Frédéric Mantegazza wrote:
This is strange, because I can see a vcgt tag in the profile. It looks like it used the LUT installed in my video card... Well, I think it is not really a problem, as it can't modify the photo frame LUT.
Sorry, you're quite right, it does do that now (it didn't originally). It shouldn't make any difference to to photo frame though.
But the tone curves are rally hugly :o( I also think I'm too far from a calibrated screen to do a correct job... I can play with the luminosity/contrast knobs, but is there a way to do that with the help of the instrument, or just with a special set of patches (and my eyes)? It is hard to figure out where are the best values...
That's actually a reason for adding the shell callout to dispcal, you could then use the control adjustment feedback to set the knobs to get closest to a certain target (if of course, you could automate sending the test image to the display). You could create a test 'chart' of a few key colors (ie. 10% step grey wedges, if you can fit 10 squares and still measure them, or coarser steps if you have to), and compute the relative Y values you expect at a reasonable gamma, and then use spotread -dw to measure the colors to guide your adjustment. Graeme Gill.