I've printed ALOT of images, on a lot of processes. Tell me more about your observations, particularly regarding Dmax!
It isn't solely about Dmax! The tests were designed to look at as much of what's happening within color space as possible.
I can't go into detail for the following reasons:
1. The project is under NDA.
2. Many of the images used to exmaine the results of the profile were developed by Karl Lang (I name I suspect you recognize) and they are his properly. But they are designed specifically to look at many areas of color space and detect issues with the color engines and they do an amazing job of this task.
3. This was CMYK output on very specific kinds of presses and substrates so YMMV depending on what you are profiling.
I can say that it was easy to see some issues with blue mapping from Copra profiles where it won in all other areas. I'll point out that sending reports to them resulted in quick fixes to improve this mapping although not enough to our full satisfaction. But it was really nice to deal with a small company that both recognizes they can improve their profiles and did so quickly enough that we could use them for this job.
For reasons I can't discuss, all the testing was done with Perceptual rendering. That opens a lot of potential can of worms as you know.
None of the various profile packages were perfect, none didn't show some flaws. In the end, it was about selecting which deficiencies we could live with and again, in some cases i1P profiles were used in process A while Copra was used in process B. PROFILER in process C, etc. Again, you have to spend a lot of time and with the right tests to determine this but I'll say, based on this work, i1P was not the winner in all cases, hence my suggestion you throw a lot of images and output into the testing before picking!