Re:Kodak CMM in Photoshop
Re:Kodak CMM in Photoshop
- Subject: Re:Kodak CMM in Photoshop
- From: Chris Cox <email@hidden>
- Date: Mon, 23 Apr 2001 20:53:02 -0700
At 10:05 PM -0700 4/21/01, email@hidden wrote:
It's a combination of problems in ColorSync and the Kodak CMM, and
>could also be a bad profile (we're finding lots of profiles that
>don't seem to grasp the ICC definition of white in 16bit LAB).
I don't think it's that simple because I've had the problem with all of
the CMM's, but less so with the Apple CMM, and almost none with the Adobe
CMM.
And you really should be using the Adobe CMM (ACE) for best results.
I've had a ProfileMaker 3.1.2 profile generate non-zero results with the
Adobe CMM; and then the exact same measurement data building a profile in
3.1.4 fixed the problemw with the Adobe CMM, but it remained with other
CMMs.
Now maybe one can effectively "blame" this on improperly built ICC
profiles, but I blame the CMM's (all of them) more than the profile. The
CMM should be smarter than the profile and not allow the profile (which
is just a text file afterall) to mandate a non-zero result with
perceptual, relative colorimetric, and saturation rendering.
ColorSync has one problem scaling 16 bit/channel LAB data -- so most
third party CMMs will have problems when called from Photoshop.
Apple is working to fix this.
I won't even START on the bugs in ICM2 and third party CMMs.
ProfileMaker has also has problems with some of it's own data
scaling. They were supposedly working on a fix (but I don't have
3.1.4 to test).
There is no good way to test the indended result of the profiles to
within a few percent accuracy.
And with some profiles, you CANNOT get a "correct" result due to the
way the tables are built (ie: if white has to be interpolated, it's
almost certain to be wrong....).
Adobe is working with everyone (ICC, OS makers, and vendors) to at
least identify the problems, but it's a slow process -- and some
vendors seem to have trouble with the basic math.......
Chris