I read this article “The Luminance of an sRGB Color”
However, there is something I don’t get.
XYZ is an absolute color space. Correct right?
sRGB is also an absolute color space, because it is specified according to specific physical hardware. Again correct, right?
An absolute color space does not depend on a whitepoint, correct?
So if sRGB is an absolute color space, and XYZ is an absolute color space, there should be exaclty one way to transform sRGB to XYZ. One particular 3x3 matrix.
However, ICC profiles operate in D50, and the red/green/blue chromaticities of sRGB in an ICC profile are given as different xyY coordinates than they are in other contexts!
What’s happening here? This is the part that I don’t get. XYZ (and xyY of course) are absolute spaces, and so is sRGB, so I would exact there to be one single well defined value for the red/green/blue chromaticities of sRGB.
How come ICC has different ones? Why does D50 matter?
For XYZ, D50 should not matter: it’s an absolute color space.
For sRGB, D50 does not matter, its whitepoint is D65.
So how can the fact that ICC uses D50 influence sRGB or its chromaticity coordinates at all?
What am I missing? All info appreciated!
My main issue, by the way, is that the RGB/XYZ Matrices from http://www.brucelindbloom.com/ are not working for me when computing from ICC.
ICC gives me a red, green, blue and white point. But plugging these in the brucelindbloom formula gives a different matrix, and the reason is exactly that those red/green/blue chromaticities seem to be given in a different way. What’s going on here? How can there be different coordinates for red/green/blue chromaticities in xyY if xyY is an absolute color space? How should I correctly interpret the red/green/blue chromaticity values in an ICC profile and get colors from it in an absolute color space (which by definition does not depend on a whitepoint)?
So the formula on brucelindbloom, as far as I know, tells how to create a matrix to convert unambiguously from the one absolute sRGB color space to the one absolute XYZ color space, with the chromaticities of sRGB space given unambiguously in the one xyY color space.
So is ICC using a different xyY color space with its D50 whitepoint, since its giving those absolute sRGB chromaticities with different numerical coordinates? How can that be possible, there is only one absolute xyY/XYZ color space.