Tilo Nitzsche wrote:
I have played with it a bit. The measurements now go through fine.
OK, good to know that this problem is solved.
I can build a matrix profile that looks very reasonable, except for neutral colors where it has some color casts. For the patch with the most visible color cast in my test picture, it creates a value of (27, 26, 22), the corresponding value from an i1 Match profile (which looks perfectly neutral) would be (27, 26, 26). (perceptual rendering intent, no blackpoint compensation)
Did you calibrate the screen first ? Commercial systems like i1 Match combine calibration and profiling into one operation, whereas Argyll is a collection of tools, and this is a two step process. A matrix profile will fit much better on a calibrated display. The result is often a bit disappointing if used with an uncalibrated display.
I tried building a lut profile (Adobe RGB as input space) but it's totally screwed up with color casts, dark colors going black, lots of banding. (The stats from profile didn't look too bad, IIRC about 6.5 max dE, 0.6 average dE.)
No idea. Send me the .ti3 and the command line options you used, and I'll take a look at it. Graeme Gill.