Sat, 30 Oct 2010 11:47:36 +1100 Graeme Gill wrote: > Dariusz ?urawski wrote: > > It is exactly what I was thinking about. :) But I'd like to use two icc > > profiles - one taken from monitor's manufacturer and one created with my > > Spyder. Of course it will not give good accuracy, but will be probably > > better than no correction at all. One thing I'm not sure of, is what is a > > real difference between monitors of particular type, so I don't know whether > > is manufacturer's profile accurate or not. > I'm not sure that will work very well, because you will not be getting > measurements from the same display in the same condition. Another > potential problem is that ccmxmake expects absolute (that is, instrument) > readings, and there are known issues with display profiles, the ICC > standard, and recovering absolute instrument readings from display profiles. > If you wanted try try this, you could: 1) Create a set of test values using > targen (a small display test set would do). 2) Use fakeread to create > .ti3 files from the two profiles 3) Manually add tags for the > TARGET_INSTRUMENT > and INSTRUMENT_TYPE_SPECTRAL YES/NO to the resulting .ti3 files :- > "TARGET_INSTRUMENT" > just needs to be set to the name of the instrument, and the reference > file should be labelled with "INSTRUMENT_TYPE_SPECTRAL YES" and the > colorimeter > labelled with "INSTRUMENT_TYPE_SPECTRAL NO". 4) Feed them into ccmxmake -f. > Graeme Gill. Graeme, question about point 1): What RGB values must be set in ti1?: only primaries and white point? Best regards,