Klaus Karcher wrote:
AFAIK that's what Heidelberg advertises as "Inverse [sic] Gamut Mapping".
Right, so it may be a Heidelberg feature, but other packages may take another approach. Currently I'm happy that Argyll's approach is at least consistent and functional, and I feel no inclination to change it's ICCV2 behaviour.
In my humble comprehesion, the PRMG (http://www.color.org/ICCSpecRevision_22_02_05_PRMG.pdf) is (or shoud be?) a agreed-upon determination of the PCS gamut boundary and a perceptual gamut mapping should map the source gamut to the PRMG without clipping in a "pleasant" way (however defined). In addition, it needs a vendor independent, invertible gamut mapping algorithm to be of use. If the user wants clipping, he has to boost the chroma of the source image to exceed the PRMG limits. Did I miss or fail to see something?
The reference gamut is still to a degree "work in progress". It may not change at this stage, but it's possible it might to some degree, since there has been more analysis and some criticism of it since it has been made public. As for a standard gamut mapping algorithm, I hardly see that as desirable. Gamut mapping is very definitely still a research topic, and will almost certainly be one for many year to come. It's just not desirable to anoint one in a standard. This situation is hardly surprising, since gamut mapping is at least 50% art and taste, not science. Do you really want color management to be frozen at the state of the art 2005, rather than further improving ? It would be far wiser I think, to settle on standards that don't depend on standardizing a gamut mapping algorithm for their interoperabillity. cheers, Graeme Gill.