What software do you use for processing?

[Olli]

Hi everyone ,

I have finally got my imaging rig sorted (now just need the clear skies!) and have started to read into the processing side of things which I’m excited to try though I am slightly overwhelmed with the amount of software available. I just wanted to see what a beginner should use. I have seen deep sky stacker mentioned a fair few times and photo shop and also seen some videos on Sirill are there any others you would recommend as well as any books or online tutorials that would help?

 

Many thanks in advanced!

[ONIKKINEN]

Siril is really good for a beginner, or for more experienced imagers.

You can do the core parts of processing in Siril quite easily, which would be (simplified) something like:

1. Crop out incomplete edges from stacking
2. Remove light pollution/sky gradients with the background extraction tool
3. Colour calibrate using the Photometric color calibration tool
4. Stretch using Asinh transformation and/or Histogram transformation, or the hyperbolic stretch tool but that one is a bit more complicated and probably not something one should look at for the first few attempts.

At this point you could export the image in 16-bit TIFF format to further adjust in Photoshop however you like. This is more or less how i process my images, although i do a couple of things in PixInsight too.

You could stack in Siril too as it is a bit better and versatile than DeepSkyStacker, but its not nearly as straight forward as the drag and drop simplicity of DSS. So for a beginner my advice would be stacking in DSS, core processing in Siril and then final adjustments in Photoshop. Could use Gimp instead of Photoshop, but i think Photoshop is well worth the money with how many different things you can do in it.

[The Admiral]

When I started I used DSS and StarTools (https://www.startools.org/). This is not expensive and can give good results, although to be fair it has a rather unique modus operandi (or should that be operandum 🙂 ), which you may or may not take to. I now don't use it as I feel it's too much of a black box, and for my stacking and processing I'm now using Astro Pixel Processor (https://www.astropixelprocessor.com/). As a subsidiary, I'm also migrating to Siril but I've not been very active of late so my Siril experience is taking a while to gain. If you do conventional photography as well, then you might take a look at Affinity Photo, which is also capable of astro processing.

Don't forget that you can get a trial period for all this software, so it's worth trying things out to see what you find easiest (less complicated 😉).

There are lots of videos around to help you with these bits of software, but the ones I find most helpful are by Nebula Photos and Deep Space Astro.

Ian

[bomberbaz]

15 hours ago, ONIKKINEN said:

 

You could stack in Siril too as it is a bit better and versatile than DeepSkyStacker, but its not nearly as straight forward as the drag and drop simplicity of DSS. So for a beginner my advice would be stacking in DSS, core processing in Siril and then final adjustments in Photoshop. Could use Gimp instead of Photoshop, but i think Photoshop is well worth the money with how many different things you can do in it.

Don't know if you are aware but Siril now has an auto stacking script which is simplicity itself to use.

Simply put the files for stacking in folders on PC and boom, click the script and sit back and wait for it to finish. 

[bomberbaz]

Ok so found the guide for Siril for you. As @ONIKKINEN says it is super versatile and easy. I stack, then do the processing before finishing it off in Gimp.

Siril - Full image processing (pre-processed with scripts)

These are the scripts I mentioned, I use OSC pre-processing with background extraction which stacks and gives an even background at the same time. You can redo background extraction after stacking if you wish.

[Olli]

17 hours ago, ONIKKINEN said:

Siril is really good for a beginner, or for more experienced imagers.

You can do the core parts of processing in Siril quite easily, which would be (simplified) something like:

1. Crop out incomplete edges from stacking
2. Remove light pollution/sky gradients with the background extraction tool
3. Colour calibrate using the Photometric color calibration tool
4. Stretch using Asinh transformation and/or Histogram transformation, or the hyperbolic stretch tool but that one is a bit more complicated and probably not something one should look at for the first few attempts.

At this point you could export the image in 16-bit TIFF format to further adjust in Photoshop however you like. This is more or less how i process my images, although i do a couple of things in PixInsight too.

