[argyllcms] Re: Understanding display profiles

Volker Sauer wrote:

I though calibration was the process of bringing the display to a known
state...

Yes it is.

I didn't know that the vcgt tag of the profile which is loaded into the LUT is called calibration as well.

That's where the calibration curve (usually) ends up being
saved. In the process Argyll uses, the calibration is initially
stored in an explicit calibration file *.cal), and it is then
used in the process of reading patches to make the profile (dispread),
and ultimately added into the resulting ICC profile as the
vcgt tag (colprof). If you use the "quick" option, all this is hidden
inside dispcal.

You're writing "if you created a calibration" - I still do not
understand this completely and I'm confused:

what I basically did is:

dispcal -y l -o -t 6500 -O "G550 NEC1990 left" -qh -gs left

Right, so you created a calibration and also a profile that
uses that calibration, and then saves it into the resulting profile
as a vcgt tag.

What exactly needs to set up in the dispaly? If I specify -t 6500, do if
have to set the display to 6500 is will he calibration do that??
What's the difference between the .cal and the .icc file?

Yes, all the aim point options of dispcal setup the calibration.
The .cal file contains the VideoLUT curves that put the display
in the desired calibrated state. The profile then represents
the color behavior of the calibrated display.

What can I do about my special situation - I cannot calibrate the second
screen and calibrating only the first gives an annoying difference in
colors - can I use colormanagment at all this way?

It comes down to what the applications are capable of. Unfortunately one
"application" that is often not color managed is the window manager. If
it was color managed, then you would be able to specify the different
GUI decoration elements in device independent color terms, and the
profile would be effective in matching it to other displays.

Color managed applications should be capable of using the profile
to display images in a manner that is (largely) independent of
the individual display characteristic.

P.S.: I read the docs over an over. Maybe the whole thing with
calibration and profiling needs to be explained differently, since I
still don't get it.

Maybe some broader "background" type explanation ?

Graeme Gill.


Other related posts: