>>3997539>20mp is 3648 pixels high>1\2.3" sensor is 4.55mm highThe post that started this sub-thread didn't mention a camera. What camera are you looking at?
>That's 802 pixels per mm, or 401lp\mmNyquist requires a value >2 for reliable signal detection. Otherwise when the signal is close to being in phase with the sampling it gets lost. 2.2 is the value used in CD Audio, as an example.
>Mtf50 is the figure we look at for resolution.No, that's the figure used to estimate perceived sharpness. Extinction resolution is typically judged by MTF10 or, in some references, MTF5.
>Lens rentals run mtf tests on all the lenses that come through, the very best they have found is the sigma 135mm, which almost hits mtf50 across the frame at 48lp\mmWhen you quote that number you are making multiple mistakes/assumptions. First, the MTF50 mistake. That is sharpness, not resolution. Second, the test is wide open at f/1.8. Third, you're looking at the lines orientated one way (Tan 50 lp/mm) and not the other (Sag 50 lp/mm). And finally, "across the frame." I don't think it makes sense to say whether or not a lens can hit a resolution target based on performance at the edges, especially wide open.
If you exclude the edge the Sigma 135 @ f/1.8 is well above 50 lp/mm at MTF50 at its worst line orientation. If you stop down the edges get a bit worse but the center gets even better. But again, this is all a measure of sharpness. When Lens Rentals tested the same lens to 200 lp/mm, the center was well above MTF10 @ f/4 on Tan lines.
>So for op to have a lens that outresolves his sensor, all he needs is for that lens to resolve a bit more than 8 times as much detail as the highest resolving lens on the market.
A 1/2.3" sensor is going to be behind a relatively small lens. Smaller lenses are easier to manufacture to higher tolerances = higher resolution. Whether his camera hits that target or not I do not know, but again, improving any component improves the final result.
