Quote:
Originally Posted by Tony Leece
I hope this question won't get me laughed at,
|
I hope the answer doesn't get me laughed at
There are no silly questions, just opportunities for learning & growth.
Quote:
Originally Posted by Tony Leece
but i see posts where people have made excellent images using hours of data.
Is there an upper limit to this?
|
Every little bit helps, but it's a matter of diminishing returns for your effort, some sort of inverse exponential function might describe it.
I've heard it suggested that
if your are using sky-limited subexposures then, perhaps there's not much point in going beyond 30 such subexposures, as there will be little improvement in the
perceivable signal to noise ratio beyond that point, pixel peeing aside. It would be good to put that to the test on a high-end rig in good conditions, perhaps comparing 1 subexposure, to 2, 4, 8, 16, 32, 64... sky limited stacked subexposures and compare the result visually and also with Signal-to-Noise Ratio measurement. Out of interest, I don't think it's a coincidence that 30 is considered a useful
minimum sample size when describing/estimating a normal/Gaussian distribution as I believe the image noise is distributed in a similar fashion.
Best
JA