[argyllcms] Re: Ideal camera target

  • From: Klaus Karcher <lists@xxxxxxxxxxxxxxxxx>
  • To: argyllcms@xxxxxxxxxxxxx
  • Date: Wed, 03 Feb 2010 09:12:17 +0100

edmund ronald wrote:
The SG may display a lot of colors, but rumor has that it is printed
with very few pigments.

... at least not less than the CC24 and much more than many other commercial targets. The SG is an inherent part of the test chart collection I use for large format scanner profiling and I'm really comfortable with it.

I guess the rumor arises from Principal Component Analysis results (Munsell Book, CC24, CCSG, CCDC, ...). It's generally accepted that few basis functions are sufficient to model even very comprehensive color collections like the Munsell Book with adequate accuracy ... however it draws an enormous distinction whether let's say 4 or 8 basis functions are required to get a comparable prediction accuracy. And there are huge differences between accuracy measures: even a tiny spectral RMSE can result in a high Delta E and vice versa.

Roy Berns's web page <http://art-si.org/> is a gold mine with respect to this range of topics.

Klaus



Other related posts: