Klaus Karcher wrote:
You assume that clipping should be avoided in either direction for every kind of profile, I assume that clipping should be avoided only in the "native" direction (device to PCS for input profiles and PCS to device for output and display profiles).
My assumption was that intention was to create a table that implements perceptual intent. Such an intent needs to avoid clipping in the PCS to device direction for output profiles, or it simply isn't a perceptual intent. Everything else flows from there + the ICC V4 spec. I may not be following you correctly, but: > In my interpretation the PCS side of an input profile is a subset of the > PRMG Since the PRMG is a subset of the representable range of PCS, and since input devices don't actually have easily defined gamuts, I can't see how you can make the above assumption. I don't think it changes anything, even if you do, because the source of a link could be either an input profile or output profile A2B. > and the PRMG is a subset of the PCS side of an output device: A real device will (in general) have a gamut that may fall inside and outside the PRMG. > but there may be > clipping in output to output transforms (for colors outside the PRMG) In which case it's not an implementation of a perceptual transform. > If a output device is able to render colors outside the PRMG, they can > only be addressed in output to output transforms. Definitely a problem then, since input devices often have very large gamuts, and you would expect to be able to address a color that is within both an input and output devices gamut through a perceptual link. My explanation does not depend on any distinction between input and output devices (there is no real distinction anyway, there are scenarios in which both forward and backward transforms could be needed of any type of device), and was in fact made on the basis of all the device profiles being output devices. Graeme Gill.