Hi Mark, Thanks for the feedback.
Regarding adding Python support to lprof... Not a bad idea. Even more, with Python I could drop entirely Qt, and use only Python for GUI building. Probably this would simplify dependencies as well. Anyway, lprof is now on hibernation until at least 3-4 months, but for sure I will take a look. About the procedure you are outlining, this is exactly what actually does qtProfileChecker. It evaluates patches across profile and compare its obtained Lab with the Target reference. Then, the "Results" table is just the error obtained for each patch. Since it works internally on 16 bits, the quantization due to bitmap is negligible. So far, the better the profile, the less dE. But a word on warning, not all patches are meant to obtain small dE. Take for example column 12. These are highly saturated patches, and maybe several are out of scanner gamut, specially on Wolf targets. On these, the dE will rise since the scanner cannot capture the true color, but error on these patches has not significance at all. Then, is just a personal taste, but I consider a INPUT profile as "good" when the in-gamut patches got a mean error of about dE < 2. A profile with error < 8 is still useable and above this, a piece of crap. But of couse, this depends a lot on the usage. Printer profiles, for example can obtain dE of about 15 and still being good. Best regards, Mart� Maria The little cms project http://www.littlecms.com [EMAIL PROTECTED] ----- Original Message ----- From: "Mark Rubin" <[EMAIL PROTECTED]> To: <[EMAIL PROTECTED]> Sent: Monday, January 27, 2003 7:08 PM Subject: [Lcms-user] lcms-1.10 beta, plus profile check error > Marti: > > Since nobody else has replied, I will say "Thank you" for the 1.10 > beta. I played around with it a little over the weekend -- the Python > interface is great (now I can get rid of all my own Python xyz2lab(), > etc., code). Any chance of also getting the IT8 stuff from lcmsprf.h > and liblprof.a into SWIG? > > As preparation for experimenting with printer profiles (I have read > your previous warnings on the subject) I tried the following: > > 1) Scanned my Wolf Faust / Coloraid IT8.7 target. > > 2) Used qtmeasurementtool to read the scanned image. > > 3) Used qtscannerprofiler to create a profile. > > 4) Used qtmeasurementtool again with the same scanned image, this > time choosing "Options -> Output colorspace -> Lab - Pick from > image and apply input profile" and the profile generated in step > #3. > > 5) Compared the resulting L*a*b* values with those in the original > Coloraid data sheet. > > Comparing by hand, they are close but not the same. (A SWIG/Python > wrapper to the lprof IT8 sheet routines would help writing easy > scripts to compute the error.) What error tolerance would you expect > from the process I described? > > > ===== > -- > MARK > [EMAIL PROTECTED] > > __________________________________________________ > Do you Yahoo!? > Yahoo! Mail Plus - Powerful. Affordable. Sign up now. > http://mailplus.yahoo.com > > > ------------------------------------------------------- > This SF.NET email is sponsored by: > SourceForge Enterprise Edition + IBM + LinuxWorld = Something 2 See! > http://www.vasoftware.com > _______________________________________________ > Lcms-user mailing list > [EMAIL PROTECTED] > https://lists.sourceforge.net/lists/listinfo/lcms-user > > ------------------------------------------------------- This SF.NET email is sponsored by: SourceForge Enterprise Edition + IBM + LinuxWorld = Something 2 See! http://www.vasoftware.com _______________________________________________ Lcms-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/lcms-user
