Nightmare in Andromeda

[Alexgc]

Hi, any tips on processing andromeda? Watched the youtube videos, read the books but the result looks a bit over-processed to me. 3hrs of clean looking 150sec subs at iso 1000. Darks flats and bias frames.

In photoshop. Used a layer mask to try to bring out some detail in the middle and then a heavily feathered selection for the galaxy to sharpen in camera raw filter, and get rid of all the vignetting. How would you all have gone about it...?

 

Edited September 25, 2022 by Alexgc

[Elp]

It's not a bad attempt. Processing is highly subjective and you need to know when to exercise restraint by doing very subtle changes at times. You may have gone a bit overboard with the contrast/levels, you'll be able to tell if you look at a histogram, the white/highlights (RHS) and black/shadows (LHS) should not be clipped off (ie the histogram graph of the image should steadily rise and fall from zero and back again and not meet a sharp cutoff on the left or right side of the histogram).

Your flats should have removed the vignetting.

If you don't use Siril, do. It's worked wonders for me and it's free, even if you simply do a pre process of your linear stack and do a background extraction to normalise the background.

[raadoo]

Uploading the unprocessed, unstretched, stack would make it easier (for me, at least) to work out a battle plan.

As much as there are general directions for galaxies and nebulae and planets, I find that the acquisition setup and its result often influence processing strategy more than going through a formula.

Looking at your process above, I can see some star bloat (overstretched, maybe?), overly noisy output in faint areas, crushed shadows and some unnatural colour cast in Andromeda itself.

[900SL]

What did you stack with? Do you have the original stacked file before any processing?

[Alexgc]

Here's the stacked unstretched or otherwise processed tif file. I'm new to autoguiding, this is only the 3rd time I've done it, but thought that the subs I was getting looked pretty good, certainly compared to the unguided efforts previously!

I'd be very interested in your thoughts and observations

22-9-24 Andromeda.tif

[Stu Wilson]

Best i can get for a quick go.

 

Stu

[Simon Pepper]

Thanks for sharing your data here is what I came up with. I think I prefer the starless version. Also whenever I see an image of Andromeda this way up I always think its up side down lol! I need to remember there is not up, down, left or right in space!

You have good data there I would be inclined to shorten your sub lengths on M31 due to blowing out the core and as I keep hearing, in post sometimes less is more. Less stretching and less of the green (probably colour balance), but as I said you got some good stuff stick at it

Thanks 

 

Edited September 26, 2022 by Simon Pepper

[alacant]

23 hours ago, Alexgc said:

How would you all have gone about it

Hi

You've produced some good quality data with which to try differrnt app(roache)s. Here's one keeping the wonderful star-field you captured intact.

 

 

[david_taurus83]

Some nice data there, better than any M31 I've tried. What equipment did you use? The stars look ever so slightly out of focus. Just though.

[vlaiv]

Yep, it is a bit out of focus.

Excellent data by the way - it goes deep.

[vlaiv]

Yep, just two small issues - a bit out of focus and core of M31 slightly clipped, otherwise, excellent data.

[Alexgc]

Holy cow! It is quite humbling to see what you've turned my data into. Some really beautiful images there. Nothing short of inspirational. Thank you.

Many comments I've read about the learning curve in this game, and this certainly applies to processing as much as image acquisition. What would your advice be for routes to improve photoshop skills? How do you remove the stars for starters...?

[Alexgc]

25 minutes ago, david_taurus83 said:

Some nice data there, better than any M31 I've tried. What equipment did you use? The stars look ever so slightly out of focus. Just though.

Thanks. I just started using my new EQ-R 6 Pro and a WO Fluorostar 95. PHD2. (New in terms of number of clear nights since purchase that is...)

[vlaiv]

31 minutes ago, Alexgc said:

How do you remove the stars for starters...?

https://www.starnetastro.com/download/

[raadoo]

@Alexgc here's my take on your M31.

I often find that it's easier to learn by deconstructing what others have done, so I've also attached the PSD file. It ended up quite large at 829MB, but I didn't want to cheat you out of pixel peeping at the full 24MP 😁

I tried to keep as much of the editing using only Photoshop's tools, so at least these will be tools you're familiar with.

The one thing that I used that's not native to Photoshop is StarXTerminator (similar to Starnet++) to break apart the stars from the galaxy. SXT comes as a Photoshop plugin for added convenience.

