My questions:
Q1: Why are the colors with DCP rendering different in experiment 1 and experiment 2, especially when the “set white balance” is “on” during DCP profile generation
Q2: How is the WB setting impacting the DCP profile rendering (especially in case the “set white balance” is “off” during DCP profile generation).
because you select different WBs ( and that overrides whatever WB slider values DxO selects for WB sliders when the “set white balance” is “on” )
read DNG standard - it is explained there… that is if DxO actually bothers to follow it
if you want WB matching target shot conditions do use WB generated off the shot target and not “as shot” or “daylight” … DxO generates a simple single illuminant matrix profile aimed for the same illumination as the shot of the target, while of course it can adapt to different WB you select it is not aimed for that purpose - so use an option to set WB when generating that DCP profile AND stick to those WB values
Don’t understand, why you want to set PL’s WB setting different
to how the picture was taken?
so what I tried
my pic (XRite color checker) was taken with camera set to As Shot
and PL7 to “2 - DxO Standard” preset (in PL6 and before = “1 - DxO Standard”)
.
then I set PL’s WB to
a) as shot
b) daylight
c) tungsten
Note that when the RAW White Balance tool is set “off” (regardless of any other WB tool setting) the result is the same as when the tool is “on” AND is set to “As Shot”
This is readily confirmed by setting the tool to “As Shot” with the tool “on” - and then setting the tool “off” … there will be no change in rendered WB.
I repeated your test but with a Spyder Checkr Photo
My pic was also taken “as shot”
I created DCP profiles for “as shot” and “daylight”, both with and without “set WB”
1/ When I set the pic ture WB to “as shot”, applying each of the 4 DCP profiles give me the same result
2/ When I set the pic ure WB to “daylight”, applying each of the 4 DCP profiles give me the same result (both other than the result in point 1/)
So either I misunderstood your test sequence, or we have different outcomes.
As long as test outcomes are different, its hard to discuss these results.
About Q1:
It’s not clear what the “set white balance” option is doing when a DCP profile is generated.
I asked the question before in the forum, but I didn’t got a satisfactory reply and that’s why I made the experiments.
The Spyder Checkr Photo has different “white like” color patches, so my understanding is that in the same light conditions the DCP profile intrinsically has the white balance inside.
If not, then I assume the “set white balance” option should do, although I have difficulty to understand the option and I neither see any difference in the profiles with that white balance enabled or not during DCP profile generation.
When I select DCP rendering, I would expect the WB setting should be void as in the DCP profile, which apparently is not the case.
It would be extremely useful if the DXO team explains how to use WB and DCP profiles!
My Q2 is about the WB setting impacts the DCP profile rendering. I do not grep the use of color profiles if a DCP profile is set with “set white balance” option enabled.
I’m heading for the way it is implemented in PL7, irrespective of the (interesting) more theoretical discussions
BTW, do you have a reference to the DCP standard?
Short version: They are independent of each other.
What does that mean? in relation to DCP profiles created with DPL7, it means that profiles don’t react on how the checkbox was set…and that we still need to apply WB for proper balance of images treated with a DCP profile.
Visual Verification: I created, from one image, profiles with WB OFF first and ON afterwards and applied the profiles to a second image and its virtual copy and got this:
If you want to repeat the test, be extra careful to not move any part of the frame that you set around the colour patches, or the profiles will be slightly different, even though the differences I got were minute (< 1/1000) when I first tested with drawing the frame for each profile separately.
In other words: The WB checkbox in the profile creation tool is for convenience only. It does not do anything to the profile.
The outcome of the DCP creation tool in PL7 is a DCP profile. If, as you stated, the WB checkbox does not do anything to the profile, then the outcome of the DCP profile creation is the same DCP profile with/without the WB checkbox set (that is btw also what I observe).
My question is: if the the WB checkbox does not changes the output of the created DCP profile, what is the purpose/convenience?
Correction (formatting error) :
You stated in your reply:
“In other words: The WB checkbox in the profile creation tool is for convenience only. It does not do anything to the profile.”
The outcome of the DCP creation tool in PL7 is a DCP profile. If, as you stated, the WB checkbox does not do anything to the profile, then the outcome of the DCP profile creation is the same DCP profile with/without the WB checkbox set (that is btw also what I observe).
My question is: if the the WB checkbox does not change the output of the created DCP profile, what is the purpose/convenience?
in a raw converter, in general, you need to use a camera profile and use some WB … DxO provide you a convenient way to set “auto” WB ( WB calculated based on specific illumination of your target when that raw file with the target was shot )
application of WB in raw converters ( how what you dial in UI sliders will be used by the code on the image data from your raw file ) that follow DNG standard ( in terms of applying DCP profiles ) is linked and dependent on the content of DCP profile ( https://helpx.adobe.com/content/dam/help/en/photoshop/pdf/DNG_Spec_1_7_0_0.pdf )
DxO PL is giving you essentially “auto” WB option ( values to use in sliders Temp/Tint in WB tool in UI that a calculated based on target and target illumination in your raw file ) - DxO PL tells you what you need to dial in WB sliders ( in UI ) for the rest of your shot with the same illumination when you are using that DCP camera profile …
look at values in Temp/Tint sliders - they are adjusted by DxO PL for you to use ( write them down, save as a partial preset, etc ) for the same illumination as how your target was illuminated
it is just a convenience for you, saving you an effort of trying to use “pick color” pipette ( PS: what a stupid thing to use word “color” for it btw )
DCP profile as you can learn from reading DNG specifications has certain tags / data that are used in calculations for application of WB by a raw converter and naturally difference in target illumination will results in different profiles … what you see when you are opening a raw file in a raw converter with DCP profile selected as values in WB sliders like K/temp & tint will be calculated based on the content ( certain tags/data ) of DCP profile ( if a raw converter follows DNG specification in that matter ) – in that sense some people might feel that “DCP profile intrinsically has the white balance inside” … I 'd rephrase ( simplification !!! ) as = DCP profile has data to use for color transform and data to use for WB application related calculation that was baked in ( by application that created this profile ) based on target illumination when the raw was shot ( if DCP profile was created that way - as they can be created for example just by having camera’s CFA spectral data w/o any shots at all , in this case that illumination - or a set for dual/triple illuminant profile - you simply use also as a pure spectral data of your choice )
Thanks a lot for the extra inputs! I think I understand better now how it works. Let me summarize:
1/ The DCP profile defines how the RAW color are rendered, but it does not define the WB (I need to further study why WB isn’t in the DCP standard)
2/ Its perfectly possible with the Spyder Checkr Photo chart to derive the WB in a specific lighting condition (using the “white” color targets) . This is done in PL7 by checking the “set white balance” option during DCP profile generation.
3/ If I want to RAW develop a picture taken in the same lighting conditions with as accurate color reproduction as the tools allow, the I need to do two things:
Do the color rendering with the generated DCP profile
Copy the WB of the Spyder Checkr Photo chart of which the white balance has been set during the DCP profile generation to the picture
The second step is a bit labor intensive, but I do not see another way at the time being
there is no “RAW color” - “color” will appear when you map some set of numbers ( after a number of operations with data recorded in raw file by camera’s firmware for sensel locations ) into coordinates in some proper color space ( I use “color” because even then some coordinates are not actually colors ) - that is usually called a “color transform” ( but in this case one side ( before “color transform” ) is not color or “color” - just some data serving as a proxy for the amount of illumination recorded by sensor during exposure
