Martin Ling schreef: > <snip> > It looked like the "fogginess" was a result of problems in LCMS, most > likely related to limited precision arithmetic. I found similar issues > in ArgyllCMS, which went away if I used its floating point math > options. > That would surprise me. I hope we are speaking about the same 'foggines', but a very instructing experiment would be :
Take dcraw, the mother of all convertors. Compile it in 2 different ways : gcc -o dcraw -O4 dcraw.c -lm -ljpeg -llcms gcc -DSRGB_GAMMA -o dcraw_srgb -O4 dcraw.c -lm -ljpeg -llcms The first (standard) uses BT709 at the moment it writes out its image. The second uses a sRGB gamma at the moment of writing. Take an image and apply both to it. Look with a standard viewer to the sRGB image. Look to the BT709 image (as your viewer will apply sRGB anyway, you get a diff between BT709 and sRGB applied). Note the standard image shows that 'fogginess'. Note also there is no lcms in the process, just the colour matrices. So also dcraw does here a 'beneficial wrong' step by inserting a wrong gamma (that somehow acts as a weird contrast enhancer). > We discussed that a bit in this thread: > > http://sourceforge.net/mailarchive/message.php?msg_name=481D7313.10009%40telenet.be > > Graeme Gill of ArgyllCMS explained the issue for ArgyllCMS here: > > http://www.freelists.org/post/argyllcms/icclink-G-and-source-gamuts-profiles,5 > > The explanation for LCMS is probably similar. > > I never filed an LCMS bug because I didn't have a simple test case > program that didn't involve the rest of ufraw, but it's probably > straightforward to write one. > > For the last year or so, I have been using an ICC profile for my Nikon > D50 which is generated by LCMS from the Adobe matrix, assuming an input > gamma of 2.2. I then use this in ufraw with a gamma of 0.45 (1/2.2), > linearity zero. > > The program I used to generate this is attached, based on Rafael > Espindola's test code from January of last year. > > > Martin > > ------------------------------------------------------------------------ > > ------------------------------------------------------------------------------ > This SF.net email is sponsored by: > High Quality Requirements in a Collaborative Environment. > Download a free trial of Rational Requirements Composer Now! > http://p.sf.net/sfu/www-ibm-com > ------------------------------------------------------------------------ > > _______________________________________________ > ufraw-devel mailing list > ufraw-devel@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/ufraw-devel ------------------------------------------------------------------------------ This SF.net email is sponsored by: High Quality Requirements in a Collaborative Environment. Download a free trial of Rational Requirements Composer Now! http://p.sf.net/sfu/www-ibm-com _______________________________________________ ufraw-devel mailing list ufraw-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ufraw-devel
