>>3985312>but as I said above, mathematically this is absolutely not generally trueIn any domain or application where shot noise dominates, more S = better SNR, period.
>I take issue with that stance from a physics perspective. Given a FF sensor, cutting out a fourth of it and slapping a lens shouldn't inherently have a worse SNR, but this is often implied.Of course reducing light gathering to 1/4 is going to produce worse SNR in a shot noise dominated situation. No implication necessary, it's hard physical fact.
>My reading disagrees, I don't give a damn about your reading. The Internet is full of people who waste time theorizing and writing when they should be collecting and verifying data. The highest DR sensors at PtP and DxO tend to be high pixel density sensors. DR is bound by noise on the shadow size. If larger pixels were generally cleaner then the 12mp A7s series would be the dynamic range kings. They are edged out by higher density sensors.
>I've found nothing that agrees with the idea that pixel density has no impact on noise. Again: on a 1-to-1 basis a larger pixel is cleaner. A 4x larger pixel will be cleaner than a 1x pixel. But it will not necessarily be cleaner than four of those 1x pixels covering the same surface area. From a shot noise perspective they will be identical. But even from a read noise perspective, it could go either way.
>Sony released a camera for low light and included a lower MP density sensor, for example.Pic related. The reason that the A7s series is lower MP is because it is computationally expensive to scale high resolution frames down to 4k while preserving detail and noise (see: R5 overheating). A 12mp camera that can oversample down to 4k will be better in low light than a 50mp camera that has to line skip/pixel bin.
>Incorrect, as you should also be capturing less noise since you have a smaller total area.Shot noise is the square root of the photons captured. Fewer photons = higher proportion from shot noise.
