I for one, really appreciate the effort put into this data collection and presentation. Data is one thing that often gets downplayed because it is not well understood by many people who are happy with impressions based on perceptions alone. Our brains, which has more to do with impression than eyes do, are not good indicators of what a camera is actually doing and what it can be expected to do in various conditions. The same with sound, our brain completely synthesizes the perceived experience based on very low bandwidth, low quality data with extremely high noise to signal ratios(highly reverberation fields are mostly noise-artifacts of the original wave...like in any real acoustic space), and often fool us as to what is really there. Even the best stereo systems can't fool a 5 year old wearing a blindfold, as to which is a real sonic experience and which is a representation of one in recording. A more experienced listener is often easier to fool because of cultural factors and familiarity with what to expect. Same with photos, a casual observer of an image has an entirely different impression of an image, as a rule, than an experienced photographer because of what additional expectations and viewing habits skew impressions. That is where data and proper testing really shines, to remove the interpretation of image taking to level the playing field for less biased comparisons or indicators of predictive performance. These charts are more useful than all the web based example of test shots. The same scene might be interpreted very differently by viewers with different experiences. I was viewing a portrait on another site and liked it, it gave me a really nice impression of the person, environment and would have been pleased to have taken it. But all the comments posted were all over the map, several saying it was terrible because of a color cast, another said it was not sharp and all I could assume was that my monitor was not high enough res or color tuning but I suspect more that is was because the overall impression of the likeness and likability of the image trumped technical flaws which I never could see. If there were technical flaws, I would not have cared but it would have been nice to know what they really were based on camera characteristics so steps prevent it the next time. I suspect there was no real flaw with the image, that 99% of non-photographer, non-pixel peepers would have liked it. Data helps more than subjective impressions of camera performance overall by isolating camera performance from the environmental issues like light, materials, lens etc. But visual impressions of a particular image, with a unique time element, environment etc and all that impacts subjective impression of a specific image are also important. Both forums of information makes up a more useful and predictive collection of information.
Overall, the D800 is measuring very very good. I will be very interested in the data points below 400ISO to see it if mimics trend line of the D7000 at 100, if so, the D800 will be the undisputed studio and landscape DSLR king for dr alone, even before adding the higher res into the mix. Stan St Petersburg Russia