shigzeo said:
And because of this, the pixels are allowed to bathe better, getting more light and, consequently, look sharper.
Yes and no, mostly no.
MF is now, and always will be, sharper than smaller formats (35mm) because an object of any given size (the whiskers on your cat) will be projected larger on the MF sensor/film. Every visualized physical object takes up more room (on the sensor/film, "in" the lens, etc) and therefore is less susceptible to lens flaws, sensor/film flaws, and diffraction effects.
This is physics. It can not be beat. It can be fought, and a valiant fight is underway, but the larger latent image will always win. The small might get better (diffraction is a hard wall they all will hit, eventually) but the large will get better too.
shigzeo said:
That is great, but it doesn't answer my quandary: the current FF sensors are only up to D200 levels, maybe a bit more in terms of sheer vertical/horizontal pixel resolution for a given physical constrain. The D7000 is 1,6x that resolution and the shards of light that hit its (smaller) sensor, do have to fight with the fact that the sensor does not collect as much light.
As I said, the primary problem (in this case) isn't that the sensor doesn't collect as much light (though that does lead to worse signal to noise ratios, which is a secondary cause of softness) but that lenses are not, and will never be, perfect and thus the "fuzzy edge" of every line is now capable of being imaged on its own photosite.
These "fuzzy edges" have always existed, it is just in lower-density mediums (medium format film/sensors, older digital sensors) the fuzzy edges fell on the same photosite as the "hard edges" and thus their secondary signal contribution was overwhelmed by the primary signal and one never saw them.
As I said in my last post you are not incorrect that some lenses are now possibly being out-resolved (though you didn't prove this point and I'm not ready to concede it). Where you got off on the wrong foot (and the part I'm challenging you to think about) is in saying this is somehow bad.
You got off on the wrong foot (and appear to still be standing on it) by presuming that pixel-level softness somehow equates to image softness in this example.
I also strongly challenge your conclusions based upon looking at those press images. Every test shows the D90 and D7000 sensors reaching the same resolution extinction point regardless of lens (within reason), strongly suggesting the limiting factor is the AA filter over the sensor, not the extinction point of the lens. If you are seeing images which contradict this the safest conclusion is that other factors than raw native resolving power are at play. Yes, bench tests are the only true way to determine the extinction point.