Integration time and detail / nebulosity

[smr]

This is probably a stupid question but something which has been on my mind now for a bit.

I was of the assumption that integration time is the holy grail of Astrophotography images. Of the Astrophotography of the year books I have I notice that all of the images, not just the winning images in their respective categories, have long integration times, how long is long you ask? Well, everything is relative but I'd say on average about 20-30, quite a few 50+ hours.

Another astrophotographer says however, that there's a myth about integration time, and that you shouldn't expect to reveal more detail and nebulosity if you can't already see it in a small stack - and that stacking more and integrating more just cancels out the noise and makes the image smoother. 

Is this fundamentally correct? I understand that if you have a longer stack you can stretch more without introducing more noise and artifacts than compared to a shorter stack.

For example, if I integrate 5 hours worth of images of M42 and I can see a nice detailed trapezium etc. but hardly any of the surrounding brown dust / gas etc. in the outer regions and around M42, then there's no point in imaging for 15-20 hours if I can see hardly any of that dust with the shorter stack?

 

 

 

[Aramcheck]

Whether  1 hour , 5 hours or more are stacked you will still end up in each case with a single digital image with a set dynamic range (about 14 bit for a Canon DSLR I think). Longer integration will give better Signal to Noise, but that's follows a law of diminishing return. At least that's how I understand it.

[Phillips6549]

Robin Glover gives an excellent summary of the factors involved in determining exposure times in this talk.  Well worth investing an hour to watch it all. 

Mark, 

 

[wimvb]

You already gave the answer in your question. More data gives a higher snr, which allows you to stretch more aggressively. This reveals weaker detail and increases apparent dynamic range.

In a camera, dynamic range = full well / read noise, which has nothing to do with images. But in an image, whether it's a single sub or a stack, the limiting noise is either dark current noise or light pollution noise, which should be higher than read noise, because you expose above the read noise floor.

[CloudMagnet]

Also is worth bearing in mind the long exposure you use, the long it will take to produce your dark images as well. If you use 20 darks for example, this will take a lot longer if you are using 5 minute exposures compared to 30 seconds which will eat into your imaging time as well.