Fundamentally a crop frame camera is every bit like an MP3 to CD audio quality. Take this comparison of a 7d vs a 5dII on the exact 50mm lens, both below the defraction limit. The 5dII image has been resized so that resolution isn't a factor. The tripod has been moved back and forth and a prime has been used so that zoom isn't a factor:
Same lens same everything, just cold hard physics. Now tell me that full frame is no longer relevant. I've actually done the comparisons both back to back in the real world and in theory and this works due to simple physics.
First off, I agree that the 5DII will generally produce superior IQ to the 7D. But, I will point out a few issues with your test:
First, diffraction begins to affect the 7D at f/6.8, so your shot at f/8 is affected by diffraction. Shooting at f/6.3 would be a better option.
Second, you aren't comparing the 'same everything' since by moving the 7D further from the subject to maintain framing, but keeping the aperture set to f/8, the shot taken with the 7D has a deeper DoF than that with the 5DII. To match DoF and thus keep everything
the same, you'd have to either set the aperture on the 5DII to f/13 (which would also be a bit past the point at which diffraction begins to affect the 5DII, f/10.3), or set the 7D to f/5.
Third, "The 5dII image has been resized so that resolution isn't a factor,
" isn't the case. A slight downsampling with a good interpolator (e.g. PS bicubic) actually increases sharpness relative to the original. To make things equal, you could downsample or upsample both images (the relative amount of resampling would be different, but it would be a fairer comparison than resampling one image and not the other).
So, of those three factors, 2 of them (diffraction and resampling) artifically skew the results in favor of the 5DII, and the third (DoF) is neutral (since your entire subject appears to be within the DoF, but if not, the thinner DoF would artifically detract from the 5DII).
I don't disagree with the conclusion - the 5DII offers better IQ - but the differential may not be as great as your testing shows.
Another factor is the AA filter - the high density of the 7D's sensor requires a stronger AA filter, and that reduces actuance (contrast), which makes the resulting images appear less sharp.
The truth is, no matter what anyone does a larger sensor will always capture more detail than a smaller sensor assuming everything is equal. On top of that a larger sensor captures more light so it has less noise, has better DOF, and has a wider FOV.
Not necessarily. In the case you mention, where you alter the subject distance to match the framing, yes. But if you don't do that, with the narrower angle of view afforded by the crop sensor, the 7D will outresolve the 5DII. That's just basic math - since sensor size does not affect magnification (only AoV), with the same focal length at the same distance, a given subject will cover more pixels of the 7D's 18 MP sensor than the 5DII's 21 MP sensor. Cropping the 5DII image to the FoV of the 7D results in an 8 MP image. The relevance is for situations where you're focal length limited (small/distant subjects), you'll get better results shooting with a 7D than shooting with a 5DII and cropping the resulting image.
Also, 'better' is a relative term. The FF sensor will have less noise (1.3 stops less), wider FoV, and shallower DoF (for the same subject framing). If you want
deep DoF, FF is not 'better' (although assuming your FF sensor has larger pixels, you can stop down further before diffraction sets in, so it may be a wash).