Processing Horsehead Nebula

I agree with Rainer, it’s hard to say much without knowing what hardware was involved. It’s important to know which scope, camera and filters were used.

Rainer Ehlert · Jan 19, 2026 at 05:37 PM

Hi,

I gave it a simple try. I am no expert.

Is this an OSC image or a monochrome RGB stack?

Please, you need to specify your equipment.

For me it looks quite normal if it is an OSC image. IMHO

HorseHead_Astrobin.jpg

I’m using an Apertura 60mm FPL-53 on a Star Adventurer 2i with a Canon 1100D. No filters were used.

I used a Pixinsight script “Star Reduction”

Ramzi Hemadeh · Jan 19, 2026 at 11:36 PM

Rainer Ehlert · Jan 19, 2026 at 05:37 PM

Hi,

I gave it a simple try. I am no expert.

Is this an OSC image or a monochrome RGB stack?

Please, you need to specify your equipment.

For me it looks quite normal if it is an OSC image. IMHO

HorseHead_Astrobin.jpg

I’m using an Apertura 60mm FPL-53 on a Star Adventurer 2i with a Canon 1100D. No filters were used.

Are you using a field flattener? That would help with those star shapes around the border and edges of your data.

📷 Horsehead.jpgI’m no processing expert either, but this is my go with your data. I think you have a reasonable signal in there. I ended up cropping the image because there was what I can only think was a gradient on the left side that I couldn’t get rid of. There were also some vertical line artefacts under some of the stars that I removed in Photoshop. This was done with my usual Pixinsight workflow and some final touch ups in PS. I am happy to share in more detail what I do if it’s of interest.

Paul

Ramzi Hemadeh · Jan 19, 2026, 04:24 PM

I’m having trouble processing this photo I got of the Horsehead Nebula in Siril. I got around 5 hours of exposure time using 30 sec subs but can barely pull out any red nebula near the Horsehead. I’m confused if I just need to get better at processing or if I need more exposure time.

This is the photo if anyone wants to give it a go:

https://drive.google.com/file/d/1llaVzSe9Z4V5TXYpwNB65nu24vGiw5TB/view?usp=sharing

In Siril, instead of Photometric color calibration, using SPCC, average spiral galaxy, Canon D350 and Nikon LPF filter (Some Low Pass Filter) bring out the Ha colors and details.

Hope this helps📷 VeraLux_StarComposer_result.jpg

Paul Sanfilippo · Jan 20, 2026, 05:26 AM

📷 Horsehead.jpgI’m no processing expert either, but this is my go with your data. I think you have a reasonable signal in there. I ended up cropping the image because there was what I can only think was a gradient on the left side that I couldn’t get rid of. There were also some vertical line artefacts under some of the stars that I removed in Photoshop. This was done with my usual Pixinsight workflow and some final touch ups in PS. I am happy to share in more detail what I do if it’s of interest.

Paul

Paul, this is fantastic, care to share some of the workflow? I’m currently learning pixinsight and I can’t seem to match this quality just yet

jack_miller · Jan 21, 2026 at 02:36 PM

Paul Sanfilippo · Jan 20, 2026, 05:26 AM

📷 Horsehead.jpgI’m no processing expert either, but this is my go with your data. I think you have a reasonable signal in there. I ended up cropping the image because there was what I can only think was a gradient on the left side that I couldn’t get rid of. There were also some vertical line artefacts under some of the stars that I removed in Photoshop. This was done with my usual Pixinsight workflow and some final touch ups in PS. I am happy to share in more detail what I do if it’s of interest.

Paul

Paul, this is fantastic, care to share some of the workflow? I’m currently learning pixinsight and I can’t seem to match this quality just yet

Hi Jack,

I’ve only been at this for 12 months so I’m still relatively new to it all, but I do have a PI/Photoshop workflow that seems to work for me now. I started out using Siril and it was ok, and I was somewhat reluctant to look at PI because there was an additional cost, another learning curve, etc. Then I did a trial of PI and (for me) the ability to produce a good image is night and day with Siril vs PI - I just find the tools so much nicer to use. Anyway, I digress…

My basic workflow for a standard OSC image is:

1. Have a quick look at the integrated image with the screen transfer function (unlinked) to get an idea of what I’ve got to work with.

2. Run Graxpert from within PI using the default settings and latest AI model (I don’t set background points - just let it run fully automatically).

3. I may or may not run SPCC at this point (mostly I do).

4. Then Blur XTerminator - I know a lot of people ‘Correct Only’ initially, but I untick this and just do the full sharpening in one go at this point.

5. Then, depending on the image and whether I want to separate the background from the stars or not, I will run Star XTerminator. Mostly I separate.

6. If separating, I will then stretch the background image using GHS - big stretches to start with and then smaller stretches after. Setting the symmetry point each time - usually to the peak - and protecting highlights with each stretch. I also turn on the log visual and aim - for the most part - for the pixel brightness to gradually decline in a linear manner from the peak at the left of the histogram to the right of the histogram.

7. Then I’ll do a lightish Noise XTerminator (e.g. denoise = 0.7, iterations = 1)

8. At this point I’ll consider applying a HDR Multiscale Transform if I want a little more sharpening.

9. That’s the background image done in PI. Now I’ll stretch the star image using SetiAstro’s Star Stretch script. I usually leave the Stretch Amount between 4 - 5, depending on how much I want to emphasise the stars in the final image; but I’ll maximise the Colour Boost up to 2.