You could stack in Siril too as it is a bit better and versatile than DeepSkyStacker, but its not nearly as straight forward as the drag and drop simplicity of DSS. So for a beginner my advice would be stacking in DSS, core processing in Siril and then final adjustments in Photoshop. Could use Gimp instead of Photoshop, but i think Photoshop is well worth the money with how many different things you can do in it.

Thanks for the advice very helpful. Will have a deeper look into Siril.

[ONIKKINEN]

2 hours ago, bomberbaz said:

Don't know if you are aware but Siril now has an auto stacking script which is simplicity itself to use.

Simply put the files for stacking in folders on PC and boom, click the script and sit back and wait for it to finish. 

Stacking multiple nights worth of data is much easier in DSS though, and all the directory creation work can be confusing for a total first timer (such as myself a few years ago).

I use my own scripts for calibration (master darks, master darkflat and a hot pixel map for 60/120/240s data on top of flats and lights, this way i dont have to stack 100 darks each time)    and then stack manually, because the auto stacking scripts is written so that you dont get to inspect the data or choose the registration/stacking parameters. Bit more involved, but this gets the best out of any dataset compared to full auto.

[bomberbaz]

3 minutes ago, ONIKKINEN said:

Stacking multiple nights worth of data is much easier in DSS though, and all the directory creation work can be confusing for a total first timer (such as myself a few years ago).

I use my own scripts for calibration (master darks, master darkflat and a hot pixel map for 60/120/240s data on top of flats and lights, this way i dont have to stack 100 darks each time)    and then stack manually, because the auto stacking scripts is written so that you dont get to inspect the data or choose the registration/stacking parameters. Bit more involved, but this gets the best out of any dataset compared to full auto.

I know what you mean and I manually check each FIT lights file before adding them to the folders, discarding any obviously poor ones.

I store my lights, dark etc folders for siril on a seperate drive so the drag simply copies them to that drive. That way I know I can delete when done safely without losing the masters.

TBH whichever you use can be a faff in one way or another and on my regular clean ups I often end up finding a folder or archive and then ask myself, wonder what that's for? 🤣

[ONIKKINEN]

18 minutes ago, bomberbaz said:

I know what you mean and I manually check each FIT lights file before adding them to the folders, discarding any obviously poor ones.

I store my lights, dark etc folders for siril on a seperate drive so the drag simply copies them to that drive. That way I know I can delete when done safely without losing the masters.

TBH whichever you use can be a faff in one way or another and on my regular clean ups I often end up finding a folder or archive and then ask myself, wonder what that's for? 🤣

Re: copying files

Give Link Shell extension a try? Its a windows plugin that allows you to create symlinks at will, which is essentially copying virtual links to the files to another location but with 0kb file size. I create a symlink of the raw data to the script directory and doesnt matter if its 5 or 500gb it all gets created instantly (and doesnt take any space on the drive).

Edited October 11, 2023 by ONIKKINEN
typo

[ollypenrice]

I'll just back up praise for Photoshop for post processing.  If it's going out of fashion for others, it ain't going out of fashion for me.

I find Pixinsight wantonly obscure but, if you don't have it, you have a problem: it's the only platform (I think) which supports Russel Croman's BlurXterminator and that is a deconvolution routine like no other I've tried. My routine, after stacking is,

Pixinsight: DBE or ABE.  SCNR Green. BlurXterminator.

Photoshop: Everything else.

Be aware, there is a lot more to learn in processing than there is in capture. It should not take you long to capture data at the limit of your equipment. But you will never reach a point where your processing gets the very best out of it.

Olly

[The Lazy Astronomer]

47 minutes ago, ollypenrice said:

I find Pixinsight wantonly obscure

Hmm, well I decided I quite liked PI's layout and structure - maybe that says more about me than anything else 😅 

[saac]

15 minutes ago, The Lazy Astronomer said:

Hmm, well I decided I quite liked PI's layout and structure - maybe that says more about me than anything else 😅 

x2 for PixInsight. I'm certainly not anywhere near as proficient as I would like to be with it or processing in general but it's PI's functionality and menu system that I like most. I also like the wealth of support re tutorials, books and youtube videos that are freely available to help.  

