[argyllcms] Re: Absolute colorimetric - dark saturated colours excessively light
- From: Graeme Gill <graeme@xxxxxxxxxxxxx>
- To: argyllcms@xxxxxxxxxxxxx
- Date: Sun, 05 Feb 2006 13:01:48 +1100
Gerhard Fuernkranz wrote:
Yes, I'm talking about Absolute Colorimetric. I was under the
And the remapping of out-of-gamut colors is rather an issue of art. As I
said, the ICC spec doesn't tell, _how_ exactly it must be done; it is
implementation dependent. Graeme has chosen to clip to the nearest color
in CIELAB or CIECAMxx space (-> lowest color difference). You cannot
say, it's right or wrong - but of course you may like it or not (as you
may like or dislike a painting or another work of art). But I do agree
with you, that I subjectively find the out-of-gamut clipping done by
ISOcoated.icc and ISOnewspaper.icc more pleasing. This applies
particularly to small printer gamuts with a high black point.
Some of the literature recommends doing gamut clipping in a manner
that weighs the luminance errors twice as highly as hue and saturation.
Unfortunately, I have not allowed for such a weighting factor in
the clipping code. It might be possible to add, but it would be
some little work, as it would essentially have to be done by adding
a value scaling factor to the rspl code (since the internals is
too tightly tied to the 3D/4D geometry.)
Graeme Gill.
Other related posts:
