This is an interesting question, because it raises the issue as to what the quality criteria should be. There is certainly more to the ‘quality’ of a profile than gamut volume, for example. The criteria probably vary according to whether we’re talking about a printer profile, monitor profile, scanner profile or camera profile as well. They also vary from one person to another e.g. portrait photographer versus landscape.
How about coming up with a list of criteria that we might want ColorThink to test for us?
For printer profiles, how about these, in no particular order:
(a) greyscale neutrality;
(b) shadow detail;
© smooth rendition of gradients;
(d) total gamut volume (in some suitable units);
(e) maximum saturation for the primaries;
(f) colour accuracy for a selected palette of colours.
In terms of ColorThink, (d) and (e) could presumably be measured in situ without using any reference data.
(a) and (f) could perhaps be done by soft proofing a test file, for example the target used to create the printer profile (tagged for this purpose with a suitable colour space profile of course).
(b) could maybe be derived from (f) by using a file containing a large number of ‘dark’ patches.
© is a bit harder to figure out. In Photoshop you can just soft proof a file containing gradients & check visually. I’m not sure how ColorThink could calculate ‘smoothness’; maybe Steve knows a way to do this.