Heart Nebula in SHO

[Tommohawk]

Hi  all.

I did this some weeks ago now -  usually I post all the hilarious details of my various screw-ups on the  night (leaving the Bahtinov mask on etc) but TBH I cant remember what I did yesterday never mind 4 weeks back, so just the basics for this one.

Am now using a TS Optics 72mm with 0.79 reducer instead of my trusty 300mm Tamron F2.8 and all looks pretty good though I don't think I have the spacing right just yet for the reducer.

Main issue with this was trying to get detail in the SII and OIII. In the end I just opted for keeping maximum detail in the Ha, and mullered the SII and OIII to breaking point - so a bit like using the Ha as you would a Lum, and using the S and O as you would the RGB. If that makes sense. It was a nightmare trying to control the colour. In the past I've used the Ha flats for all SHO but because the O was so wayward I did separate flats this time. Its helped, but there's a lot of blue toward the top. Similarly a lot of red to the right - but I've fiddled with this for way too long, and that's partly why it's taken so long to post.

So - TS Optics 72mm F6 with 0.79 reducer, ASI1600MM Cool,  32 x SII, 26 x Ha, 36 x OIII all 300 sec, high gain (300) DSS PS 

Hope you like it - as ever criticism or tips welcome, thanks!

[Jkulin]

First thing I noticed was the magenta, so I took your image into PI and used the magenta removal tool, hope you see the improvement.

There some gradients as well.

[Tommohawk]

14 hours ago, Jkulin said:

First thing I noticed was the magenta, so I took your image into PI and used the magenta removal tool, hope you see the improvement.

 

There some gradients as well.

Hi!

One problem I have when processing is that my colour vision isn't good - red green deficient - so I really struggle to balance things. I use the sample tool a lot to make sure things are about right, but usually I run the end result past my wife who then tells me where I've got it wrong! Your version looks better -she tells me!

Actually I can see the difference once its pointed out. I think there are a couple of things causing problems. The Ha is overwhelming, so in SHO there's loads of green. To keep the balance I rack up the S and O but then the balance goes. To make things worse, the star sizes are smaller in Ha, so the bloaty B and R give a magenta-ish halo. 

I don't have PI, but in PS I tweaked the G curves to push the dark and light shades more, rather than just the mids, and this has improved things. I need to find a way of processing O and S whilst keeping the star sizes small - this would help no end I think.

Maybe I just work in Ha/Mono!!

Thanks for the input.

[david_taurus83]

Hi. It looks a bit noisy, more so than I'd expect with that camera. I recently done the Flame and Tadpoles in narrowband, unity gain and 5 min subs. I was a little underwhelmed tbh with the Oiii and Sii as signal was weak and weather wasnt great either. Ha was obviously better. I gathered a similar amount, around 3 hours each. It came out much smoother when I combined channels and added Ha as a Luminance.

Sorry if I'm getting this wrong but how did you apply your data? Ha as L and Oiii and Sii as RGB?

[carastro]

I agree with the colour gradients comments above and the Magenta, but as you have a degree of colour blindness, then of course you need a second pair of eyes to help you with this.

Agree regarding the spacing.

 As regard the Oiii and Sii, I was given a tip some years ago, that if your Ha data is detailed and fairly smooth with small stars, that in fact the Sii and Oiii is merely a colour, and that you can blur them (in Photoshop you can use dust and scratches), If I need to do this, especially if the Oiii and Sii data is more noisy, I blur the final image and then place a further copy of the Ha over the top of the blured image and blend it as Luminance and it works quite well.  I find this particularly useful if I am imaging from home where the LP can make the Oiii and Sii VERY noisy.  Might be worth trying.

Personally I don't think the Ha is overwhelming in this image.   Presume this was blended SHO. 

Carole 

[carastro]

Hope you don't mind but I took the image into Photoshop and had a play.  

Gradient exterminator helped a lot.  I fiddled with the colours and did a small despeckle and High pass on the detailed bits only.

Did you bin this data, as some of the small stars are very blocky?

Carole 

 

[Tommohawk]

Hi David/Carole and thanks for the comments and reworked version.

None of the data was binned, and the SHO was used as RGB with no blending or luminance. When I said I used the Ha like Lum, what I meant was I relied on the Ha for the detail, but I didn't actually use a Lum layer

As you say, the OIII and SII channels (Red and Blue) do look pretty horrible but that's the result of stretching them to blazes. 

Interesting David that you use unity gain for NB. I've found even with high gain its difficult to get any detail in the SII and OIII and so have been reluctant to try unity gain - I don't want to waste time experimenting with  the limited clear sky we have! That said, I recently did the horsehead as Ha(HaR)GB and accidentally left the gain at unity for one session of the Ha and that worked pretty well - it was possible to just stretch it a little more.

I'm not sure why I didn't try using the Ha as Lum -it does seem an obvious way to go, and having tried that it's much better certainly from the noise point of view, and it also helps keep the star sizes down. TBH I think I spent so long (I mean several sessions at the PC) trying to force the raw OIII and SII data without bloating the stars that I got a bit fed up with it!

I'm sure I read somewhere that using Ha as Lum was a bit naff - as it rigs the data. Maybe I'm muddling this with using the Ha as red. I thought Olly P said he wasn't much on using Ha as Lum? Maybe I remembered wrong.

Re gradients - first, Carole you said you used PS to get rid of gradients? Is that a feature I haven't seen or a plug in or something? Oh... I think I misread it - is gradient exterminator a separate app? Part of PI?

Also, I'm not sure I understand why folk get so exercised about gradients! If the gradient is an artifact, say caused by a light leak in the filter, then I guess you'd want to correct it - fair enough. But if the gradient is due to a genuine variation in colour across the image, why would you want to remove it? For that matter, even if the gradient is due to LP, if that's the way it really is, that's OK... isn't it? Maybe I'm missing something!

Thanks for the tips - I'll have a redo and post again. I've done a quick redo with Ha as Lum and blurred the RGB but now the stars have blue haloes because the blurred B has increased the star size. Might take a while!

 

[carastro]

Quote

I'm sure I read somewhere that using Ha as Lum was a bit naff - as it rigs the data. Maybe I'm muddling this with using the Ha as red. I thought Olly P said he wasn't much on using Ha as Lum? Maybe I remembered wrong.

I think what Olly was talking about with when using RGB filters, not to use the Ha as luminance as it has a tendency to wash out the colours.  

Carole 

[carastro]

Quote

Oh... I think I misread it - is gradient exterminator a separate app? Part of PI?

Gradient exterminator is a plugin for Photoshop by Russell Croman.  Well worth getting if you are using photoshop.

Carole 

[Tommohawk]

Thanks for that Carole - I'll check out the plug in.