[argyllcms] Re: Use of calibrated display to profile digital camera

Ben Goren wrote:
My concern with using a printer is similar to Graeme's concern with using a computer monitor. Sure, you can create a wide range of tristimulus values, but you might be using just a few spiky sources to do so. Ideally, you'd want your target to reflect every part of the spectrum so that a real-world object that reflects a very narrow portion of spectrum -- in a range that your printer's inks reflect nothing -- will be accurately accounted for by the profile. 

This is much less of a concern with reflective targets. It's really
hard to make inks with steep spectral transitions, hence the
popularity of fluorescent inks for really edge of gamut colors which
can achieve such things.

The issue with using a printer is similar to a photographic
target like the IT8 - the patches are composed of 3 or more
primary inks, and therefore may not represent the spectral
composition that you want to capture in the real world (but
how do you know what that will be anyway ?). But they
won't be as radically different as a display device though.

The more your capture device resembles the human eye in terms
of its spectral response (the Luther condition), the fewer
problems there will be with this aspect anyway.

I've certainly made scanner test charts using a print process,
although this was for the particular purpose of characterizing
that print process with that scanner, to see whether the
"scanner as colorimeter" approach can work in the ideal case.

Graeme Gill.

Other related posts: