How image quality increases with number of subs

[MarkS]

I've been asked a couple of times recently how much difference increasing the number of subs makes to the final image quality.  The technical answer is that if you double the number of subs (and hence the total exposure time) then the signal to noise ratio improves by the square root of 2 i.e. 1.4 and hence you can pull slightly more faint detail out of the image.   The aesthetic answer is that doubling the number of subs leads to only a marginal improvement.  The question is best answered by looking at a practical example - so I've taken my recent Cave Nebula image and stacked 1, 2, 4, 8, 16, 32, 64 subs of 5 minutes so you can see the effect of each doubling.  The images are here:

http://www.markshelley.co.uk/Astronomy/2010/cavesum1.jpg
http://www.markshelley.co.uk/Astronomy/2010/cavesum2.jpg
http://www.markshelley.co.uk/Astronomy/2010/cavesum4.jpg
http://www.markshelley.co.uk/Astronomy/2010/cavesum8.jpg
http://www.markshelley.co.uk/Astronomy/2010/cavesum16.jpg
http://www.markshelley.co.uk/Astronomy/2010/cavesum32.jpg
http://www.markshelley.co.uk/Astronomy/2010/cavesum64.jpg

If you less than double your number of subs then you will see even less difference.

Hope you find this useful!

As a preview here is a single sub:


Mark

[Carole]

A very good post Mark, thanks.

I have read about this before, but seeing it is really good. 

Carole

[Rocket Pooch]

Hi,

This is interesting, I have never taken 64 subs before but there is definately a difference, I guess also it comes down to how noisy your camera is as well.

Chris

[MarkS]

In this particular example 80% of the background noise originated from the background flux i.e. the light pollution.  The remaining 20% came from dark current and CCD read noise.

In other words, even though it was taken from a darkish location, the final image quality is still constrained by light pollution.

Mark

[PhilB]

That's a very interesting set of images. How do you differentiate between background flux and instrumental noise?

[MarkS]

That's a very interesting set of images. How do you differentiate between background flux and instrumental noise?

A very brief summary of a talk I gave at OAS:
http://forum.orpington-astronomy.org.uk/index.php?topic=5488.msg31580#msg31580

Happy to answer further questions ...

[PhilB]

That's quite a list, Mark. Think I'm going to have to buy you a pint, or 2 ........

[MarkS]

Phil,

I've put my Powerpoint presentation in http://www.markshelley.co.uk/Astronomy/technical/
It's called Noise.ppt

It's big (72Mbytes) but it might be useful.

Mark

[PhilB]

Thanks Mark. I've downloaded it. I'll have a look through it later.