I love DxO mark reviews but there is one area I remain unenlightened (perhaps in both theory and practice).
My understanding is that lens diffraction will result in loss of detail (resolution) at f-stops higher than a mathematically determined point based on pixel pitch.
If I've got things right an APS-C sensor at 16MP does not suffer diffraction until somewhere between F8 and F11, but at 24MP loss of detail begins somewhere between F5.6 and F8. A FullFrame sensor will, of course, do much better because pixel pitch is larger.
I would *love* to see DxO mark test this. Seems simple (to my possibly challenged little mind), just shoot resolution chart at successive f-stops and record where diffraction kicks-in and limits or reduces resolution. Might be hard to normalize as a number and integrate into overall score, but immensely helpful as a photographer to know for each camera where stopping down begins to trade off not only longer exposure but also loss of detail [probably negating any actual DoF improvement].
I mention this here of course because the Nex-7 vs. Nex5-N is a good place to care [what I really want to know is whether I get equal DoF AND better resolution from the N7, or that (at some F-stop) my equal DoF comes at the price of resolution]. So if I do a lot of high f-stop work to maximize DoF perhaps the Nex-7 doesn't actually give me better resolution? Even more interesting to compare FF to APS-C to get a sense of how much more DoF the FF can capture....
Hope my post isn't too clueless. In any event keep up the terrific work - DxO is a fabulous resource!