Back to forums

[RCC] Sh2-136 Ghost Nebula LRGB

Pleiades Astrophoto PixInsight Processing LRGB Request for feedback
18 replies•987 views
patrice_so avatar
patrice_so avatar
Dear fellow astrophotographers,

While searching for a broadband target at this time of year, I ended up capturing the Ghost Nebula. I managed to capture 24 hours and 30 minutes of LRGB monochrome data with which I am satisfied. I stacked and processed the data. The outcome is as follows:

https://app.astrobin.com/?i=zm4asw 

I am usually happy but not fully satisfied with my results, as I usually see possible improvements that I am unable to implement (either because the data does not allow it or because I am not good enough at processing). However, this time I am pretty happy with the result, and I cannot identify any significant improvements that could be made in the processing. I feel that the limitation lies in my own critical and aesthetic judgement, rather than in the data or the processing. As we progress in astrophotography, we develop techniques and skills. We also need to improve our aesthetic sensitivity and judgement. To develop my sensitivity and judgement further, I would love it if you could point out possible improvements that I am not even aware of yet.

The entire processing was done in Pixinsight:
- WBPP for stacking
- Graxpert on L, R, G and B autofocused layers
- RGB channel combination
- SPCC
- multiscale gradient correction
- BlurX (correct only)
- BlurX
- StarX to process the stars separately: Seti star stretch, colour saturation, curve adjustment, NoiseX and unsharp mask
- GHS is used to stretch the RGB, with NoiseX applied once in the middle of the stretching iteration to improve the black point calibration
- Curves and selective colour correction are used to achieve a vivid RGB basis that I like
- BlurX (correct), BlurX and StarX, as well as GHS/NoiseX/GHS on the L layer, are used to create a luminance layer that I like
- LRGB combination
- Curve editing with a lightness mask to work on the bright and dark areas separately.
- Star combination.

Any comments or criticisms, especially regarding the general aesthetics, would be much appreciated.

CS

Patrice
Well Written Helpful Insightful Respectful Engaging
andrea tasselli avatar
andrea tasselli avatar
Stars lack significant color (which means the luminance overwhelms the color contribution)  and, while I approve of the general aesthetics and processing, I am wondering whether the subject really looks like this (all this whiteness). As I haven't tackled this one yet it is just musing at this time yet one is left wondering…
Daniel Cimbora avatar
Daniel Cimbora avatar
I am far from an expert, but your image looks really good to me.  Lots of detail in the nebulae but not overprocessed.  Stars look great, I don't perceive a lack of color that Andrea mentions..  Background pleasingly dark but not too dark.  Nice composition.

Your aesthetic matches mine, and I'd be very pleased with this outcome.
Well Written Respectful Supportive
Francesco Meschia avatar
Francesco Meschia avatar
I imaged this same region last year with almost exactly the same framing. We’re talking about subjective aesthetic taste here, but for me it’s stretched a bit too much. As Andrea noted, this may have the side effect of washing out the colors (which aren’t that strong to begin with). I had elected to go for a more subtle rendition of the dust, I’m thinking it may work even better for you, since your data appears to be quite good.
Helpful Insightful Respectful
patrice_so avatar
patrice_so avatar
Dear all,

Many thanks for your positive comments and very valuable inputs. That is much appreciated. 

First, I have to admit @andrea tasselli point about stars colors. These are stars taken from the RGB layer, without the L layer having been applied. However, I believe that my image suffers from using 300s exposures. I should wait for the next opportunity to shoot one hour of R, G and B 30s subs to see if this goes better. In the version below, I tried to use a very light arcsin stretch to get a better starter at the beginning of the process. But as you can see, the pair of blue stars next to the Ghost has some orange. I tried various strategy, but I alway end up with worst than that. 

Second, regarding Andrea's and @Francesco Meschia point about the object colors, I also admit that the with style may be a bit exagerated, even if I tend to like it. The area is quite white in itself. It is by far less saturated than the region next door of NGC7023. But indeed, many other images of the Ghost Nebula exhibit more colors. I tried to improve that with a new stretch, that I combined with the stars mentionned above. But this is very difficult. As the data are quite whitish, I get gradient appearing whenever I increase saturation, calling for the kind of selective correction I don't like to do. Besides, I also  believe that the overstretch pointed out by Francesco is even more problematic. 


I am quite convinced about the increasing saturation of the lighter areas, espectially on the main object. This is great progress. I will try to prepare a third version with a light saturation increase using the first version to see if this gets betters. For the stars, I need the sky to get clean. 

Many thanks. Further comments always welcome. 

CS

Patrice
Francesco Meschia avatar
Francesco Meschia avatar
But as you can see, the pair of blue stars next to the Ghost has some orange. I tried various strategy, but I alway end up with worst than that.

I may have an idea for this. Can you run the “FWHMEccentricity” script on the three R, G, and B integrated masters before any processing is applied, and post the values for FWHM?
Well Written
patrice_so avatar
patrice_so avatar
Thanks @Francesco Meschia .

Here they are : 





I suspect what you are aiming at, but I am unsure how to interpret these measures. Could you enlighten me?
Darryl Ackerman avatar
Darryl Ackerman avatar
I think it looks great. When I combine LRGB lately I do it without stars, then add the RGB stars and do not use the luminance stars at all. My skills are not so great, but it seems to be giving me better stars overall.
patrice_so avatar
patrice_so avatar
Darryl Ackerman:
I think it looks great. When I combine LRGB lately I do it without stars, then add the RGB stars and do not use the luminance stars at all. My skills are not so great, but it seems to be giving me better stars overall.