The greatest challenge I have with processing is simply skill fade, I go from one winter season to the next and spend most of the time just relearning what I did last last year. I would be lost without those tutorials. 

Jim

[The Admiral]

1 hour ago, ollypenrice said:

Be aware, there is a lot more to learn in processing than there is in capture. It should not take you long to capture data at the limit of your equipment. But you will never reach a point where your processing gets the very best out of it.

I usually say that imaging is 20% data collection, 80% processing. Processing can have a profound effect on the appearance of an image; it can make or break an image. That probably explains a lot 😉.

Ian

[ollypenrice]

40 minutes ago, The Lazy Astronomer said:

Hmm, well I decided I quite liked PI's layout and structure - maybe that says more about me than anything else 😅 

It does suit some people and not others. The reason I'm more at ease in Ps is that I like layers. I can copy an image onto a new layer, modify it and then decide where I do and don't want to keep the modification. I don't have to struggle to make a mask that covers just what I want it to cover, I can just erase small areas of one of the images.

Olly

[Olli]

Thanks all for the help, seems to be a lot of options out there.might be a bit of a long shot but  are there any places I can get data from so I could play around and practice while it’s cloudy?

[ONIKKINEN]

16 minutes ago, Olli said:

Thanks all for the help, seems to be a lot of options out there.might be a bit of a long shot but  are there any places I can get data from so I could play around and practice while it’s cloudy?

Some excellent quality datasets can be downloaded in the data release threads here:

https://stargazerslounge.com/forum/294-iki-observatory/

[powerlord]

I've been doing this nearly 3 years now. I've tried most stuff out there.

I stack with Astro Pixel Processor. it's a bit more flexible that Siril, and does mosaic really well too. It's also not too expensive for a perm licence.

I use Siril for astro bits and bobs like stretching, plate solving, photometric colour correction and deconvolution.

I use Affinity Photo for editing with the addition of Russell's StarXterminator and NoiseXterminator.

I hate Pixinsight with a passion on many levels. I won't go into them here, make your own mind up.

[alacant]

8 hours ago, powerlord said:

I hate Pixinsight

+1

We were relieved when the trial period ended!

Edited October 12, 2023 by alacant

[Clarkey]

Much as I find Pixinsight a right PITA, I have started using it - mainly for the Russell Croman tools. Also, I do believe if I ever get proficient it will give the best results. However, if I was ever to design a user interface it would be the opposite of PI. I did use Astro Pixel Processor and Affinity photo for some time (Affinity is a cheaper version of Photoshop but very similar functionality). Even with PI, I still use Affinity after to get the image looking more how I want. I suspect this is more to do with lack of ability in PI, rather than the software itself. If you want quick and easy results, give Startools a go with the free trial version. Personally, I don't like what it does to stars - but just going through using all the defaults will give you an acceptable image.

[edarter]

ask 100 people this question and you'll get 100 different answers. I think you have to find your own path to be honest, and there is a fine line between battling on with one piece of software because you are picking up the odd bit of extra 'goodness' out of it and chopping it in to try something else. For me that was photoshop, I know my round photoshop pretty well as a 'normal' photographer but REALLY struggled to get anything I was happy with out of it astro processing wise. I stayed with it probably 18 months longer than I should have because I didn't want the expense of using Pixinsight. I finally bit the bullet on PI and it was probably the single biggest jump in my processing results I've had. The interface can be a little querky but to be honest I don't find it that bad at all. FWIW my journey so far is as follows:

DSS & Photoshop - all sorts of issues stacking DSLR images in DSS (EOS600d)
Astropixel processor (APP) - the muts danglies for stacking as far as I'm concerned, all my issues went away, but didn't really like it for processing.
APP & photoshop - this is where I was for ages! APP for stacking, photoshop for processing
Siril - didn't like it that much, interface wise etc
Startool - Just don't like the results, or the interface. As others have said, too much of a black box. The stars look fake to me no matter what I try. I can tell a startools image at 1000 paces!
APP & Pixinsight - APP for stacking PI for processing. My word what a jump in output that was, instantly better images (though still nowhere near as good as many others on there!). 
I intend to try PI for stacking at some point but TBH it scares me - I read all the questions about it on facebook and in forums and it seems really complicated, so I tend to stick to APP for that!

