Check my noise please

[kirkster501]

Got my wonderful FSQ and Atik 460 capturing M33 as I speak (type).  All goign lovely!

Could someone check what they think of one of my stretched 1x1 luminance 300 second subs please?  They are taken at -21C.  Does the zoomed in - zoomed in to the top right of the image -  look noisy?  it does to me but i am not an expert on these things.

This is the sub on FITS liberator..

And this is zoomed in:

Thanks for looking.

Steve

[narrowbandpaul]

Hi Steve

Be wary of judging noise just by looking. It's an unfair test.

Better is to take an area of the background and measure the standard deviation in a raw image. This is the noise in DN (ADU). Even this is not fair. The only real unit is the electron. To convert DN to electrons you can approximate with e/DN=Well capacity/Saturated Pixel Value. Well capacity you can get from the camera specs and for the saturation level look for a saturated star and measure the peak brightness.

Then noise (e)= (Well Capacity/Saturation)*Std Dev

If the noise is around the quoted value for read noise then things should be OK. If it's really high then it must have come from a bright background sky.

Hope that helps

Paul

[Gina]

It "looks" normal enough to me for a single sub.  Get a couple of dozen of those subs stacked and you should have a nice image

[kirkster501]

Many thanks Gina!  I was calibrating what was "ok" and "not OK"!

Steve

[davew]

What do you mean by noise Steve ? As there's only one really hot pixel showing then I doubt it's that.

As a specific test I would advise listening to Paul. As a very general test let DSS be your friend. As someone who uses Maxim for capture and calibration, I still use DSS for a " Quick look " re background noise and FWHM drift. If you take your subs ( I do it as I'm capturing sometimes ) and only register them in DSS it gives two very useful results. One is average FWHM and you can use this to watch for focus drift. I only get it with my refractor. The other one is sky background levels. I've seen 0.5% on a good night using Ha and around 5% plus, if I remember correctly, on a bad night with OIII. When it climbs up there, be thinking more about your bed !

These figures are found on the far right of the page. Ignore score !

DSS is free and won't infect your subs As Paul says, you can't really judge a sub by looking at it with a screen stretch.

Dave.

[kirkster501]

Ok thanks guys will look at DSS - very familiar with it from my Canon EOS days (still going those days btw for widefield! )

When I say "noise" I mean that "fog" in the picture when zoomed in.

[narrowbandpaul]

Well that "fog" as you put it is just background. I assume you live somewhere with at least a little light pollution. In that case with an FSQ and a sensitive camera you will pick up these photons. They will inject noise in the image and will almost certainly be the overwhelming source of noise. Even with the LPR filter in it won't filter out all the LP.

So actually the test I proposed could be worthwhile. If the noise is well above the read noise then it certainly indicates that you could do with better filtration. Narrowband imaging is one possibility. Some who live in areas with horrendous LP have this as their only option.

However, looking at your image it seems OK. In all probability you have some noise from the background and a bit of vignetting.

I do recommend to evaluate things numerically where possible. Appearances can be deceptive. Also having a good understanding of noise is pretty useful when imaging. After all that's what it's all about!