Re: [Gimp-print-devel] 6-color generation

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

   Date: Mon, 31 Jan 2005 20:51:35 +1300
   From: Mark Tomlinson <mar...@xt...>

   On Sun, 30 Jan 2005 00:54:20 +0000
   "Alastair M. Robinson" <bla...@fa...> wrote:

   > Have you considered doing the separation in HSL space?
   > 
   > Perhaps you could use an HSLtoXXX function, with an array storing the 
   > hue angle of each colourant instead of the hard-coded hue angles in the 
   > existing HSLtoRGB function?
   > 
   > The luminance could probably be taken care of using the existing tuning 
   > methods.
   > 
   > One problem I can forsee with that approach, though, is that it will 
   > assume that all 360 degrees of fully-saturated colours can be 
   > represented by mixing two inks at a time.  I'm not 100% sure that would 
   > be true for CMYRB colour.

   I can't see this being easier than what is there currently. I do
   see a difficulty that it may be tricky to choose M+Y instead of R
   for light colors (same as GCR).

Well, let me ask: what's wrong with picking R for light colors?  The
luminosity of the red ink isn't that radically different from magenta;
only black is much darker than everything else.  At least if we use
the same mix for any given hue we'll avoid color shifts.

This idea should also minimize ink consumption, which will make a lot
of people happy.  We won't know its limitations until we try it.

   Even if Red isn't halfway between Magenta and Yellow, can't it
   simply be set to min(1.0*M, 0.7*Y) (or whatever other combination
   of M and Y give R)? If it's not perfect it could be fixed up by
   altering the hue-maps. Only problem here is needing different maps
   for different ink sets...

How do we express this kind of thing in a general way?  How do we
decide whether to use red or blue ink?

   On a side note, I did try representing the CMY inks as different
   hues and using a matrix to convert from "genuine" CMY to ink
   levels. I did get a transform which produced curves similar to the
   hue map that is there. I didn't really pursue it that hard - I felt
   it was inferior to the hsl maps.

This is an extension of the hsl maps that we already use, which is
appealing in that sense.

-- 
Robert Krawitz                                     <rl...@al...>

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 Gimp Print   --    http://gimp-print.sourceforge.net

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