On Jul 7, 2013, at 4:03 PM, Gerhard Fuernkranz <nospam456@xxxxxx> wrote: > A (CLUT) profile would now map this RGB triple to an XYZ color which is > somewhere between the XYZ of the painted patch and the XYZ of the inkjet > patch. Ah -- I understand. Thanks for the clarification. First, if the painting has both metameric matches -- for example, if it's a mixed-media work with both printed and painted samples -- then you're left with a significant problem. The obvious answer is to use a camera that doesn't suffer from that type of metameric failure. The next answer is to see if the failure still exists when using a color conversion filter or a different light source. The answer after that is to apply one of Dr. Berns's multispectral imaging techniques -- presumably using the camera you have and one or more filters. Or, if you're creating a profile just to reproduce a single work of art, you can use the art itself as your chart; make spectrometer measurements of a bunch of bits of solid color and match those up with the RGB samples from the image. HP, I think, has a system to do that automatically, but you could also make an overlay with a piece of paper, cut holes in it for where you want your samples, and sample through the holes and photograph the work with the holey paper in place. And, for a general-purpose profile, a color between the two metamers is probably what you want. Cheers, b&