Someone commented on a forum concerning RGB → CMYK transforms using the relative colorimetric rendering intent:
Is this accurate? I thought that relative colorimetric (and perceptual) rather subtracted the paper white from the tables within the profile. Sort of an attempt to have the output exist apart from the substrate it was printed on.
Hi Rich,
I just recently discovered the nature of the whitepoint adaptation in ICC profiles. It is not so much an assumption of L100 a0 b0, but an adaptation or scaling. I believe all three intents in both directions have the same whitepoint. Absolute and relative share the Colorimetric tables. When the profile is built, the whitepoint from your measurement data is recorded as a tag, and adapted in some manor to L100 a0 b0, and so is the rest of the data (Scaled). When the absolute rendering is used, all the data is then mapped back to the media whitepoint tag.