Rick Gordon wrote:

That being the case, might there not be a case to be made for a hybrid colorimetric rendering that is essentially AbsCol except for pure whites, but pure white where c0,m0,y0,k0 (or R255,G255,B255) exists in the file? I would assume that, if you are converting prior to printing, a blend-if layering of a AbsCol and a RelCol rendering, where the AbsCol would be visible if the value on RelCol layer is 0 to 254.

We played with this type of thing for a couple of customers when at Colorbus, (using Argyll to generate the absolute device link with a fudge white point see: <http://www.argyllcms.com/doc/icclink.html#w>), but really we regarded this as something used for those who really didn't know what they were after. Yes, I suppose it's better than a pure relative colorimetric rendering, but we were usually after a visually identical match, not just something "close". To achieve it needed a whiter proofing stock than the target, a good calibration system, and fine consistent control over light tint levels by the proofing RIP and printer. Often the white point accuracy achieved by the profiles isn't quite good enough, and a fraction of a delta E nudge was needed to get a good enough visual match. Yes, trimming the non-printed areas from the print was important, if viewers were not to be distracted. Overall it is a doable, but slightly painful approach. Printing on the same media as the press (ie. for Newsprint), or choosing a proofing stock with an identical white point was a far more livable proofing solution. Some stocks are close enough, that a slight, almost imperceptible tint is enough to make it a good solution. Graeme Gill.