[argyllcms] using ArgylLcms to work around colorSync issues

  • From: michael cinquin <safarlebu@xxxxxxxx>
  • To: argyllcms@xxxxxxxxxxxxx
  • Date: Thu, 17 Jun 2010 20:32:16 +0000 (GMT)

Hi all I'm trying to work around ColorSync oddities on mac os x, with the help 
of ArgylLcms.
I use Apple Color, which has a strange behavior depending on the Icc Display 
Profile that is chosen in the System Pref -> Display pane.
On Snow Leopard, I have noticed using a scope that when using "Generic RGB 
profile", Color displays accurately the primaries (Red, Green, Blue) and 
secondaries (Cyan, Magenta, Green). But the gamma is wrong : the image is too 
dark.If I use "Adobe RGB" as a display profile, Color displays accurately the 
gamma, but not the colors (has been checked with a scope).
I have found a temporary solution : using the display profile "Generic RGB", 
and using dispwin to load a 2.2/1.8 gamma .cal file : this way the gamma and 
the colors are correct.
I would like to make this temporary solution more elegant : either by (A) being 
able to modify the gamma of the "Generic RGB Profile", or (B) by permanently 
adding a vcgt to the icc profile.
Solution A : modifying the gamma of the "Generic RGB" Profile or building a 
generic one from scratch- I saw in ColorSync Utility that the Generic RGB 
profile has response curves with a 1.8 gamma, and I was hoping I could find a 
way to turn them into 2.2 gamma response curves (rTRC,gTRC,bTRC)- I was hoping 
there would be with synthcal and colorprof some way to build a very simple 
synthetic icc profile, but I failed at finding a way to turn a synthetic .cal 
file into a .ti3 file that could be used by colprof. Is this possible ?
Solution B : adding a vcgt to the "Generic RGB profile"- I thought I might be 
able to use `applycal` to add the "2.2/1.8 gamma" .cal file to the "Generic RGB 
profile" (the .cal file is here : http://pastebin.com/VAf8M8y0)- I tried 
`applycal -v -a ~/gamma_shift.cal "Generic-rgb_profile.icc 
Generic-rgb_profile_modded.icc`, but got a seg ault- so I tried to use a new 
linear .cal file as a test, built with synthcal -t d linear.cal ; but I got a 
seg fault again- I then tried to compile ArgylLcms from source (instead of the 
1.1.1 binaries for Os X on Snow Leopard) : installed jam, cd to ArgylLcms, ran 
`jam -q -f Jambase -j 3`, but only a few binaries compiles (the log of the 
compile is here : http://pastebin.com/xyp5rmSb)

I would appreciate if someone could suggest something to make solution A or B 
work, or another solution I had not found.
Michael Cinquin 


Other related posts: