I tried not to jump to the discussion, since I am trying to finish an essay that in many way echoes Michael's words. But it is not very often to see level of ignorance in some of the posts. So, just a very few quick points.
We should all calm down and to try to understand what DxO really said. There is a wealth of information in the DxOMark site, but poorly organized and presented. How many bothered to read their technical definitions, formulations, testing methods, and the most important measure, the "3D" "Full SNR" charts, that was inproperly burried deep?
Nikon's clamping the black point at the mean read noise level is not an NR, and does not affect the noise in the shadows, as in the DxO SNR measurements. The lower read noise of D3x is confirmed by eyes in many carefully conducted tests, when shadows are pushed up 2-4 stops, and the different is anything but subtle. Weather the SNR over the 0.1% luminance region important is though totally a different question. If you do not push the shadows up, if you do bracketing on every shots, then the answer is clearly no. But the numbers and the pushed-up shadows do tell the differences the careful numerical analysis could reveal, even it is only an academia issue. And the carefully measured and carefully tested did happn to agree almost perfectly.
Are you aware that, despite of the common quick believe, actually DxOMark revealed even at low ISOs the Canon's noise is better than the D3x's from the midtone up, as noise performance can not be described as a single number even for a fixed ISO? Did you see DxOMark actually said the 5DM2 is the least noisy camera at high ISO? DxO does fail in several very important fronts, for example, to address the resolution issue (so the D3/D700 received overall scores improportional to their image quality, mostly the true resolution and the "looking"), to analysis not only the standard deviation but other noise measures such as luma/chroma noise spectrum and energy-spatial frequency distribution, and to intelligently construct sensible performance merit scores out of these raw numbers, which, exactly as MR said, by themselves can be valid while tend to grossly mis-lead (such as the 18% gray SNR).
When DxOMark said they used ACR for the Phase One P45+ analysis? Calm down, read and think more, understand what the numbers really mean, and do they really help or harm your school of arts.
"Only suckers buy paid reviews", so all the magazine subscribers and buyers are in that category?
Also very interesting to me (but most did not agree), the MF lens quality is a very important contributor to the "MF look" -- my Pentax 67 prime and zoom lense raised my 5DM2 image quality one big step up.
I just want to stop here, as we seem all agree pretty well that DxOMark only provides one side of the story.
P.S.: Happen to be also an audiophile, I fully understand what MR said, and totally love LPs and tubes. Many great photographers are musicians, Ansel Adams and Charlie Cramer included. Photography is a form of art, not science about the numbers, although from time to time we need ways to justify the amount of money we ponder weather to spend, so we want to spend time to fully understand some numbers rather than treating them as total nonsense.
I read following statements in Eyes vs Numbers:
I feel following needs to be said, for Michael's statements may be taken by some readers too literally; this is not for or against any camera but for the better understanding of these issues.
1. All Nikon DSLR cameras cut off a portion of the pixel data, thereby removing pattern noise (one could not name this "noise reduction" without provocing the international Nikon community to violent responses). All Canon DSLRs keep the raw data "uncut". This difference may have caused the ridiculous assertion by DxO, that the D3X has two stops greater dynamic range than the 5D2.
2. Noise reduction in the sense most people understand it will be performed on raw data only by Sonys, at least by the A700 and A900. This step could not be disabled with an earlier firmware version of the A700, but now OFF means OFF, apparently.
3. On-chip (on-sensor) noise reduction occurs with all CMOS sensors, which are inherently more noisy than CCD. However, that is single-pixel noise reduction, which has absolute nothing to do with the noise reduction in the sense as it is done in raw processing software or by the mentioned Sony cameras, namely based on the context of pixels.
4. The raw data of the Phase One cameras is far from being really raw.
Finally, there is nothing wrong in analyzing raw data and making certain conclusions based on that; at least I am doing just that without feeling guilty. In fact, one can make certain findings only based on the raw data. However, one needs to be careful with those findings. I'm afraid, in the DxO evaluation of the P45+ they used ACR for the analysis; if so, that invalidates all of their findings.