Tips for processing workflow in SiriL and Photoshop

[INeedSomeHelp]

2 weeks ago I created a post asking how to improve my Andromeda image. Turns out I made a classic beginner mistake and Andromeda was just out of frame. I have, however, repeated this experiment, and have quite some light frames of M31 now! I could use some help with the post processing, I can manage myself but I got such good tips in my last post that I figured I might as well ask for advice once more. Attached to this post are the output of a stack performed in SiriL AND 2 .tif files. I already performed some post processing steps in SiriL, which resulted in the 2 .tif's. In Siril I: 

1) Applied the automatic Color Calibration (for M31 ofcourse).

2) Corrected the green noise.

3) Performed background extraction on 1 image (and one image without). The background extraction (as I understand it) is to get rid of the gradient in background color. In my non-background corrected image you can see that the background at the bottom is quite orange (light pollution from a road I think). The background extraction fixes this quite nicely, BUT it also causes some very heavy graining to occur. 

4) Lastly I applied the stretching algorithm that is build in in SiriL under "Histogram Transformation" and saved the .tifs as a 16 bit image.

5) Exported both a .tif without and one with background extraction (2 images are attached). I also attached the image that you obtain directly after the stacking (called result.fit).

 

So what happened with the background extraction causing this graining? Should I skip it, or perform it differently? Furthermore: how to proceed with a .tif in photoshop.

I think, with good editing and processing, there might be a good image there!

Cheers.

 

stretchedwithbackgroundcorrection.tif stretchedwithoutbackgroundcorrection.tif.tif result.fit

[Budgie1]

To get you going in Photoshop, have a look through this tutorial on Basic Photoshop Image Correction. It's based on CS2 but is still relevant and easy to follow.

[AstroMuni]

1 hour ago, INeedSomeHelp said:

I already performed some post processing steps in SiriL

Just incase you havent read it already, this tutorial helps a lot in suggesting a sequence of steps https://siril.org/tutorials/tuto-scripts/#tuto-3

If the background light pollution is excessive, you may benefit from applying background extraction using a 1st degree polynomial function to all the images before stacking. Siril has a function to do that and its mentioned in one of the tutorials. https://siril.org/tutorials/tuto-manual/#background-extraction

Edited August 3, 2021 by AstroMuni

[INeedSomeHelp]

23 hours ago, Budgie1 said:

To get you going in Photoshop, have a look through this tutorial on Basic Photoshop Image Correction. It's based on CS2 but is still relevant and easy to follow.

Nice, I will have a look at this!

[INeedSomeHelp]

23 hours ago, AstroMuni said:

Just incase you havent read it already, this tutorial helps a lot in suggesting a sequence of steps https://siril.org/tutorials/tuto-scripts/#tuto-3

If the background light pollution is excessive, you may benefit from applying background extraction using a 1st degree polynomial function to all the images before stacking. Siril has a function to do that and its mentioned in one of the tutorials. https://siril.org/tutorials/tuto-manual/#background-extraction

I haven't read that, I will start with this! Do you think that the graining which occurs after background extraction is caused by excessive light pollution? BTW, do you guys think that the automatic stretching incorporated in SiriL yields good results? Or should I do it manually (I already tried in photoshop, but couldn't get a good image that way, the automated SiriL stretching performed better).

[Iem1]

As a fellow beginner, I found SiriL auto stretch a very good way point to see what the basic data is capable of, and sometimes use it as a starting point to then be picked up and finished in PS etc.

I found the methods in the above video great for doing a full 'manual' stretch in PS as I struggled too 

 

Hope this helps, take care 

[INeedSomeHelp]

On 03/08/2021 at 14:47, Budgie1 said:

To get you going in Photoshop, have a look through this tutorial on Basic Photoshop Image Correction. It's based on CS2 but is still relevant and easy to follow.

This is what I can get following your tutorial. How would you improve this image? I still have to do some star reduction.  

[happy-kat]

Did you take a look here

https://stargazerslounge.com/topic/355585-ep1-youtube-making-every-photon-count-by-steve-richards/

[AstroMuni]

22 hours ago, INeedSomeHelp said:

Do you think that the graining which occurs after background extraction is caused by excessive light pollution?

The graining that is left behind is generally noise and using darks, flats should help reduce this. Gradients left behind by light pollution are handled by the background extraction. Taking more images should help increase the signal to noise ratio too.

[INeedSomeHelp]

29 minutes ago, AstroMuni said:

The graining that is left behind is generally noise and using darks, flats should help reduce this. Gradients left behind by light pollution are handled by the background extraction. Taking more images should help increase the signal to noise ratio too.