10. After this I export both images as 16 bit tiffs and open them up in Photoshop.

11. I then run Camera Raw Filter (as a smart object in the layer so that I can go back and edit more if wanted) on the background image, adjusting exposure, contrast, saturation, vibrance etc to my liking. I may, at this stage run another denoise using Noise XTerminator from within photoshop.

12. I then copy and paste the stars image as a layer onto the background layer and Screen the star layer over the background layer.

13. I think that’s pretty much it.

I think a lot of the processing is about finding what works for you. There’s so much information about there and people will suggest all sorts of rules you should follow (and perhaps some are important), but just experiment. That’s what I did. Mind you, I am still learning stuff every day and no doubt some of what I’ve described above will change. Having fun (amidst the frustration) though is what makes it worthwhile for me.

Paul

Ramzi Hemadeh · Jan 22, 2026 at 03:26 AM

Paul Sanfilippo · Jan 21, 2026 at 09:53 PM

jack_miller · Jan 21, 2026 at 02:36 PM

Paul Sanfilippo · Jan 20, 2026, 05:26 AM

📷 Horsehead.jpgI’m no processing expert either, but this is my go with your data. I think you have a reasonable signal in there. I ended up cropping the image because there was what I can only think was a gradient on the left side that I couldn’t get rid of. There were also some vertical line artefacts under some of the stars that I removed in Photoshop. This was done with my usual Pixinsight workflow and some final touch ups in PS. I am happy to share in more detail what I do if it’s of interest.

Paul

Paul, this is fantastic, care to share some of the workflow? I’m currently learning pixinsight and I can’t seem to match this quality just yet

Hi Jack,

I’ve only been at this for 12 months so I’m still relatively new to it all, but I do have a PI/Photoshop workflow that seems to work for me now. I started out using Siril and it was ok, and I was somewhat reluctant to look at PI because there was an additional cost, another learning curve, etc. Then I did a trial of PI and (for me) the ability to produce a good image is night and day with Siril vs PI - I just find the tools so much nicer to use. Anyway, I digress…

My basic workflow for a standard OSC image is:

1. Have a quick look at the integrated image with the screen transfer function (unlinked) to get an idea of what I’ve got to work with.

2. Run Graxpert from within PI using the default settings and latest AI model (I don’t set background points - just let it run fully automatically).

3. I may or may not run SPCC at this point (mostly I do).

4. Then Blur XTerminator - I know a lot of people ‘Correct Only’ initially, but I untick this and just do the full sharpening in one go at this point.

5. Then, depending on the image and whether I want to separate the background from the stars or not, I will run Star XTerminator. Mostly I separate.

6. If separating, I will then stretch the background image using GHS - big stretches to start with and then smaller stretches after. Setting the symmetry point each time - usually to the peak - and protecting highlights with each stretch. I also turn on the log visual and aim - for the most part - for the pixel brightness to gradually decline in a linear manner from the peak at the left of the histogram to the right of the histogram.

7. Then I’ll do a lightish Noise XTerminator (e.g. denoise = 0.7, iterations = 1)

8. At this point I’ll consider applying a HDR Multiscale Transform if I want a little more sharpening.

9. That’s the background image done in PI. Now I’ll stretch the star image using SetiAstro’s Star Stretch script. I usually leave the Stretch Amount between 4 - 5, depending on how much I want to emphasise the stars in the final image; but I’ll maximise the Colour Boost up to 2.

10. After this I export both images as 16 bit tiffs and open them up in Photoshop.

11. I then run Camera Raw Filter (as a smart object in the layer so that I can go back and edit more if wanted) on the background image, adjusting exposure, contrast, saturation, vibrance etc to my liking. I may, at this stage run another denoise using Noise XTerminator from within photoshop.

12. I then copy and paste the stars image as a layer onto the background layer and Screen the star layer over the background layer.

13. I think that’s pretty much it.

I think a lot of the processing is about finding what works for you. There’s so much information about there and people will suggest all sorts of rules you should follow (and perhaps some are important), but just experiment. That’s what I did. Mind you, I am still learning stuff every day and no doubt some of what I’ve described above will change. Having fun (amidst the frustration) though is what makes it worthwhile for me.

Paul

Wow thank you so much I will definitely reconsider using Pixinsight. I have one question. Would you say the cost is worth it and also how long did it take you to learn how to use PI to see a night and day difference between that software and Siril.

For me the cost was worth it (but then I’m a 51 year old with a decent income to be able to afford it). If cost is an issue then Siril is definitely good software. I just found PI so much easier to get good results with to be honest, and I think I was able to do that after only a few weeks of playing around with it . PI is extremely extensible with ‘processes’ and scripts - allowing you to do a variety of things with your images. I’m not sure how extensible Siril is. I would give the trial a go and then you can evaluate whether it’s worth it from there. I knew within a few days of using the trial that I would like to continue using the software. One thing I will say that Siril has the advantage for is stacking - it seems to be much faster than PI’s WBPP. Having said that I still use PI to keep the workflow in the one software and I also like the ‘Master’ files (flats, biases, darks) it generates which I then save (instead of the individual fits) with my lights (fits) for future tinkering if I want.