I'm not buying the sensor contact test as you can't be sure the aperture didn't change unless you setup something to compare depth of field. I think this might be a great test for a manual lens.
Other tests to do:
- 2 versions of the same lens, like a Sigma or something (i'm presuming that the sensor invisibly boosts iso for 3rd-party lenses too). 1: a native Canon EF mount, 2 the same lens in Nikon mount via adapter, so the camera doesn't know what aperture really is.
- A manual lens, like a zeiss, samyang, m42-via-adapter, so the camera doesn't know what the aperture is. Take a shot on a 5Dx /1Dsx, and take the same shot on a film-body (if you can develop the film to exact-enough standards),the film should vignette less because film doesn't have the 'wells' like digital sensors have.
- Or same test as before, without the slight-disconnecting of the lens, take one shot normally, then just tape the pins so the camera can't see the aperture for the second shot (if you're using an EF85/1.2L, focus first because you can't with taped pins).
Also, Leica's M-mount digital lenses a) are faster (like 50mm f/0.95), b) have a shorter flange-mount distance, so c) the angle of incidence of photons hitting the sensor is a lot worse than that of your typical dslr. They've got a patent along the lines of micro-mirrors around the edges of the sensor to redirect the light properly, that cuts down wide-aperture digital-sensor vignetting just like invisible-iso-boosting (but without the increased noise).
Back to OP, same here, 14-24 and 200-400mm. That Repro-Nikkor 80mm f/1.0 1:1 macro lens would be nice if it weren't a few grand (and it can fit on Canon anyway via adapter.
The opposite is more fun to speculate, like if I could mount a TS-E 24L on a D800 (if i could afford both, of course).