Why wouldn't this hypothetical 50ke- camera have a recommended exposure value different than 20x the read noise? This guideline doesn't care about the FWC. It is FWC agnostic. It only cares about the read noise. It works for any camera. This is because once you have sufficiently swamped the read noise, further exposure has greatly diminished and further diminishing value, in the face of mounting detriments. It doesn't matter if the FWC is 10k, 20k or 50k...if you have 1e- read noise, then you only need a background sky of 20e- to sufficiently swamp the read noise. Consider that a pure signal of 20e- affected only by it's own intrinsic shot noise has an SNR of 4.47:1, while a 20e- signal that also has 1e- read noise has an SNR of 4.58:1. This is a difference of 0.2dB. That is meaningless.
If you are not clipping anything once you reach the 20x point, then you could certainly continue to expose. However, the longer your exposures, the more blur they will experience due to seeing, tracking and environmental effects (i.e. wind). The longer your exposures, the more you risk when you have to toss a sub. Such risks can be mitigated by spending more money...and if you have a highly reliable $10k (or more expensive) mount, then you might have little issue with using longer exposures because you may never have to toss any. However at some point, you are going to start clipping information.
I think the bottom line is: are stacking and increasing exposure time equivalent? Both are ways to increase SNR. But does increasing exposure time beyond the minimum required to exceed 20x read noise give you something else that stacking does not?
The reason I ask this is because it seem that the fainter an object is, the longer you need to expose in order to capture an adequate amount of signal. I understand why you wouldn't want to go below 20x read noise. But are there instances where you need to exceed it in order to get enough signal? I belong to a club where the prevailing wisdom is to increase exposure time to "go deeper". Stacking above say 25 subs is considered a waste.
Once you have sufficiently swamped the read noise, for the most part, yes. Stacking and increasing exposure are effectively equivalent. It boils down to a matter of stacking efficiency. You might have 93% stacking efficiency if you follow the 20x guideline. If you expose for longer, say twice as long and stack half as many subs, you might get up to 95% stacking efficiency. That is, of course, unless you start losing subs for any reason. Then, those half-length subs are suddenly giving you an advantage! Lose just one of those double-length subs, and you are now at a signal deficit vs. the single-length subs (even if you also lost one of the single-length subs, too! )
However that is also not really what I've been talking about. I've been talking about the difference between using longer exposures at a dark site, vs. shorter exposures at a light polluted site. The point is, if you expose long enough (whatever that may be for your camera, and the 20x guideline is just a means of figuring out what is long enough), then the total shot noise from a dark site will be the same as the shot noise from the polluted site. If you expose to swamp the read noise the same, then the total shot noise from either site will be the same. For the total shot noise to be the same, then the signals have to be the same, which also means the histograms have to be the same.
Regarding going deeper. The only reason you have to expose for "long" in the first place is because of read noise. If we removed read noise from the equation, then we would have shot noise limited subs. Dark current, skyfog, object signals. These all grow with time. So, 3600 seconds, 1200 seconds, 300 seconds, 100 milliseconds. It wouldn't matter how long your exposure was. You could stack 50000x100ms subs, or expose for 5000 seconds in a single sub. If you have a photon flux of 1e-/min, then you are going to end up with the same 83e- signal regardless. If you have 0.01e-/s dark current, you would end up with 50e- dark current in the end, regardless. You would end up with an SNR of 83/SQRT(83+50), or 7.2:1, regardless.
The only reason to expose longer to go deeper is because of read noise. If you have a lot, then sure, you probably need to expose longer...however, the 20x rule will tell you that. If you have very little read noise, then you don't need to expose as long, and again, the 20x rule will tell you that. It's just a tool, though. All it tells you is how much signal you need to sufficiently swamp the read noise. It doesn't have a perfectly linear relationship with different read noise levels. It actually might under-estimate a bit for cameras that have lower read noise, and over-estimate a bit for cameras that have higher read noise. At least that has been my experience. It's just a tool to give you an idea of what signal you need, and from that it just takes a bit of testing to figure out how long you need to expose. When you compare similar integration times, the differences in final SNR are usually quite small. The differences are usually only a few percent, and in my experience so far with the ASI1600, I am usually well into the 90% stacking efficiency range by using the 20x rule. At low gain, I've exposed for too long in a few cases, and ended up swamping the read noise by over 100x (around 160x, IIRC) in only 60 seconds of L exposure. I had around a 560e- mean background sky signal vs. 3.5e- read noise in my Andromeda L subs. I also clipped over 60 stars...so 60 seconds was clearly too long!
I will say this. At some point, a matter of practicality comes into play. I am always measuring how large my signals are, even while I am imaging. In the case of the Andromeda subs, I decided to stick with what I'd started with, as I did not want to end up stacking a thousand 5-10 second subs. I can't help anyone with where the subjective cutoff is. Personally, with this camera, I aim for somewhere between 200-300 subs, as I feel that's the sweet spot. It recovers enough bits to give me good precision, high quality data...but not so much that I start running into any real problems with FPN (assuming I dithered properly). Not everyone wants to stack 200+ subs. Some people don't even want to stack 100 subs. Those are personal decisions, and have little to do with what might be statistically relevant.
If you still PREFER to stack only 25 subs, then more power to you. With a 16-bit camera and very long exposures, that should be fine. However that doesn't necessarily always work, statistically, for every camera. Using 30 minute subs with the ASI1600 runs into some other practicality issues. Amp glow can get quite significant with subs much longer than 10 minutes, and the additional noise can be somewhat rabid at 20 minutes and beyond. You are probably going to lose more than you'll gain by using 30 minute NB subs with an ASI1600. Plus, stacking only 25 subs with 12-bit data ain't going to do much good for any posterization if you encounter any. You need to stack a good deal more than that to smooth out the quantization error from lower gain settings and the 12-bit ADC.