Created attachment 6838 [details] 2b.pdf With Ghostscript 9.00 and head (r11850) the color in the attached file is rendered much lighter than with gs8.71 when -dUseCIEColors is specified. With 8.71 the image is closer The command line I'm using for testing: bin/gs -sDEVICE=ppmraw -o test.ppm -dUseCIEColor ./2b.pdf
I will dig into this to find out what is going on. -dUseCIEColors should not really be used any longer but rather the new ICC work flow should be used which is the default case.
(In reply to comment #1) > I will dig into this to find out what is going on. -dUseCIEColors should not > really be used any longer but rather the new ICC work flow should be used which > is the default case. That won't work with pdfwrite until we can write version 2 ICC profiles. We need to be able to set UseCIEColor in order to convert inappropriate device spaces when producing PDF/X or PDF/A. Although as noted yesterday on IRC this doesn't seem to be working when the input is a PDF file (for me). Though clearly it is having some effect here, and the input is PDF....
OK. Yes. The exception on -dUseCIEColor being the pdfwrite case, which Ken and I spoke about at the meeting
OK. So the issue appears to be that the profile that I am generating from the default sRGB CIEABC color space defined in the postscript initialization code is not a good representation. I will dig further into this. However, for optimal color, I would still discourage using -dUseCIEColor unless as discussed above, the user is trying to do something special with the pdfwrite device.
I found the issue with this. Regression testing now and fix should be committed in the next couple days.
Fixed with rev 11934