-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 12/09/2015 03:56 PM, KOVÁCS István wrote: > I've just done an experiment. > Someone mentioned on Google+ that he uses > sRGB_IEC61966-2-1_black_scaled.icc from > http://www.color.org/srgbprofiles.xalter. Since I noticed some weird > 'inversions' (dark areas becoming brighter than their surroundings) in > high-contrast versions of a photo with the default 'web-safe' if LCMS2 > was turned on, I decided to follow his advice. I exported 3 versions > of a photo (each version has a different set of operations applied). I > tried all combinations: > LCMS2 on vs off > web-safe vs black_scaled > perceptual vs relative > > So I ended up with 24 output files. I found that the 'perceptual' and > 'relative' files were the same size; they looked the same in Geeqie > (colour management turned on, display profile loaded); and after using > "exiv2 rm *jpg" to delete all metadata, the pair of images was binary > equal.
Are there any pixels in your test photo, which color values are out of the output color space? If not, isn't rendering intent irrelevant? > [...] Šarūnas Burdulis http://math.dartmouth.edu/~sarunas -----BEGIN PGP SIGNATURE----- Version: GnuPG v2 iEYEARECAAYFAlZplRMACgkQVVkpJ1MUn+aQWgCbBq6dCwWehexvpusTwugEMgro TScAn3SPUyixaGrwAixhsCjX6wuMoX0u =j2nb -----END PGP SIGNATURE----- ------------------------------------------------------------------------------ _______________________________________________ Darktable-users mailing list Darktable-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/darktable-users
