Light Pollution: Filter vs. Post Process

[frugal]

Trick question I know, the correct answer is of course Dark Site, but that ain't gonna happen

I have been having a trawl through the archives to see if this has been answered, and I could not find anything answering this specific question. I can find lots of people asking if LP filters were good, and numerous comparisons of shots taken with and without an LP filter (and they certainly look effective). But I could not find anything about the effectiveness of LP filters vs post processing.

What I would like to know is this: Assuming I have tools like PixInsight and PhotoShop, can the light pollution in my images be processed away? Will I get a better S/N by cutting out some light at source? Or will cutting it out during processing be just as good? If I have to raid Frugal's CCD piggy-bank (due to be full about the time Sony stop making CCDs in 2026 at this rate) for £270 for a good LP filter and an adapter to fit it onto the scope, I want to make sure that it is going to be worth the investment and not just "making DBE a bit easier".

To give you an idea of what sort of light pollution I am up against, I live in a semi-rural town with a small city about 20 miles to the North and a large one 30 miles to the South. I have attached a 300s image of the Iris Nebula from a couple of nights ago, and a 600s and 900s image of the Rosette from a while ago.

300s

600s 

900s 

From those 600s and 900s shots I managed to pull out this:

 

[Guest]

Stunning!

[vlaiv]

Yes, LP signal can be removed in post processing. And yes LP filter will help you get better images.

Let me explain. LP has two components - signal and associated noise. Signal part is rather easy to remove (there are some tougher cases when there is gradient over nebula with no clear sky as reference, but then again it can be done as well).

Noise part is problematic, you can't remove it from the image easily and it impacts image along other noise sources. If LP was only signal we would not need LP filters at all. But since there is associated noise with LP signal, and it is associated in the following way: more LP signal - means more LP noise - this is where filter helps, by blocking most of LP it also blocks most of LP associated noise. You still need to remove that little LP that remains in the same way you would with high LP levels but, noise that remains and impacts the image is much less then in the first case.

This is why people use LP filters.

[wimvb]

Generally, a problem should be solved at the root. If too much of the wrong type of light is recorded, you should stop it from reaching the sensor. An lp filter will block some light. Narrow band or narrow rgb filters should be more efficient.

The problem with light pollution isn't just that it introduces noise, it will affect star colour and make registering faint detail impossible, becauae it eats dynamic range. Try catching a dark nebula or ifn with broadband filters from a light polluted site. No matter how many subs you take, I don't think you'll succeed.