I remember reading somewhere that your CCD exposure times for DSOs are ultimately limited by your local light pollution levels. Is this right, and if so what method would you use to determine the maximum exposure time for your local light pollution level? Is it absolute, or can you recover the situation with software? Is it different for LRGB and NB?
In my experience there's only so much you can do about light pollution. It'll always create a background level in your images, which will often be an oddly shaped gradient and potentially a bugbear to remove. I suspect it's easier to remove in LRGB than single-shot colour, but I have no experience to back that up.
I've never done any narrow band imaging but it would seem that the narrower the band, the less likely light pollution will strongly affect your images (particularly in H-alpha - which doesn't match common artificial wavelength).
To determine how long you can image just image an emptyish patch of sky at different exposure lengths and check the histograms of these to see how much back contribution there will be to your images. That's the amount you'll be wanting to cut off your actual images.
The problem with sky glow is that it introduces additional shot noise. You can image in LRGB from a light polluted location but you need to do much longer total exposures to get the same SNR as you can get in a shorter time under dark skies. You may still be able to get decent results for bright objects in a reasonable time but you won't be imaging dim galaxies.
Imaging in narrowband works a lot better because the narrow bandwidth of the filters rejects most of the sky glow.
The gradients you get from light pollution are manageable using software, e.g. DynamicBackgroundExtraction in PixInsight or GradientXterminator for Photoshop.
Thanks for the feedback. I've read the link, which I'm not entirely sure I understand. Is he saying that if the calculation indicates a TargetADU of 2150 I should choose an exposure time so that the background (area 'between' the stars) should have an ADU > 2150 ?
Beg to differ. Taken from Sydney (my heavily light polluted back yard)
All you need is a low noise camera ( ie cooled) good QE, reasonable aperture, accurate calibration frames
and the patience to take a lot of subs to cut through the glow.
Beg to differ. Taken from Sydney (my heavily light polluted back yard)
All you need is a low noise camera ( ie cooled) good QE, reasonable aperture, accurate calibration frames
and the patience to take a lot of subs to cut through the glow.
You poor *******, I though my new light pollution was bad but thats beyond light pollution, thats an environmental hazard. Have you thought about suing them for lack of sleep due to the second Sun they just installed in the sky?
I have had pretty good results using light pollution filters. Yes, I need to take longer subs but I see an increase in the SNR. The two filters I've used are the Hutech IDAS and the AstronomiK CLS. The latter filter really needs longer subs than the Hutech but does a fine job in pretty heavy LP. All the photos on my PBase pages are from North Curl Curl. My most recent is here and used the CLS for luminance and no filter for colour: http://www.pbase.com/prejto/image/155539657
Thanks for the feedback. I've read the link, which I'm not entirely sure I understand. Is he saying that if the calculation indicates a TargetADU of 2150 I should choose an exposure time so that the background (area 'between' the stars) should have an ADU > 2150 ?
Hope you don't mine me dropping in Rick.
Hi Graham. If you get 2150 for the targetADU calculation, then that is what you should aim for in the sky background. You could use longer, but then you lose dynamic range and run the risk of saturating the sensor. You could use less, but then the read noise will become a more significant fraction of the total noise and the SNR decreases. The rule of thumb accounts for varying sky brightness and applies equally well when using filters.
As Peter has stated, you can get around the problem by increasing the total exposure time, but that will not be just a little bit extra. To put some numbers on it, if you have sky which is 10x as bright as true dark sky (eg bright suburban), you will have to expose dim targets for about 10x as long (in total, not sub length) to get the same result as you would from a dark sky - this is what Rick was getting at - you won't be imaging dim galaxies in any reasonable time unless you have a very big aperture. Apart from pollution reducing filters, there is no way around it - software can't help, since the noise is part of the background light and it cannot be separated from dim signals.