...You seem to extract from the equation that inherently lower levels of sharpening applied during on-board processing (or off-board, via raw converter), do contribute with less noise and less detail. I do not see any major optical impairement on this test, though.
The hellacious amounts of noise reduction being employed, nowadays, by Nikon can also be verified in some Imaging Resource samples from Nikon D90, for instance.
The overall best performer, with best balance of detail and noise (albeit a relatively small or non-existent amount of Noise Reduction) is, without a doubt, the EOS 450D (or XSi), also shown in those crops.
Nope, what I'm saying is that you can't draw any conclusion one way or another from the tiny images on DPReview where some are in focus and some aren't. It's junk science. If you'd loosen up on taking your anti-Nikon bias to such extremes you'd show a lot more credibility.
You've done some very impressive work to develop your routines to vastly improve the output from the 1D MKIII, and from another similar thread it's clear that after your processing the results are even marginally better than from the D3 (the D3 showed some banding at extremely high ISO's).
I'm just happy that the two majors finally have cameras that compete head to head with respect to noise performance. Even DR is getting reasonable with both systems if you shoot RAW. Finally we're at the point where we can start choosing systems by body controls, AF performance, lenses, build quality and cost instead of sensor technology.
Competition is good for photographers.