Tue, 24 Apr 2012 17:55:11 +0400 Alexey Gribunin wrote: Thank You, Alexey! Yes, I've performing analogue tricks in the past but I hope it is possible to improve the robustness of the colprof code in RGB-printer profiling aspect. I'm asking Graeme to take a look at the problem again (I remember i've asked but didn't find old message at a glance). Is there a way to obtain inverse LUT for colorimetric transform with neutral black point which would be capable to produce uniform gray gradient? May be with some ICC hacking, may be optional... But automatically! The aim is to propose the user the profile with perceptually neutral gray gradient from paper white to darkest and perceptually-neutral "black" in relative intent. In this case the profile black point shouldn't to comply with device R=B=B=0 "black" point which may be not neutral or may not be reachable through (along) a uniform neutral gradient. Is such "hack" possible with ICCv2 profile? > Hello Nikolay! > > I think this strange behavior is because of the following issues of your > device's dark area: > 1) Quite dark R0,G0,B0 - L=3.7, a=-0.65, b=-3.6. > 2) At the same time the darkest neutral point is quite light - > approximately L=10. > > I don't think you want to use this device as color proofing device, right? > So I suppose that in this case you don't need good colorimetric > precision, you need just good visual quality. > > So I would try to edit measurements to make it more "correct". I would > try to "move" darkest point R0,G0,B0 from L=3.7, a=-0.65, b=-3.6 to > L=3.7, a=-0.65, b=0. > Certainly, in such case it's necessary to "move" some neighbor values > too e.g. by using some linear equation. > > By using this method, I think, you will get more deep and correct > shadows. Such small blue cast in such deep shadows will be invisible, > for sure. > > Best regards, > Alexey Gribunin, UNIT Color Technologies, Moscow, Russia.