Thank you very much @Darryl Ackerman for your kind words. I did the same : I remove the stars from the RGB to keep them aside for reintroduction at the final stage. The stretched RGB and L layers were starless layers when applying LRGB combination. I totally agree with you.
Respectful Supportive
Francesco Meschia avatar
Francesco Meschia avatar
patrice_so:
Thanks @Francesco Meschia .

Here they are :

[cut]

I suspect what you are aiming at, but I am unsure how to interpret these measures. Could you enlighten me?

Thanks! I just wanted to see if the R master had much “fatter” stars than the B one. That doesn’t seem to be the case here, which makes sense since you’re imaging with a reflector.
The eccentricity seems a bit high for masters, but I don’t know whether the measurement is affected by the diffraction spikes of a reflector.
Well Written Insightful Respectful
Drew Hackney avatar
Drew Hackney avatar
I really like your rendition on this object. With regards to the orange fringing on your stars, there are two things I do with stars on my LRGB images:

1. I run BlurXTerminator as correct only on each individual channel to correct any issues with my stars before combining the channels. Then I register the images to my luminance channel, and crop out the edges as necessary. Then I combine the RGB channels using the newly registered R, G, and B images.
2. Once I separate my stars, I invert the image and use SCNR to remove green noise. This corrects for any magenta present. Orange isn't magenta, but it's close so maybe it'll help?

Mind you, I consider myself okay-to-good at the hobby so this might be bad advice. I also shoot with refractors so I don't know re-registering your RGB channels will  screw up your diffraction spikes.
Helpful Respectful Engaging Supportive
patrice_so avatar
patrice_so avatar
Thanks @Drew Hackney  for your input and your kind words. Glad you like it.

I stopped running BlurX on R, G and B channels seperately because at some point I read that running BlurX on RGB image would prevent BlurX from making a mismatched correction on different channel as it may do in when the layers are processed separately. I will check the literature again to be sure.

I considered applying SCNR to the star layer and the inverted layer. However, I like the rainbow effect of the diffraction spikes. I know that my spikes are not that big and rainbowy. This is due to mechanical imperfections in the scope, but also to the fact that my scope has extra thin spider vanes, resulting in modest spikes. I would like to keep what is there of the rainbow effect, so I would prefer to avoid removing the green and orange altogether. But if there is no other choice, I might consider it. I will see.

Thanks and CS,

Patrice
Helpful Insightful Respectful Engaging
Francesco Meschia avatar
Francesco Meschia avatar
I stopped running BlurX on R, G and B channels seperately because at some point I read that running BlurX on RGB image would prevent BlurX from making a mismatched correction on different channel as it may do in when the layers are processed separately. I will check the literature again to be sure.

Yes that's correct – and I'm not advocating running BXT only on the individual channels. My idea was to give BXT a better fighting chance by first using it to match all channels to the best one, then again on the RGB combined image.
Well Written Insightful Respectful Concise
sydney avatar
sydney avatar
I think the original looks great.  I wouldn’t change anything unless there are things that bug you.  I also love the detail and color in the little background galaxies.
Well Written Respectful Supportive
patrice_so avatar
patrice_so avatar
Thanks @sydney for your kind words. I completely agree. I was not able to take the time to give it another round. At this point, I believe that I will try to increase very slightly the saturation of the original processing and look for gethering 1 hours of short R, G and B exposure to improve the stars.
Tim Ray avatar
Tim Ray avatar
Lately I have been combining my Luminence image with my RGB image with the ImageBlend script. Gives more control than the LRGB combine process.  

CS
Anderl avatar
Anderl avatar
Darryl Ackerman:
I think it looks great. When I combine LRGB lately I do it without stars, then add the RGB stars and do not use the luminance stars at all. My skills are not so great, but it seems to be giving me better stars overall.

Only my opinion but i think that deep exposed luminance stars add depth to a picture that is hard to get with rgb stars only.
patrice_so avatar
patrice_so avatar
Dear fellows astronomers, 

Many thanks for your comments. I decided for a very slight improvement in color saturation. I also added a very prudent additional step of color denoising as I noted that the initial version (i.e. the V2). The new version is cleaner without manifesting any symptom of overdenoise. I considered adding constrast in the mid-tones to highlight details in the dusts, but finally renounced as this was creating artifacts that I don't want on this image. 

The image is here for pixel peeping : https://app.astrobin.com/u/patrice_so?i=zm4asw#gallery



The additional denoise benefit is clearly visible here, just as the fact that I tried not to make it excessive: 


Many thanks for your inputs. Obviously, any additional inputs are most welcome. 

CS

Patrice
Helpful Respectful Engaging Supportive
patrice_so avatar
patrice_so avatar
Dear fellow astronomers, 

I cannot help it, I had to try to reprocess the whole thing once more. This gave me V5.

V5 is more vivid. Colors are more saturated and dark lights are more contrasted. The fraiming is also slighly angled, so that the spikes are not vertical and horizontal, which I like. 

V4 is however more soft, which I like more when it comes to the dark lights. 




V5 is not online yet, as I am unconvinced by the background. 

I would be glad to get a last round of comments. I think that I am progressing a lot with this exercise. You help is greatly appreciated. 

Thank you very much and clear skies to all of you, 

Patrice
Respectful Supportive