Recommendations on how to improve my image of the heart nebula - post processing

[AstroLearnerWill]

Hello, I'm a beginner and have been following some tutorials on pre and post-processing, but I'm sure I can do better with this nebula!

I managed to get some clear skies last night and took around 30x160s lights and the same biases, flats, and darks. Using an 80mm SW refractor, HEQ5 Pro, flattener, 600d astromodified. I'm using an l-enhance filter for the first time too.

I've used siril and this guide to both pre-process all of the frames into one image, and then some basic post-processing as mentioned in the tutorial. Which produced the image attached to this post.

After pre-processing and photometric colour calibration, I see a lot of detail in the red channel, more so than comes out in the final post-processed image. I'm not sure but based on this can I do a better of job of getting all of that detail from the red channel to show in the final image?

Any advice on how to produce a better image with this nebula would be much appreciated.

I've used Siril and Gimp so far but I also have a trial licence for PixInsight I can use too.

Thanks!

[Rallemikken]

1 hour ago, AstroLearnerWill said:

I've used Siril and Gimp so far

Siril and Gimp are all you need, but it takes time to master both. And don't rely to much on tutorials, try to find a workflow that suits you and get the most out of your data. In this case it looks like your color balance is way out.

As a start, you need a good fits-file. I have added a script to Siril, and runs it in semi-automatic mode. See attached file. Prepare stack as usually, after it finish run "cd process" on the command line and "Search sequences" under the sequence tab. Choose the last one, r_pp_light_.seq, and head over to the "Plot" tab. Learn how to exclude frames, and prepare the stack with only the best left. DO NOT be to eager to dump frames with high FWHM. When skies drift over the FWHM drops, but the signal gets worse. Many times I drop the frames with the lowest FWHM of this reason, the signal sits in the darkest frames. When it comes to roundness, be a little more brutal. Then the stacking tab: I mostly use Average stacking with rejection, Additive with scaling, Sigma Clipping 2.5 / 2.5  Once you have a fits-file, copy it out of the Process folder and start to play. Take a look at http://www.agle.no/astro/vis_enkeltbilde.php?indeksnummer=54  Pretty similar gear.

This is the exif of what I did in Siril, in this order:

Photometric CC
SCNR (type=0, amount=1.00, preserve=true)
Background extraction (Correction: Subtraction)
Median Filter (filter=5x5 px)
Asinh Transformation: (stretch=   8.0, bp=0.00000)
Asinh Transformation: (stretch=   7.0, bp=0.00000)
Asinh Transformation: (stretch=   6.0, bp=0.00000)
Asinh Transformation: (stretch=   5.0, bp=0.00000)
Asinh Transformation: (stretch=   4.0, bp=0.00000)
Asinh Transformation: (stretch=   3.0, bp=0.00000)

The SCNR is the "Remove green"-filter. The new background extraction in Siril has become really good, with preview and no loss in signal. Rest is done in Gimp. Play around in Siril with the same fits, and save different versions from this as tif's.  See what you can make the most of in Gimp. Every image has it's own quirks. Don't drift to much between different applications. They all do the same, it's just a matter of learning to master each one. There is nothing in Pixinsight or Photoshop that can't be done in Siril and Gimp. Just take your time, and DONT DELETE old datasets. In a year or two you will brush the dust of them, and present new images as you learn.

OSC_Preprocessing_no_stack.ssf

[AstroLearnerWill]

10 hours ago, Rallemikken said:

Siril and Gimp are all you need, but it takes time to master both. And don't rely to much on tutorials, try to find a workflow that suits you and get the most out of your data. In this case it looks like your color balance is way out.

As a start, you need a good fits-file. I have added a script to Siril, and runs it in semi-automatic mode. See attached file. Prepare stack as usually, after it finish run "cd process" on the command line and "Search sequences" under the sequence tab. Choose the last one, r_pp_light_.seq, and head over to the "Plot" tab. Learn how to exclude frames, and prepare the stack with only the best left. DO NOT be to eager to dump frames with high FWHM. When skies drift over the FWHM drops, but the signal gets worse. Many times I drop the frames with the lowest FWHM of this reason, the signal sits in the darkest frames. When it comes to roundness, be a little more brutal. Then the stacking tab: I mostly use Average stacking with rejection, Additive with scaling, Sigma Clipping 2.5 / 2.5  Once you have a fits-file, copy it out of the Process folder and start to play. Take a look at http://www.agle.no/astro/vis_enkeltbilde.php?indeksnummer=54  Pretty similar gear.

This is the exif of what I did in Siril, in this order:

Photometric CC
SCNR (type=0, amount=1.00, preserve=true)
Background extraction (Correction: Subtraction)
Median Filter (filter=5x5 px)
Asinh Transformation: (stretch=   8.0, bp=0.00000)
Asinh Transformation: (stretch=   7.0, bp=0.00000)
Asinh Transformation: (stretch=   6.0, bp=0.00000)
Asinh Transformation: (stretch=   5.0, bp=0.00000)
Asinh Transformation: (stretch=   4.0, bp=0.00000)
Asinh Transformation: (stretch=   3.0, bp=0.00000)

The SCNR is the "Remove green"-filter. The new background extraction in Siril has become really good, with preview and no loss in signal. Rest is done in Gimp. Play around in Siril with the same fits, and save different versions from this as tif's.  See what you can make the most of in Gimp. Every image has it's own quirks. Don't drift to much between different applications. They all do the same, it's just a matter of learning to master each one. There is nothing in Pixinsight or Photoshop that can't be done in Siril and Gimp. Just take your time, and DONT DELETE old datasets. In a year or two you will brush the dust of them, and present new images as you learn.

