Narrowband Filter advice for DSO imaging

[alan4908]

I'm looking for some advice on how to improve my DSO images through the use of narrowband filters.  Specifically, I'd like to improve the images taken from emission nebula with my colour CCD camera.  I understand that a common way to capture emission nebula images with mono cameras is to use a Ha filter to capture the Luminance and then use R, G, B filters to capture the colour. I'm thinking of using the same technique with my CCD camera eg insert an Ha filter in front of my CCD, capture a Luminescence image and then post process the result into my colour image. At the Ha wavelength (656.281nm), the relative sensitivity of my CCD is about 0.78, if I assume that only 90% of the light is transmitted by the Ha filter, this would be reduced to about 0.7, which seems acceptable. I've attached the relative sensitivity of my camera and my latest attempt at the heart nebula taken with my colour CCD (7.5 hours exposure).

Assuming that this would work in practice and given that you can buy Ha filters in 3nm, 6nm, 7nm, 12nm and 35nm variants, which pass band would give the maximum image enhancement ?

Alan

[alan4908]

Is there any chance of a reply from someone who might of tried this ?

Alan

[ollypenrice]

I have never tried this but plenty of DSLR users use Ha filters, depite capturing on only a quarter of their pixels. Obviously a mono sensor works far better but a colour one can benefit from Ha. You'll then need to discard your green and blue channels and, presumably, resize the Ha to match the RGB. I haven't done this but it certainly can be done.

I wouldn't use Ha as luminance to any great extent, though. A good few years ago we used to see lots of images done like this, all very pale pink with blue star haloes, dodgy star colour and emaciated reflection nebulae! There is a much better way. Add Ha to the red channel in Photoshop's Blend Mode Lighten, so where the Ha is brighter than the red it's applied. It isn't brighter for the stars so these remain unscathed. Afterwards a very low opacity Ha layer can be used as luminance on some targets but if you have reflection nebulae in there as well they will obviously bne subdued since Ha is red.

In terms of which filter, I can't say how they would all work on a colour chip. I use 3nm Astrodon and 7nm Baader. The 3 is incomparably more moon-proof, gives far smaller stars and sometimes find more structure and contrast but not always. It varies from nebula to nebula. If I had to pick only one it would certainly be the Astrodon.

Here are some 'Ha into red in blend mode lighten' efforts...

https://ollypenrice.smugmug.com/Other/Best-of-Les-Granges/i-3bjPNHd/0/X3/Mopdified%20Straton%20FIN%20web-X3.jpg

https://ollypenrice.smugmug.com/Other/Best-of-Les-Granges/i-dwsBgd9/0/X3/Cocoon%20to%20Tulip%20HaRGB%2022%20hours-X3.jpg

https://ollypenrice.smugmug.com/Other/Best-of-Les-Granges/i-4pF8WgJ/0/X3/IC443%20HaOIIILRGB%2032%20hours%20web-X3.jpg

It might be worth posting your question under the title 'DSLR with Ha filter?' in order to generate a response. For your purposes the answers will carry over.

Olly

[carastro]

I have done similar with a DSLR for colour and added at a later date Ha from my CCD camera, so what you are suggesting can be done.  

I can't comment on the Ha pass band with a colour camera but I would have thought it would be quite difficult to capture much data with the the lower band widths and actually seeing a star to achieve focus will be more difficult in the lower bandwidths.

I know there are people around who regularly image in Ha with DSLRs, which is essentially and uncooled OSC camera, hopefully some-one will give the information you need.

Carole 

[alan4908]

Olly - thanks for the reply.  I believe that I (almost) understand your suggestion. To paraphrase: make the Ha as an additional red channel (since it will be capturing part of the red spectrum), disregard the green and blue channels (since they will just contain noise) and blend the resulting Ha channel with the OSC red channel using PS blend mode lighten. Effectively, this will create a

"a enhanced" red channel. The item that I don't understand is why would I have to resize the Ha image to match the RGB ? - I can understand that Ha stars would be smaller and presumably dimmer that the stars captured in a OSC image but if you use the PS technique you suggested wouldn't this effect be largely eliminated ?

