As you can see the 7D actually does pretty well. I'm not saying that it doesn't do ok (I think I might have ruffled some feathers before , I'm just saying that I would prefer something that is better than that.
I'm not a big fan of 7D, but it's nothing to do with image quality. So, I am not upset by what you say about the 7D.
However, the belief that high pixel density automatically translates into poorer IMAGE quality is a myth.
This was the sample I had seen a while ago showing the noise difference between the 7D and the 5D mkII.
... I think that is one of the reasons why the 5DmkII has better image quality. The "more pixels is better crowd" might argue that the 5DII has more pixels, but come on, it is only 3mp more (which isn't that much) and when you look at the pixel level the 5DII really is doing better.
While the 5D2 may be better because it has a larger sensor (nothing to do with its absolute pixel count), the difference is really not astounding (i.e., definitely not 2 to 3 stops better like what many people think). Let me quote from your web link:
"While the differences aren't HUGE, the 5D Mark II's full-size sensor definitely made a difference at ISO 3200 and beyond. Of course that means both cameras were fairly well matched up through ISO 1600, not an easy feat for an 18MP cropped-sensor camera."
I should also point out that in the above link, the 7D shots are darker (i.e., lower exposure) than the 5D2.
Another problem with a more tightly packed sensor is diffraction, but that is a whole other topic. I don't know all of the technical end of it, and I don't want to start a big argument, so I'll quit w/ it here. Just wanted to show that the less density/better pixels crowd has a good point too
That is a myth that can be debunked most easily. If you compare images from the 12 MP 450D against the 18 MP 7D side-by-side, all the way from f/2.8 to f/16 using the same lens,
you'll see they are EQUALLY affected by diffraction beyond f/8. The 7D is not in any way worse. Some other points to note:
(a) diffraction occurs at the lens, and the sensor merely records what has already taken place
(b) diffraction believers forget about the AA filter in front of the sensor, its strength can be lowered with increasing pixel density
Oh, and I just found one more thing that proved me wrong on an earlier thought I had...DxO Mark is an independent lab that tests camera sensors for all kinds of things (sn ratio, dynamic range, tonal range, color sensitivity, etc).
DXOMark is full of garbage. Their numbers have no real world meaning (in fact, they contradict what you actually SEE from DPReview etc) and they are not even self-consistent. I do not wish to go into all the gory details here unless you really want me to.