[argyllcms] Re: Bad gamut mapping from CMYK
- From: Graeme Gill <graeme@xxxxxxxxxxxxx>
- To: argyllcms@xxxxxxxxxxxxx
- Date: Tue, 25 Aug 2015 10:51:37 +1000
Nikolay Pokhilchenko wrote:
May be the problem arises only with higher resolution profile, the -qh
parameter?
Coated ceramic media measurements was attached to my previous message. The
profile of same printer with the same inks but textile media with less
contrast (another and much more less bumpy data set) gives no inversion even
with less smoothing factor.
I wonder why the inversion took a place at such high smoothing factor as -r3
I've used.
Sorry, I made a mess of transcribing the options (it was rather late):
colprof -v -qm -sISOcoated_v2_eci.icc RGB_Thermotransfer
+
cctiff -ip ISOcoated_v2_eci.icc -ip RGB_Thermotransfer.icm
RGB_Thermotransfer.icm
RGBsurfall_cmyk.tif test.tif
and
colprof -v -nP -sISOcoated_v2_eci.icc -qm RGB_Thermotransfer
+
cctiff -ir ISOcoated_v2_eci.icc -ip RGB_Thermotransfer.icm
RGB_Thermotransfer.icm
RGBsurfall_cmyk.tif test2.tif
My guess is that it is the clipping that is behind this behavior. If you are
feeding
an expanded perceptual gamut in, but the gamut mapping is expecting the more
compact
colorimetric source gamut, then there is a mis-match, and the clipping behavior
will be exposed. Something about that is mapping the black point to a red color.
I don't know why without a lot of detailed investigation, although it is
probably
related to how bumpy the device gamut is. If the input gamut is the one that the
gamut mapping is set for, then there doesn't seem to be a problem.
Graeme Gill.
Other related posts:
