[Openicc] Drop size calibration

[Robert Krawitz]

   Date: Mon, 28 Jan 2008 23:24:24 +0100
   From: Gerhard Fuernkranz <nospam456 at gmx.de>

   Graeme Gill wrote:
   > Right, yes you need to be able to mix three drop sizes to avoid
   > banding issues as the screen gets close to 100% of one dot size
   > (which was what I meant by overlap of different dot sizes).

   But if we close down the first first drop size before reaching 100%
   area coverage, this implies that the next drop size should
   basically start dithering into the "holes" which have not yet been
   filled with the previous drop size, in order to obtain the least
   visible dot pattern.  Wouldn't this - in the optimal case - require
   a threshold array for the 2nd level which is particularly tailored
   to the ink limit of the first level? Or do you think, it does not
   matter?

It doesn't appear to be an issue in practice -- we use a single dither
array, and the result is smooth.  I can see the larger drops start to
fill in with a loupe, but the pattern of the larger drops is smooth.
The smaller drops are already at 50% coverage, so there aren't a lot
of holes to worry about.

Incidentally, I've tried constants other than 50%, and 50% seems to
yield the best results, at least with the cases I've tried.  60% or
higher yields little difference from the standard 2-level dither,
while anything much less than 50% yields too many large drops at too
low a density.

   > Sufficient overlap should smooth some of the worst of this out,
   > since it's the near 100% fill that has the highest dot gain,

   Isn't the "dot gain" (i.e. effective area coverage (computed from
   measured densities) minus prescribed area coverage (in %)) zero by
   definition, at 0% and 100% fill, and has its maximum rather
   somewhere in the middle?

It's not zero "by definition" at 100%; it's possible to use 100% of
the largest drops and not actually completely fill the page.  This
happens at low resolutions.  Depending upon the printer and paper
combination, it requires anywhere between 360 dpi and 1440x720 dpi to
really completely fill the page.

-- 
Robert Krawitz                                     <rlk at alum.mit.edu>

Tall Clubs International  --  http://www.tall.org/ or 1-888-IM-TALL-2
Member of the League for Programming Freedom -- mail lpf at uunet.uu.net
Project lead for Gutenprint   --    http://gimp-print.sourceforge.net

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