Carole - on the data capture and focusing - they are good points. Having had a think about this, I don't see why it would be too dissimilar from the situation presented with Ha capture with a mono camera except for the fact that I will be at approx. 0.7 of the relative peak spectral efficiency of my OSC - so, although it will be an impact, hopefully it will not be large. If it doesn't work then I guess this is a good excuse for an investment in a new mono CCD and a filter set

Alan

[ollypenrice]

Your paraphrase does say just what I meant.

Regarding the size, I don't know how the stripping out of R and G works. I was thinking that with a quarter as many active pixels the image might be downsized after this process but maybe the software interpolates the missing pixels spaces and retains the same image dimensions as before. This is rather basic but I'm afraid I don't know it, never having tried it!

Plenty of people on here will know though.

Olly

[JohnC64]

Hi Alan,

I've done this with a DLSR, both to take the RGB and then to take the Ha (with a clip in filter).  With the Ha image I needed to zero the G and B fields of the image as they just contained noise.  As your OSC CCD is cooled I would expect that this may not be required but it doesn't harm as the Ha should all be contained in the red channel.  I used pixel math in PI to do manipulate the Ha image.  

At that point, if i remember correctly, I blended the Ha into the red channel as Olly describes above to enhance the reds.  I didn't get as good a result if I used the Ha as a L as I ended up with a salmon pinkish image with a cyan fringe around the stars which looked unnatural somehow.

As for Ha filters, I use an Astronomic 12nm clip in filter for  the Canon, which gives reasonable results.  I also use a 3nm Astrodome with my CCD and it is markedly sharper and gives better contrast.

Cheers John

 

[alan4908]

Olly - Thanks for the reply. On the size point, unless anyone suggests otherwise,  my working assumption will be that I don't need to do anything eg I will assume that the PS software is intelligent enough to do anything required. 

John - it's encouraging to hear that you obtained the most optimum image enhancement if you went for a "Ha enhanced red" approach, as suggested by Olly.  I think I will invest in a 3nm Ha filter for my OSC CCD, in the hope that this will maximum the contrast with the OSC image. I presume that I will need to expose for a longer period when the Ha filter is in place, do you have any guidance on the relative increase ?

Alan

[JohnC64]

Hi Alan,

When using the DLSR, I am hitting an issue with noise of the sensor for long exposures so am limited in my exposure time to around 5 mins.   So I end up with RBB of 3 mins and Ha of 5 mins . So Ha around ~2x longer that the RGB.

As you are using a cooled CCD, you should be able to go for much longer subs so would aim for 3x as long if possible.  I don't think that there is a hard and fast rule, as  to some extent, you can bring the levels on the Ha up using processing before doing the merge to get an aesthetically pleasing enhance.

Look forward to seeing any results

Cheers John 

[alan4908]

John

OK - I normally use 10min subs so I'll increase this to somewhere between 20 and 30min for the Ha.  I've now ordered the 3nm Ha filter, so I just need to wait for it to arrive and for the clouds to clear.    

Alan

[ollypenrice]

The 3nm is phenomenally effective against unwanted light, be it moonlight or LP. I would expect you to be able to image for 30 minutes without difficulty. That's my standard sub length, though I can do that in luminance as well since I have no LP to contend with.

Regarding how to combine it, there are a few things on your side if you add to red in blend mode lighten. Firstly the noise in your Ha image will be mostly in the dark stuff so it won't be added in this blend mode because it won't be as bright as the same region in red. This allows you to overstretch, considerably, your Ha and still not import its noise into the red.

(When you have the Ha on top of the red in BM Lighten you can see what it's adding. If it is adding little or nothing you can attack it in Curves as it is stitting there. Pin the low part of the Curve and lift it above that, watching for when it starts to lift above the red. Even though you may now be wildly overstretched it is not the noisy parts of the Ha you are adding.

Secondly the Ha will not be as bright as the red on the stars so it won't affect them. This is important.

Olly

[alan4908]

Olly - thanks for the advice.

On the light pollution/moon front. Fortunately, the main light pollution that I get is from the moon.  Given that I've just purchased ACP Expert, I'd like to explore the best configuration option for imaging at 3mn with my OSC. For a particular target, I'm able to pick from three options: a) moon down;  any moon; c) moon avoid. The latter will wait until the moon is at a certain angular distance from the target with the distance varying with the moon's phase. Given the low number of clear moon less nights we get in the UK, I'd appreciate your thoughts on the best option.

Alan