[argyllcms] Re: How many patches?

Roger Breton wrote:

Yes, so much so that I always had to throw the towel every time I tried
profiling through the Epson RGB driver. This in spite of all the good
results everyone I know tell me they get.


I would suspect the Epson RGB driver to be strongly non-linear, cf. previous posts about this. I have been trying to profile an Epson Stylus Pro XL+ (antique A3 4 colours model). I am on Linux, and the RGB driver is doing little more than [c,m,y]=[1-r, 1-g, 1-b]. This gives horrible results of course, with a wrong grey axis, but it's much better behaved for profiling.

My experience so far has been to profile with a flatbed scanner (Agfa Snapscan). I purchased an IT8 from Wolf Faust and profiled my scanner. Then ran a profile with 700 patches. So far, the results are better than with the Windows driver. These are the problems I have:
1. there is a green cast on the neutrals, very likely due to the use of the scanner as a colorimeter. The light source in the scanner is a cold cathode tube. I guess it has a spectrum similar to a fluorescent tube.
2. there is a break in the shadows and the black. I tried to scan in [16bits linear] rather than [8bits - gamma 2.0], on a UMAX A3 scanner but there was hardly a difference.
Next things I want to try:
1. neutralize the grey axis with the rgb driver *before* profiling. It should not make any difference, but it may solve the shadow problem.
2. use a webcam (Philips Toucam Pro, Sony CCD) + halogen lamp as a colorimeter.

Can I ask you what colorimeter / spectrometer you are using?
Stephan.


Other related posts: