101 question - number of subs

[kirkster501]

Guys, I've had a brain stall..... (not uncommon at my advancing years.....?)

So the mantra (and truth) is collecting more subs is better and the more the better.  Yep, I'm sold.  And I see the results. 

But I struggle with one question.  Why not use *the same* sub and stack that one sub numerous times instead of taking different ones?  The same could be said of daytime photos as well, only in terrestrial there is much less advantage in doing so since a 1/1000 exposure is not a big deal.  However, collecting 30 x 20 minute subs IS a big deal.

Is it all about noise being averaged across *different* subs?

[vlaiv]

Yep, we stack subs because of the noise, we need same signal, each sub has same signal, so one sub would have same signal, but each sub has different noise - noise is random, but one sub would always have same noise.

When we average - signal just stays the same, but what is different, sort of cancels out - that being the noise. So no point in stacking single sub - nothing will be different

Edit:

I've just read my post and it is hilarious 

But yea, on each frame signal is the same - average of the same thing is that thing. Noise is always random - average of random tends to 0 (if proper distribution centered around 0 like truly random noise).

[wxsatuser]

SNR!

If you stack two subs, you reduce noise by a factor of 1.4x, if you stack 4 subs 2x, 9 subs by 3x, 16 subs by 4x, 25 subs by 5x, 49 subs by 7x etc etc.