Light Pollution Filters on Mono Cameras

[Whistlin Bob]

I bought my ASI1600mm back in December 2019 and have been very pleased with it ever since- it comfortably outperforms my cooled, modded Canon 550d, especially on narrowband.

However, one area that I've had to work hard to match my previous camera is with Ha regions on galaxies. I used an IDAS D2 filter on the Canon, and it picks out a bit of Ha as per this M81/82 from 2 years ago with a 200p:

I returned to this target last year (in a 130pds) for the image below. As you can see- the hints of red are much weaker:

This is partly down to a smaller scope and I can compensate for this by taking some Ha subs, but I don't understand why that should be needed when the DSLR managed quite well without them. As an experiment I dug out the IDAS filter and attached onto the front of the Coma Corrector and ran an LRGB sequence against the same target (back on the 200p). This was the result:

I'm really pleased with this, to the extent that whilst I took a 40 minute stack of Ha subs, I haven't bothered incorporating them, because I can see that they're not going to add a lot that isn't already there. I'm still not sure about my attempts to leave the dust in, as it looks a bit noisy and patchy, but I can have a play with that later.

The reason for the post is that I've always understood the received wisdom to be that you don't add an LP filter over your LRGB filters and I've never understood why that is. The usual domain for these filters is on an OSC camera over a bayer matrix, and I know from my DSLR days how much they can help with picking out detail. Following this experience I'm definitely inclined to continue using the IDAS filter for LRGB work, even if that's not what the textbooks tell you to do, but I wondered if anyone could shed light on what the thinking is behind all this... is it just a misunderstanding on my part?

[vlaiv]

I'm for using LP filter in combination with regular filters.

Depending on your optical train, there are several options.

You can use LP filter instead of lum filter. I used to do this because I had 1.25" version of LP filter and had problem of using two 1.25" filters stacked. This prevents most of the gradients in luminance - but color data suffers from light pollution.

You can use 2" LP filter on top of your regular imaging train (say first thing after focuser - like screwed into CC or similar) - and that is probably the best option in light polluted location - however, you need to properly manage color shift that results from using LP filter - otherwise you get very funny colors.

Also, make sure that your LP filter blocks UV/IR part of spectrum as not all do and if you don't use luminance with refractors - it can bloat your stars.

[Ken82]

Fantastic result whistling Bob so we’ve been doing it wrong with our mono cams!

I had wondered this for some time, why do I always need to take additional HA subs to bring out those red areas when I didn’t do any such thing with my OSC/dslr. 
 

@vlaivcould it be somewhat of the colour shift you mention causing the lovely red colour ? Or is it likely just due to the higher SNR or something? I wonder if it’s just the IDAS filter causing it and a different LP filter would be different? 
 

Ken 
 

 

[Whistlin Bob]

Thanks @Ken82- yes it looks that way! I think the IDAS has an especially high transmission for Ha.

The colour balance is pretty awful (despite what the marketing says!), but Pixinsight sorts that out easily. 

@vlaiv that's really handy- thanks!

 

[vlaiv]

2 hours ago, Ken82 said:

Fantastic result whistling Bob so we’ve been doing it wrong with our mono cams!

I had wondered this for some time, why do I always need to take additional HA subs to bring out those red areas when I didn’t do any such thing with my OSC/dslr. 
 

@vlaivcould it be somewhat of the colour shift you mention causing the lovely red colour ? Or is it likely just due to the higher SNR or something? I wonder if it’s just the IDAS filter causing it and a different LP filter would be different? 
 

Ken 
 

 

Color shift is due to the fact that LP filters have quite a bit of gaps in their response curve. This throws off color balance and even prevents you from properly recording some colors (lowers camera gamut).

Here is comparison between couple of LPS filters. Notice that R, G and B are not as much affected as for example - neutral grey/sand color of the background. This is because designers of filters tried to get primary colors properly balanced - but that is not the problem - problem is in mixed colors.

You can do couple of tricks to restore this color balance - but they are quite a bit tricky. You need reference device (like a good DSLR or spectroscope) in order to derive color transform matrix that will give closest match to actual colors. Most people will not bother with that.

Ha is very difficult color to properly render in astronomical images for several reasons.

First - it is pure spectral color, meaning that it consists out of single wavelength. No display can produce that color. In fact - only way to really reproduce that color (or any other spectral color) - is to emit that exact wavelength.

Any red that we use in our images will simply not be as saturated and deep red as Ha red is.

Besides saturation - our vision tends to be less sensitive at that wavelength - and perceived brightness tends to be lower. If we want to render Ha wavelength next to regular red - and we assume the two are of the same intensity as light sources - we need to render Ha darker.

It looks something like this:

If left is like pure red from RGB palette - (1,0,0) - then Ha would be darker and deeper and richer (more saturated).

There is very nice trick that you can use to see what Ha color looks like - just take your Ha filter and hold it against the light bulb (or any other source of light - well, except the Sun - it will still be too strong to look at) - you'll see world in monochromatic Ha light and everything will be Ha colored - that way you can get the sense of what Ha color looks like.

You'll often see images like this trying to represent visible spectrum of the light (none of the colors is actual rainbow color as no display can produce accurate spectral colors - they are just close enough match - or in this case - not even that close).

but this one is much better representation of spectrum (found on wikipedia):

Here green is more vibrant, next to green you get that teal of OIII wavelength both violet and red fade to black because we loose sensitivity at ends of spectrum. You'll notice that Ha tends to be that deep rich red color - not bright flat red as in above spectrum.

 

[Whistlin Bob]

Thanks Vlaiv- I like that second spectrum above- that all makes sense. I'm a bit guilty of taking the pretty picture approach to astro images, but I might have a go later at seeing if I can match the Ha to what I can see through a filter held up to a light.