Am 28.08.2011 13:04, schrieb Stephen T:
I'm still chipping away at the camera profiling challenge. The white point (wtpt tag) of a camera profile has important effects on the results. From Argyll docs: "the white point of the test chart, will be mapped to perfect white when used with any of the non-absolute colorimetric intents." Now the white point of my IT8.7/2 target (patch GS00) is slightly bluish: Lab = 92.04 -0.69 -2.10 (RGB (D50, gamma 2.2) = 230 232 235). Firstly, I have used the colprof parameter -U to scale the profile white point (wtpt) to L=100. This avoids exposure increasing when mapping the white point of the test chart to perfect white (non-absolute intent). But then there remains a slight reddish colour shift. This slight error can also be seen when processing the test chart - neutral patches are not perfectly neutral: reference data R080505.txt GS00 = 230 232 235 R080505.txt GS06 = 161 161 161 ufraw outputs (WideGamutRGB, custom WB on GS06): ufraw GS00 = 226 225 226 perceptual 224 225 231 absolute (should be bluish, error with perceptual) ufraw GS06 = 160 158 156 perceptual 159 158 160 absolute (should be neutral, error with perceptual) Now, can I force the white point to be perfectly neutral? Is their a utility to edit the wtpt tag in an ICC profile? Is this even a good idea?
AFAIK, the -u option is supposed to assume a perfect diffuse reflector as media white point (instead of assuming the brightest patch on the calibration chart as "white"). However, if you want to use the profile for "general purpose shooting" (and not only for a very particular studio lighting condition), then the shots need to be white-balanced anyway after applying the profile - and in this case the media white point in the profile does not really matter. Regards, Gerhard