I have 50 dark frames, 50 flat frames and 75 bias frames. This should be sufficient right? Also, there is no grain before I perform a background correction, it really is the background correction which causes it.  I also took about 650 lights at 2.5 seconds each. This should mean (if I did the math correctly) that I have a total integration time of 27 minutes. I don't know if increasing the amount of light frames will improve the image much right?

Edited August 5, 2021 by INeedSomeHelp

[Iem1]

11 minutes ago, INeedSomeHelp said:

I have 50 dark frames, 50 flat frames and 75 bias frames. This should be sufficient right? Also, there is no grain before I perform a background correction, it really is the background correction which causes it.  I also took about 650 lights at 2.5 seconds each. This should mean (if I did the math correctly) that I have a total integration time of 27 minutes. I don't know if increasing the amount of light frames will improve the image much right?

That should be more than enough calibration frames for sure, I have read 20-30 of each is more than sufficient.

I have also noticed that on occasion the background extraction alone does more harm than good, I have just thought "If it is not broke, don't try and fix it"

27 minutes is a relatively short imaging time, longer imaging time should for sure help just about everything I believe (less noise and more detail)

Here's some links to help you out, specific to the Andromeda galaxy on an untracked system. The second video covers processing, a guide I followed when processing my own, it involves removing the stars using Starnet which isolates the galaxy, preserving the background and stars during processing.

Part 1 (Deatails on capturing the data)

Part 2 (Details on processing the data)

This guy does a lot of work involving untracked images, might be worth taking a look at some of his other videos for hints and tips as he covers his own workflow in pretty easy to follow, step by step guide.

Best wishes

[AstroMuni]

35 minutes ago, INeedSomeHelp said:

I also took about 650 lights at 2.5 seconds each. This should mean (if I did the math correctly) that I have a total integration time of 27 minutes. I don't know if increasing the amount of light frames will improve the image much right?

Taking more images will definitely help. Most of the lovely images you see out there are several hours worth of imaging. AND learning what I call the dark art of post processing helps a lot!

Folk split out the object, stars and background into different layers and then enhance each one seperately - A skill in itself which I havent mastered very well. My images have grainy backgrounds too (see link in my signature). One simple way to get rid of the graininess is increase the black point, but that may get rid of the wispy bits of galaxy too.

[INeedSomeHelp]

6 hours ago, Iem1 said:

That should be more than enough calibration frames for sure, I have read 20-30 of each is more than sufficient.

I have also noticed that on occasion the background extraction alone does more harm than good, I have just thought "If it is not broke, don't try and fix it"

27 minutes is a relatively short imaging time, longer imaging time should for sure help just about everything I believe (less noise and more detail)

Here's some links to help you out, specific to the Andromeda galaxy on an untracked system. The second video covers processing, a guide I followed when processing my own, it involves removing the stars using Starnet which isolates the galaxy, preserving the background and stars during processing.

Part 1 (Deatails on capturing the data)

Part 2 (Details on processing the data)

This guy does a lot of work involving untracked images, might be worth taking a look at some of his other videos for hints and tips as he covers his own workflow in pretty easy to follow, step by step guide.

Best wishes

Yep, I found those tutorials and actually followed this one! However, I use the combination of SiriL and photoshop, as well as a 100mm macro lens. Nice that I can take less calibration shots, it feels kind of bad to burn away your shutter with a lens cap on your lens. 

Not doing the background correction means a "cleaner image" (i.e. less grain), but then I have an orange tint in the bottom of the image, which fades to deep blue at the top. 

I will maybe try some nebulae soon, such as the eagle or lagoon nebula!

[INeedSomeHelp]

I thought I was quite a big boy with my 27 minutes ;). As for the processing, this is why I made this post. I felt that there is definitely ways to improve the image post-hoc. Nice images btw.

[happy-kat]

with your very short exposures taking calibration frames takes no time. You did not take any dark flats they generally help background noise a lot.

Also you could experiment as darks don't always improve DSLR images, YMMV

[INeedSomeHelp]

This is the result. I think more skilled post-processors would have improved the final outcome but I still think it is quite cool!

[alacant]

On 09/08/2021 at 10:09, INeedSomeHelp said:

the result

Nice. It's good to experiment to see which style suits best an image.

I think you may be losing some detail though, especially the outer bits of the galaxy.

But hey, your star colours are great.

Edited August 10, 2021 by alacant