HTH and no offence intended to any that like the software I've not got on with, its such a personal thing that I think everyone has to find their own path by trying them and seeing how they get on.

Ed

[Ouroboros]

I feel the weighted batch preprocessing in Pixinsight has been a game changer.  I like the way it sorts and associates different lights, flats and darks. Mind you, I’ve only used it for OSC so fairly straightforward. I like the pre-run check with flow diagrams showing what it’s going to do with the different files. The free-to-view YouTube videos by Adam Block on WBPP are excellent.

I decided some time ago that it suits me to try and learn one piece of dedicated software rather than to do processing across several.  At the time PI seemed to be the application of choice. So I went for that, although more alternatives are now available. 

I can now use PI for everything and can even (!) end up with a pleasing result. I am amazed though at the superb results some people can achieve with it.  I think some people are just natural virtuosi and others like me are not. 

I like the logical methodology of PI.  And it’s become considerably easier since the development of AI add-ons  like BlurXterminator etc. 

I’ve said this before on SGL, but the issue I have is that although I’m pretty much fine up to and including the point where the image is stretched to the non-linear state, after that  the options broaden into what feels like a huge multi-dimensional parameter space. OK I can finish off by doing curves and pushing up the saturation, applying a few masks and adjusting and enhancing local contrast and sharpness and so on. But I’m often left feeling I can’t really see what’s wrong with my image nor do I know how to make it look better. I think that probably comes down to a fault in myself rather than PI.  I have not found any of the reference books like Warren Keller’s  (excellent though it is) much help in this regard.  Other suggestions welcome. 

 

[edarter]

22 minutes ago, Ouroboros said:

But I’m often left feeling I can’t really see what’s wrong with my image nor do I know how to make it look better. I think that probably comes down to a fault in myself rather than PI.

 

exactly this!  I'm getting results that I'm happy with (way better than I could in photoshop) but don't know how to improve from here. It seems to be a case of fumbling around in the dark, or should I say processes, and just trying things, which feels really hit and miss.

[DaveS]

When I first moved from DSLR to CCD I went with AstroArt, version 5 at the time, and am now using version 8 SP2. I still find the stacking exceptionally good and blindingly fast. I was a long time PI refusenik until the RC AI tools came out which caused me to bite the bullet and stump up for a license. I still use AA* for some of my processing as I find the DDP stretch often gives better results that either ArcSinH or Masked Stretch though the latter is useful. The "Attenuate Single Colour" in AA8 I find very useful for its control and visualisation of the areas being attenuated, I've had a better reduction of magenta stars in SHO palette images than in anything else.

Occasionally AA8 will throw a wobbly when trying to stack NB images (I'm still not sure why) in which case DSS will do the job.

I have also had WBPP turn its nose up at my subs, but AA8 will do the job.

[powerlord]

I think it's just to be clear for beginners -  when @edarter say moving from PS to PI was a massive difference I don't deny that personal experience. I just want to state though that that is specific to that Op. (so no prob with what you said there - hope that is clear too).

What I mean is that if you understand the tools, and know how to use em, there is 100 ways to skin a cat. If you don't, then nowt will work - but maybe at that stage someone shows you how to skin that cat with whatever the latest tool you've bought is.. and voila! you might think - if only I'd bought that in the first place. When really, it was just that you didn't know how to skin the cat.

