[argyllcms] Re: dispwin vs xcalib loading LUT

  • From: Graeme Gill <graeme@xxxxxxxxxxxxx>
  • To: argyllcms@xxxxxxxxxxxxx
  • Date: Mon, 01 Jan 2007 15:26:48 +1100

Frédéric wrote:

I attached a profile coming with xcalib package. This profile has a 2.2 gamma. When loaded with xcalib, it does not change anything to my uncalibrated screen (no previous profile loaded). But when loaded with dispwin, the screen becomes very bright.

I'm not sure what xcalib is doing then. The profile clearly has a vcgt
tag with a power curve, power 0.45. Compared to linear, this can't
be loaded into the display and have no effect !

*But*, when I load an Argyll generated profile (also attached), they both give the same result!

The Argyll profile has a very mild boost in the vcgt. This might be typical
of correcting a CRT to have a gamma 2.2 response.

I had a look in the profiles, with iccexamin, and I saw differences; it seems that the vcgtag is linear in Argyll generated profile, even if it is a 2.2 gamma profile.

What do you mean by "2.2 gamma profile" ? The Argyll profile looks appropriate
for a CRT display with a native gamma slightly higher than 2.2. The vcgt tag
corrects this to gamma 2.2, and the profile device curves reflect an overall
response of gamma 2.2. The other profile only make sense if it is working
with a display with a native gamma of 4.8! (ie. native gamma 4.8 * correction 
0.45 =
calibrated behavior 2.2, as reflected in the TRC's gamma 2.2) - ie., it is a 
peculiar profile as far as I can tell.

>  But I also see that the [rgb]TRC tags are, in both
cases, the same (2.2 gamma curved).

Graeme, can you explain me how exaclty dispwin loads LUT? Do you know why xcalib does not give the same results only on its sample file?

Not being the author of xcalib, not without some effort. Can you clarify
a little how xcalib and dispwin behave on these two profiles (ie. description of
four cases).

Graeme Gill.

Other related posts: