[argyllcms] Display profiling problem

  • From: George Pop <gapop@xxxxxxxxxxx>
  • To: <argyllcms@xxxxxxxxxxxxx>
  • Date: Sun, 25 Mar 2012 02:34:31 +0200










Hello,

I am in over my head trying to solve a display profiling issue. I hope that 
someone more knowledgeable on this mailing list will help me.

I have a monitor (Samsung F2380) with a black crush problem that is not subtle. 
A lot of details in the shadows in photos and video are lost.

For a while I've used this monitor with a profile that I found on the Web. That 
profile has worked wonders, completely solving the black crush problem. Photos 
with a full tonal range can now be displayed with excellent contrast and no 
loss of details. But it has a color cast which I wasn't able to fix using the 
RGB settings in the monitor's menu.

So recently I bought a colorimeter (the Colormunki Create/i1Display2) and have 
used it with dispcalGUI and Argyll 1.3.6. The resulting profiles get rid of the 
color cast, the grays are spot-on. But the black crush problem is back.

I have tried both "Curves + matrix" profiles and "XYZ LUT + swapped matrix". On 
the latter, I activated gamut mapping for perceptual intent (using sRGB from 
Argyll as a source profile) and have successively tried "Luminance matched 
appearance", "Perceptual", and "Perceptual appearance" as gamut mapping 
intents. The "default intent for display device profile" is Perceptual, and I 
also made sure that my color managed apps use perceptual. The differences 
between these profiles are marginal as far as the black crush problem is 
concerned.

My calibration and profile settings are: D65 white point, native white level, 
native black level, gamma 2.2, high quality calibration & profile, BPC on. I've 
also tried native white point, gamma 2.4 (worse), 2.0 (also worse) and L* (this 
did open the blacks a little, but the midtones and highlights are too high).

I don't know what else to try. I've been brute forcing the calibration settings 
with considerable waste of time and constant disappointment.

Can any information be recovered from the profile I found on the Web to figure 
out what makes it work so well for tonal range?

Any suggestions would be welcome.

Thanks,

George




                                          

Other related posts: