Quoting Kai-Uwe Behrmann <k...@gmx.de>: > Am 02.06.10, 14:09 +0200 schrieb marti.ma...@littlecms.com: >> The point here is that no gamut information is encoded in the profile. >> If you want to take the 0..255 cube as the gamut, that's fine but does >> not mean the monitor gamut is such cube. For example, a monitor can >> deliver the same physical color on [20, 20, 250] that on [20, 25, 253] >> just because the blue channel cannot be more chromatic. Then you got >> an apparent gamut the real device cannot render. > > So one has to rely on the profile class, which is unfortunedly of > "monitor" for sRGB for historical reasons. The other not always > possible way is to ask the context for a hint if the profile decribes > a device or not a device.
Well.. you can still use this method on all profiles. It is just a matter of using two separate transforms when checking each color, one transform from TYPE_XYZ_DBL to TYPE_RGB_8 and other from TYPE_RGB_8 to TYPE_XYZ_DBL. Then you will detect the gamut introduced by the RGB_8 encoding (as well as the quantization errors) Marti ------------------------------------------------------------------------------ _______________________________________________ Lcms-user mailing list Lcms-user@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/lcms-user