OSC_Preprocessing_no_stack.ssf 1.3 kB · 2 downloads

Hey

Thanks for your advice and feedback, that's exactly what will help me improve.

It's a good experience actuallly doing some of the stacking myself, I've learnt a few new things about the process.

I went through the process you mentioned step by step and came up with the attached after some minor editing in GIMP. What do you think? The colours appear to be a bit better balanced, not just full on deep red. A bit more of a spectrum of colours visible and some extra detail. Of course I'm only expectations minor improvements here, but as long as I'm heading in the right direction.

 

Thanks.

Edited December 8, 2022 by AstroLearnerWill

[Rallemikken]

6 minutes ago, AstroLearnerWill said:

What do you think?

Big improvment, much more details, but red still to dominant. Have shot a lot with astromodified Canons myself, and the mod in itself does not surpress blue/green. Something other must have caused the obvious lack of signal in these two channels. What filter did you use? Don't use filters myself, so I don't know much about them. If you still have the fits you started with, try a manual color correction instead of a photometric or vica versa. Pay attention to the RGB preview in Autostretch mode and zoom in on the background somewhere in a corner. With a neutral colorbalance the background should also be somewhat neutral. Regardless of what type color correction I use, most often I also choose to run the green filter. There are ways to balance colors in Gimp, but if the signal is lost in one or two channels you don't have much to work with.

[The Lazy Astronomer]

The L-enhance is a narrowband filter, with a bandpass at Ha and another at Oiii & Hb. Narrowband images are false colour (although the HOO combination afforded by these dual/triband filters gives a relatively close approximation of true colour, with Ha mapped to red and Oiii mapped to green and blue), but the fact that most of the light spectrum is not recorded means photometric colour calibration won't give the optimum result.

The heart is quite strong in Ha emission and Oiii is generally much weaker for all objects, so I would say a fairly strong red bias is to be expected.

[Fegato]

The colour shot looks very noisy. I suspect the best thing you can do to improve this is to get more data. I'm guessing you're around F/6.5 with your setup? So particularly with the filter you want many hours of data to get that noise down. You could probably increase your exposure quite a bit if your mount / guiding can cope, and certainly lots more subs.

The Ha is very strong with the Heart, and the OIII needs teasing out in the best of conditions. It's probably easier with a false hubble like SHO type palette (although personally I don't like this much). However, with the moon strong, you're unlikely to get much of it in your data anyway, as that part of the spectrum is more prone to interference by moonlight and LP. And the L-enhance has a very wide band in the OIII / Hb (24nm) so it will pick up loads of the extraneous light. I have to say that although I also use a dual narrowband filter (IDAS NBZ) I tend to only go after Ha when the moon is bright (although I am spoiled with dark skies, so can afford to be picky).

[Ratlet]

Tidy work mate.

That's a really nice improvement imho.  You are definitely moving in the right direction.  Can't really add much to what everyone else has said.  I suspect you've got a strong foundation in terms of equipment and technique (and attitude) and you're just missing more time on the targets.

Edited December 9, 2022 by Ratlet

[AstroLearnerWill]

2 hours ago, Fegato said:

I'm guessing you're around F/6.5 with your setup?

That's right. I do get fairly good guiding with my setup so I can afford to increase the exposure time (clear sky time permitting). I'll give that a go and get some more detail. I landed on 164s subs after using Nina's optimal exposure calculator, plugging in data from a sensor analysis with Sharpcam.

 

13 hours ago, Rallemikken said:

What filter did you use?

L-enhance. First time using this filter. I'm in bortle class 5 skies and on the day had a 100% full-moon very close to where I was pointing the telescope (brilliant timing with the clear skies). I used the l-enhance because I wanted to isolate the Ha which as others have mentioned is strong in this nebula. So perhaps I can expect a stronger red detail due to that choice. On another night with no full-moon and darker skies, I'd like to try without a filter and see the difference with this nebula.


Thanks everyone for your feedback, I am amazed and grateful for the constructive critical feedback. Really is the only way one can improve.

I've got some new post-processing techniques to try and also probably the biggest factor is to get more data as mentioned by some of you above. Ideally on a night without a full-moon close-by. Due to being impatient and lack of clear-skies I've tried to get everything done in one evening, I should consider multiple evenings or leaving the scope out capturing images overnight.
 

[Rallemikken]

33 minutes ago, AstroLearnerWill said:

I used the l-enhance because I wanted to isolate the Ha which as others have mentioned is strong in this nebula.

If I'm not mistaken, that filter surpress other wavelenghts than Ha, which explaines the lack of blue and green signal. If you want to use that filter on a DSLR, you must run two sessions, one without the filter, and one with. Then you blend them, many ways to do that, even stack them together. But with a modded DSLR there is seldom lack of red.

[Fegato]

29 minutes ago, Rallemikken said:

If I'm not mistaken, that filter surpress other wavelenghts than Ha, which explaines the lack of blue and green signal. If you want to use that filter on a DSLR, you must run two sessions, one without the filter, and one with. Then you blend them, many ways to do that, even stack them together. But with a modded DSLR there is seldom lack of red.

No - it's dual narrowband. It passes Ha in a 10nm band, and then a 24nm band that covers both Hb and OIII. You can separate these out by separating the channels in processing, and then mix them as you want. But as I said above - hard to get the OIII to come out in very bright conditions.

[Fegato]

1 hour ago, AstroLearnerWill said:

That's right. I do get fairly good guiding with my setup so I can afford to increase the exposure time (clear sky time permitting). I'll give that a go and get some more detail. I landed on 164s subs after using Nina's optimal exposure calculator, plugging in data from a sensor analysis with Sharpcam.

 

Well I guess 164s might be fine, depending on conditions. In strong moonlight it only makes sense to expose for so long unless you've got very narrow filters.