 

22-9-24 Andromeda.psd

[Alexgc]

That's spectacular, thanks! Day off tomorrow so I'll spend the time working through it, and look up the star removal tools as well. Apart from aesthetics, presumably if you remove the star field you can alter the DSO without the stars bloating etc

[Simon Pepper]

6 hours ago, raadoo said:

@Alexgc here's my take on your M31.

I often find that it's easier to learn by deconstructing what others have done, so I've also attached the PSD file. It ended up quite large at 829MB, but I didn't want to cheat you out of pixel peeping at the full 24MP 😁

I tried to keep as much of the editing using only Photoshop's tools, so at least these will be tools you're familiar with.

The one thing that I used that's not native to Photoshop is StarXTerminator (similar to Starnet++) to break apart the stars from the galaxy. SXT comes as a Photoshop plugin for added convenience.

 

22-9-24 Andromeda.psd 828.96 MB · 2 downloads

Oh my… Just awesome! Goes to show we all have our own take. This one is my favourite so far 👌

[ollypenrice]

Most of us have learned that our first and most fatal error as beginners lies in black clipping. We bring in the black point (the left hand slider in Levels moved too far to the right) and this clips out the fainter signal. The temptation to do this often arises from the desire to get rid of gradients, but the black point slider is not the way to do this. This screen grab shows your current histogram in Levels (Photoshop.) It is massively black clipped, meaning the left had side of the histogram peak is jammed up against the left hand edge. Your faint data are now lost and cannot be recovered. There should always be a small amount of flat black line on the left before the peak begins to rise. The way to check this is to look at the Levels histogram after every operation in stretching. You can also use the eyedropper in Ps to place Colour Sampler marks on the background. Keep it high, around 23 to 25, during processing. You can always bring it in as a very last operation but you need the elbow room above the black point for many operations. 21 to 23 is a decent background value because the night sky is not jet black.

The next thing would be colour balance. Green is too strong, as often happens. Pixinsight has SCNR green but an alternative, on Rogelio Bernal Andreo's website, is free Hasta La Vista Green. These routines are vital. They pull out unwanted green and green noise.

Do not shoot short subs for the core if you don't need them. Open the linear image in Photoshop, open Curves, and click the cursor on the core. If the point appearing on the curve is not right at the top you are not saturated. The need for short subs is greatly overestimated n my view, though I don't know here.

I downloaded your data but can't open it, unfotunately. ('Incompatible File Format' for some reason.) However, here's the histogram on my own M31. Note the distance between the histo peak and the left hand edge of the graph.

Olly

 

 

[Laurieast]

I usually do Lunar and Planetary, but could not resist a go with this.
Photoshop levels and curves, and at every step look at your levels as Ollypenrice said above, repeat several times, and up the saturation. Applied the HLVG filter to reduce the green. I did also do a reduce stars action with Astronomy Tools action set. 

 

 

Edited September 30, 2022 by Laurieast

[Alexgc]

"My god it's full of stars"

Brilliant. I'm so impressed with all the above results and am inspired to keep at it. Thank you all of the above, especially for the advice. I shall have a go tomorrow and if I'm feeling brave enough I'll post the output!

[Alexgc]

In the meantime, here's one I took earlier. My 1st ever DSO, 3rd April last year. Burnt out my wife's hairdryer keeping the lens dew free... 

[DaveS]

My (Second) version after working in AstroArt 8

Split RGB

Individual gradient removal

Recombine in Trichromy so AA could rebalance the chanels

DDP of the combined RGB with saturation boost and high pass filtering (May have left the latter a bit too high)

Richardson-Lucy deconvolution for the slightly out of focus stars.

Slight hue and saturation adjustment

Unsharp mask with dark halo control.

I'm sure there's more in the data

 

[Marvin Jenkins]

What a fantastic testament to SGL membership this thread displays. 
So many people helping a beginner asking questions, long may it continue.

M

[Alexgc]

Ok, I think it's an improvement. No clipping at least. StarXterminator is quite incredible but need to play around with it a bit more. Not got my head fully around layer masks as yet. Thanks for all the above advice 🙌

 

22-9-24 Andromeda 5th 8bit.tif

Edited September 29, 2022 by Alexgc

[tomato]

Apologies, I'm late to this thread. Here is my effort processed through PI, StarTools and Affinity Photo. I separated the stars with StarXterminator  and used StarTools to try and sharpen them a little as the focus was a little soft. The background is super smooth however, the starless image looked like it had been pasted onto a brand new blackboard. Are you imaging from a dark site?