1. I'll try. You're seeing the aperture through the changing optical geometry, ie as the aperture changes, so does the lens' light-bending, so it looks the same, automagically compensating. Except on the DA 18-250, where it looks the same from the back but nor from the front. Oy.
2. When my sensors are noisy, I stick in earplugs and pay no attention.
But I was going to ask, does shifting to a lower resolution - setting my K20D to 5mpx, for example - affect the diffraction limit (DL)? (Yeah, that again.) And are there any other (side)effects of down-rezzing? And is there a tech reference somewhere? But then I note that setting resolution ONLY affects JPG output - with RAWs, the final resolution can be selected in development (like PhotoLab3). That answers the DL question - DL is a function of sensor size/density, and that hasn't changed, only the way the pixel site data is put together. Does JPG down-rezzing really increase sensitivity? Is there a reference somewhere?
Since Sony was mentioned, I'll bring up a curious point. Down-rezzing the K20D does not change the crop factor. The full area of the sensor is still being used. With my old 5mpx Sony V1 and P10 PNS cams, however, the crop factor DOES change, with the SmartZoom feature, basically an in-camera crop. Down-rez from 5mpx (2592x1944) to VGA (480x640) and the V1's 4x becomes 16x, the P10's 3x becomes 12x. The effect is less extreme at the 4.5, 3.1, and 1.2mpx resolutions, of course, and it's just a trick, some pre-processing to save on PP work. Hey, it's in-camera, so it's Pure! Somewhat less of the sensor is being used. This SHOULD affect the Diffraction Limit somewhat. Always trade-offs, of course. TANSTAAFL.