[argyllcms] Re: Problem with 1.1.0 RC2 ?

  • From: Graeme Gill <graeme@xxxxxxxxxxxxx>
  • To: argyllcms@xxxxxxxxxxxxx
  • Date: Tue, 15 Dec 2009 10:00:25 +1100

andrzej duda wrote:

On 14 Dec 2009, at 07:55, Nikolay Pokhilchenko wrote:

You should use Ekta Space as Your source gamut while profiling display.

I did it. It's actually BestRGB (Ekta Space was grayed in dispcalGUI).

(c) 2009 adddr. Created with dispcalGUI and Argyll CMS:
dispcal -v -d1 -c1 -P0.672,0.22,2.75 -H -qh -t6500 -b100 -g2.2 -f1 -k0 -A20
colprof -v -qh -al -s "/Library/ColorSync/Profiles/BetaRGB.icc" -cmt -dmt

And nothing changed.

I think you will find that there is no practical significance to
what you observe, since the colors involved are outside the
spectrum locus, and therefore can never occur in real world

They are a result of the CIECAM02 transformation, which
is known to have problems with non-real world colors.

Doing a gamut mapping from a super-big source space
(such as ProPhoto) is not useful though, since this will
over compress anything in that space. For large source
encoding colorspaces you need to use a different
workflow in which you first do any desired manual
image adjustment, and then gamut map from the
specific image gamut to the destination print space.
(ie. this is an "input referred" type of workflow
 rather than re-rendering from one output space
 to another.)

I will (once again) take a look at the CIECAM02 conversion
to see if there is anything that can be done about it,
but this is really a research problem (ie. there is
a whole CIE technical committee TC8-11 composed of some rather
eminent color scientists currently looking for a solution
to this particular issue.)

Graeme Gill.

Other related posts: