[Gimp-print-devel] Mike Porter's image

[Robert L K. <rl...@al...>]

I've analyzed it a bit more.  If anyone wants to see what's going on,
either download his image or
http://www.tiac.net/users/rlk/colors3.tif.  Try printing it at various
saturation levels.  It's very obvious what's happening.

I think that what's going wrong is that we're treating each channel
as independently scaling between 0 and 65535.  This means that any
input with zero red prints the same cyan level, whether its
RGB=0,255,255 (pure cyan) or RGB=0,0,0 (pure black).  This results in
too much cyan being used in pure cyan, so it winds up being very dark,
and then it can't get any darker as we move toward black.  This simply
doesn't work with subtractive CMYK inks (it works with additive RGB
light sources).

I'm not entirely certain what to do here.  There needs to be some kind
of correction.  I'm going to try a few ad-hoc ideas, and if some of
the folks with a better grounding in the theory want to jump in, that
would be great.  While this doesn't affect every image, I do think of
it as a showstopper for 3.2, because for certain reasonable images the
effects are very severe.

-- 
Robert Krawitz <rl...@al...>      http://www.tiac.net/users/rlk/

Tall Clubs International  --  http://www.tall.org/ or 1-888-IM-TALL-2
Member of the League for Programming Freedom -- mail lp...@uu...
Project lead for The Gimp Print --  http://gimp-print.sourceforge.net

"Linux doesn't dictate how I work, I dictate how Linux works."
--Eric Crampton