I recently got a Datacolor SR print profiling system and have been making some profiles to compare to profiles from Epson, Red River, and one made by Eric Chan (which seems to be really excellent) in preparation to switching my 3880 over to Inkjetfly inks. I figured it would be smarter to learn how to profile with known paper/ink than on something completely new.
The first attempt was for Epson Premium Glossy with the DC 729 target plus grays set - 9 pages in all. After learning how to use the system and kicking the SpyderGuide out the window I was able to get a nice profile that is at least as good as the Epson one but which I was able to tweak to match my profiled monitor quite well. All looked good.
Then I tried my favorite paper - Red River Polar Matte with the same 729 plus grays set and got a total disaster. The overall gamut was significantly smaller than the other profiles and worse, showed really gross banding in B/W prints. Completely and consistently useless even after reprinting the grays pages and remeasuring a couple of times. I contacted Datacolor who have yet to answer.
After considering what was going on, I tried doing another profile with just the 729 color patches (the same I used before) but without the extended grays pages. The result was an excellent profile with no banding and with a gamut that slightly exceeds the other profiles I have for that paper/ink combination. I've done several reprints of images I've done with the other profiles and the DC one is at least as good and is tweakable.
My guess is that the extended grays pages are too finely spaced for that paper/ink combination so the software thinks the paper can't produce good gradients since different patches wind up being measured with essentially the same values and whatever curve fitting function they use goes crazy. That in turn produces a profile that becomes quite non-linear in shadow areas of low saturation. On the other hand if only the 729 color patches are used the grays aren't extensively tested, so my guess is that the software just assumes it's all linear there and that winds up working much better. After that I also tried the 2 page 225 patch target and got a decent profile, but the gamut was a bit better with the 729 patches.
I don't know if this applies to all matte paper, but it is something to keep in mind.
