[argyllcms] Re: targen -c terminally slow

  • From: Graeme Gill <graeme@xxxxxxxxxxxxx>
  • To: argyllcms@xxxxxxxxxxxxx
  • Date: Fri, 18 Dec 2009 01:49:09 +1100

Ben Goren wrote:
targen -v -d 3 -e 8 -s 128 -g 256 -f 2142 -c ../../Red\ River\ 65lb.\ Premium\ Matte\ 
C2S\ 2.0/2009-11-18/Canon\ iPF8100\ -\ Red\ River\ 65lb.\ Premium\ Matte\ C2S\ 2.0\ -\ 
2009-11-18.icc  "Canon iPF8100 - Red River 60lb. Premium Matte 2 sided - 
Still running....

Hmm. I think I can fix that so it will run a whole lot faster. The trigger is
using so many regularly spaced fixed test points (-s 128 -g 256). This stresses
the voronoi computation. The fix will be to not add all the fixed point before
the movable points. This may still not be very fast for the pathological
case of adding very many fixed points and then only one movable point though...

By the way, since this is a printer, you should (in theory) be using -d2,
although this will change the -s 128 to be CMY wedges from white
rather than RGB wedges from black. CMY is likely to be closer to
the actual pure colorant for a printer.

Graeme Gill.

Other related posts: