Matching bandpass for narrowband

[codemonkey]

Last night, along with my new EQ8 (!!), my amazing wife brought home a late birthday present; a 7nm Baader H-alpha filter.

I'd been planning on buying myself a 5nm Ha along with 3nm OIII and SII as I've read that the star sizes are typically different with the different filters and it can cause difficulties when processing them as false colour. 5 for ha and 3 for the others seems to be commonly recommended to get star sizes closer to matching.

Now that I have a 7nm instead of the 5 I was thinking about, what size bandpass should I be looking at for the OIII and SII? Still 3nm? Or would I be better off with something like 5nm now?

[Amaranthus]

How bad (or negligible) is your light pollution? Will you be imaging regularly in moonlight? These are the things to really consider. Narrower bands will reduce extraneous inputs, at the cost of longer integration for the same SNR. I think 5 (or 7) is a pretty good compromise - 3nm is for seriously constrained NB! Heck, I image at 12nm (http://www.astronomik.com/en/photographic-filters/h-alpha-12nm-ccd-filter.html etc.), but I am in dark skies.

[pluto]

I did a couple of comparisons when I got my narrower filters here:
http://www.iceinspace.com.au/forum/showthread.php?t=125922
And here:
http://www.iceinspace.com.au/forum/showthread.php?t=126305

Makes a huge difference from my extremely light polluted skies in urban Sydney. Balance wise I went with 5nm Ha and 3nm SII and OIII which definitely made it easier to balance the magenta stars from the SII channel.

[codemonkey]

Thanks Barry :-)

I'm in pretty dark skies, on 5 acres halfway between Kilcoy and Woodford, so I'd estimate about 15km to the nearest streetlight.

Imaging when the moon is out would be an advantage, but not 100% required. I'm certain the 5nm would be fine in terms of light pollution.

The main thing I'm not sure about is matching star sizes when processing the images. Will a 7nm Ha and a 5nm OIII/SII match well? Will I have colour halos that are difficult to process out?



Whoa! I'd found your Ha thread this morning but hadn't seen the OIII. That made a massive difference there.

[Amaranthus]

Alternatively, don't worry about star sizes and instead replace them with RGB stars in post processing. It's actually easy to do and quick to acquire.

[RickS]

You won't match star sizes just by picking "matching" filter bandwidths. The cause of magenta halos is that most objects are much brighter in Ha than OIII and SII, so you need to stretch the OIII and SII data more regardless of filter bandwidth. You can fix this during processing in a variety of ways.

For Ha, a 7nm (or 5nm) filter will include NII as well as Ha which can be a good thing.

If you can afford it, a 3nm OIII filter will do the best job of rejecting moonlight.

For SII a 5nm should be fine.

Cheers,
Rick.

[codemonkey]

Hadn't really considered doing that; I tend to like the more reduced stars you see in narrowband, but that's definitely an idea worth considering. Thanks Barry :-)



Thanks Rick :-)

I've not seen much mention on this forum of doing so, but on CN and SGL it seems to be a very common approach, supposedly for the reason you specify: the theory is that you tend to stretch Ha less so stars appear to bloat in OIII and SII because of the amount of stretching you need to apply. That's the theory anyway, and so people seem to recommend going 5 for Ha and 3 for the others. I was basing my posts above on this.

Thinking more critically about it, maybe it doesn't matter that much. I can probably tweak the star sizes in PixInsight using MorphologicalTransformation, I just wanted to make sure that I wasn't making my life more difficult.

[RickS]

The problem is that the star size differential will vary depending on the object. A few things I have tried that worked despite size mismatch:
using MT or deconvolution to reduce star sizes in OII and SII
JP Metsavanio style tone mapping which removes the stars altogether for the colour processing
replacing the stars with RGB stars as suggested by Barry


Cheers,
Rick.

[RickS]

Thinking about it, I can even provide examples :)

Matching star sizes with MT: http://www.astrobin.com/122474/

Tone mapping: http://www.astrobin.com/138813/

RGB stars: http://www.astrobin.com/full/60763/0/

Cheers,
Rick.

[codemonkey]

Ah, really? That's interesting! I would have thought it would be consistent and the target irrelevant. To alleviate me of some of my ignorance, would you mind briefly explaining that?

Edit: Nevermind, reasons are obvious. You stretch them differently depending on the makeup of the nebula which causes different star sizes.



Cracker shots mate! Personally I prefer the stars in the MT version. I think I'll try and get what you've suggested on this one: 3nm for the OIII and 5nm on the SII which I believe will have less of an issue with the moon?