Re: [Gimp-print-devel] Stylus photo 870

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

   Date: Wed, 3 May 2000 23:21:37 +0000
   From: Jean-Marc Verbavatz <ver...@if...>

   > OK, probably the inks are different.  I should put an Epson cartridge
   > in my printer and see what happens.

   Aha, that would explain. I'm using Epson inks.

I should quit being such a cheapskate.  Then again, I sometimes burn
through a cartridge a week, doing all this testing.

   >    Primarily, red looks too yellow and blue too magenta to me. Gray is
   >    not quite gray either. This can probably be fixed by adjusting the
   >    respective densities but I haven't tried yet.
   >
   > I've noticed that too.  Does your gray turn out slightly greenish?

   Yup.

We have to solve this.  The repository is much better than even 3.1.3,
but it's still not perfect.

   Same thing with dither=4. The rendenring is slightly different but
   about as grainy.

   I just printed the exact same photograph (a rather difficult one
   from previous experience) with my version of gimp-print-3.1.2 (a),
   with stock 3.1.3 (b), yesterday's 3.1.3 with dither=2 (c) and with
   dither=4 (d).  b, compared to others is lousy. That's history.  c
   and d don't differ significantly. They're good. A bit grainy,
   slightly oversaturated, greenish grays (see above).  At this point
   (a) is best in my opinion. A little less saturated than c,d (but
   whether this is worse or better is rather subjective and can be
   adjusted).  Better grays. Less grainy.

Huh.  That's very interesting, and doesn't accord with my experience.
I would expect stock 3.1.3 to be the least grainy, followed by the
current repository and then 3.1.2.  3.1.2 might be denser, which would
reduce the perceived grain.

   The greenish thing should not be difficult to fix. That's color
   balance again and probably involves no specific flow in the
   algorithms.

I hope not, but it's been exceedingly difficult to get right thus far.

   One issue in the grain, I've seen it while working on 3.1.2, is
   what I would call "undersampling of smallest drops". Because I'm
   new here I don't know whether this has been discussed before,
   sorry. Anyway, I think that smallest drops in variable size (or
   smallest ink levels) are so tiny that they don't fill in. Take 50%
   gray. If printed with only smallest dot size black will shoot
   everytime at 720 dpi and yet will be grainy because the diameter of
   the drop is less than 35 microns. This particularly shows for dark
   inks (K,C,M). The fix is to mix in larger volumes of lighter inks
   (Y, LC, LM) instead and to use some CMY instead of K (larger
   volumes again) for midtone grays. The risk then is to damp the
   paper. So we need to keep track of the volume of ink laid down. I
   can tell for sure that this problem occurs and I can tell where
   exactly from the printout, particularly for grays, because it shows
   despite randomization.

Actually, gray IS printed as CMY (and LC and LM as appropriate) until
it gets quite dark.  And currently, there's some CMY mixed in all the
way to almost 100% black.

I think that the black threshold is a bit too low right now.  It's
also possible that the density value that's being used to compute the
size of the black dot is wrong.  Try changing this:

	  tk = print_color(d, &(d->k_dither), bk, bk, k, x, row, kptr,
			   NULL, bit, length, 0, 0, 0, 0);

to

	  tk = print_color(d, &(d->k_dither), bk, bk - d->k_lower, k, x, row, kptr,
			   NULL, bit, length, 0, 0, 0, 0);

What this will do is decide what kind of black dot to print based on
the difference between the black value and the lowest value at which
black is printed, rather than the absolute black value.  So that would
use smaller dots.  Let me know how it works.

   Let me know if I am too verbose. I realize that the above is rather
   descriptive and does not actually fix the problem. I'm still a bit
   overwhelmed by changes in dithering in 3.1.3. But I realize they're
   for the better; dithering plays a huge role in a good rendering.

Not at all.  You're coming up with entirely too many good ideas and
expressing them entirely too well for that :-)  Please, keep those
ideas and code flowing!

-- 
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