And as I have a whole 2kg family pack of McCanes oven chips on my shoulder about PI... I know I'm not best place to be impartial here. However - I have not seen a single image EVER edited in PI that could not have been edited just as well with something else - and like everyone else - I'm not an expert in everthing out there - so I can only say for sure that I've not seen a single image edited in PI that I know I couldn't have matched using Siril and Affinity Photo (+ russels filters which bar blurx work fine in affinity photo). I will make the one notable exception - which is that nowt beats blurx for deconvolution and it's only available in PI - but it's blurx doing the work there - PI is just the container! I wish Russell would bring it out standalone or as a filter but he won't even though I keep pestering him to).

So, for beginners all I'll say is don't get taken in by PI's mystique - it just does the same stuff siril and affinity photo (or PS) does - and in nearly all cases, in a decided more limited, slower, non-layered, prehistoric sort of way, hidden behind techno-babble. There's lots of PS and affinity photo tutorials on youtube - though less than I'd like to see - and it's the best way to learn imho.

I keep meaning to start an astrostu channel myself focused only on stacking/processing with my chosen tools - I'm sure there would be folk interested - but it seems a lot of effort and I tend to work really quickly - and for me for it to work as a channel, it would need to be posting videos of processing I'm doing in real time as I do em - if it's too fast for you - use the youtube controls to slow it down, etc. And when I did a trial one a year or so ago the feedback on here was that I WOULD need to slow down - and that just wouldn't be a long term solution for me.

Anyhoo that's an aside - the main thing I'm saying is:

1. do not for a second believe that PI is some sort of mystical panacea - king of the heap. It is not. It is an image editor (and stacker and a few other things). Just a very limited one that can only edit astrophotography pics. Treat it as such. Im reminded of as a biker, when folk say 'well I paid £800 for my ARAI helmet. I mean I only have one head, so it's worth getting the best'.. as if there is any evidence whatsoever that it's any better than a £50 one. If it has a BSI kite mark, and an ACU gold it's passed all the testing out there. the extra 750 might be getting you better safety but is just as likely to be emperors clothes mince.

2. Following on from 1 then - look at what is out there (including PI if you are willing to demonstrate you are not a criminal and are 'approved' for a trial licence..) , and decide for yourself what works for you. None of it is 'press a button. Some is more like that than others (e.g. startools), but to be honest, there's no way of avoiding some esoteric software somewhere in the processing - for me that's Siril: initially a black box of infinite misery to me that I ignored, until I bit the bullet and watched some tutorials and eventually worked out what it could do much better than affinity photo and Astro pixel processor (which I use for stacking - again a decision I made after trying DSS and Siril, though being on a mac, DSS needed to be run in a VM which limited it appeal to me out the box). In the end you work out a number of processes that work for you based on the type of processing (wideband, narrowband, SHO or HSO or HOO or whatever). and once you've got em, and yer happy with em - it's like anything else - it becomes really easy with practice.

 

Edited October 13, 2023 by powerlord

[ollypenrice]

2 hours ago, Ouroboros said:

But I’m often left feeling I can’t really see what’s wrong with my image nor do I know how to make it look better. I think that probably comes down to a fault in myself rather than PI.  I have not found any of the reference books like Warren Keller’s  (excellent though it is) much help in this regard.  Other suggestions welcome. 

You have it in one.

I've been running imaging workshops, giving tutorials, demos, whatever you want to call it, for years and I've been trying to improve my images for more years than that. The one thing I insist on is learning to look at the image.  When you have learned how to look at it you can see what needs attention. We can all stare at an image and fail to see that it's green. Or clipped. Or over saturated. Or just ruddy hideous!!!

To combat this, I built in certain rules in my workflow. 1) Measure the background sky at regular intervals. Ps lets me see its brightness and its colour balance in RGB at a click. I want it between 20 and 23 and equal in R, G and B. 2) Keep looking at the histogram. Is it clipped?  3) Keep doing hard test stretches.  You're not going to keep these stretches but is there any faint stuff that you've failed to drag out? 4) Take a break and look at other astrophotos you like, but not of the object you're working on at the moment. You are not trying to replicate existing images but extract the best from your data.

I'd also look at good astrophotos and ask yourself what's good about them. This will make you a better critical observer.

Olly