22 replies to this topic

[jmillsbss]

Just beginning to work on my relationship with Siril.  A couple of weeks now, and I've watched a ton of YT vids and tutorials.  Getting better, the more I try.  But one thing I don't understand, when do you do the starmask/star removal?  Not all the tutorials even use this function.  Do you remove the stars before you stretch the linear image?  That seems like the right time so the stars don't get overstretched.  Any difference whether with galaxies or nebulae?  Is the workflow the same?

 

Again, just trying to get a handle on Siril for now.  Will add other programs as I get a little more adept at Siril, but it seems to be a very versatile piece of software. And although I don't understand it, the ability for users to write and share scripts and make improvements to the software seems like a positive.

 

Thank you for your time and helpful comments about the questions I have!

 

CS!    Jay

[rlmxracer]

Yes, You do star removal before stretching. You can also use this feature to remove haze from the core of globular clusters. 

[wkight]

I would add that the reason you run Starnet before stretching is that you will want to use different types of stretches for the stars and nebula.   If you were to stretch the stars as much as you stretch the nebula, you might blow out (saturate) the stars

 

Check out some useful tips on the Starnet website:  https://www.starneta...om/tips-tricks/

[BrickInTheSky]

Just want to add that if you image a Star Cluster, you also can do the star removal, discard the starless image and switch to the star image and just stretch it. Do not recombine with starless. This will get you a cleaner cluster image.

Edited by BrickInTheSky, 18 June 2025 - 01:34 PM.

[jmillsbss]

I would add that the reason you run Starnet before stretching is that you will want to use different types of stretches for the stars and nebula.   If you were to stretch the stars as much as you stretch the nebula, you might blow out (saturate) the stars

 

Check out some useful tips on the Starnet website:  https://www.starneta...om/tips-tricks/

Cool!  See, I didn't know about this!!!  You don't know what you don't know til you ask.  Thanks!

[jmillsbss]

Yes, You do star removal before stretching. You can also use this feature to remove haze from the core of globular clusters. 

I didn't know how to clean up a decent image I'd captured of M13.  I did the starmask then tried to work on the haze in the middle of the glob and I should have just discarded that data altogether.  It was a mess after I got through with it!  Using the starmask alone sounds like an interesting option.  I'll try that method.  Thanks!

[rj144]

The image needs to be stretched for star removal though.  If you do it on a linear image, it will autostretch, remove the stars, and then reverse the stretch to get back to a linear state.

 

So, even if you do it before stretching, it is "technically" still stretched.

[OzViking]

I've found that doing some limited stretches on the whole image, then running StarNet gives me better results. Using the GHS transform on the "half stretched" starless / starmask layers before recombining them works best for me.

 

I found that if I did not do this then I got all manner of strange artifacts (square patterns in the background and scatterings of random coloured pixels) around brighter stars. 

 

V.

[rlmxracer]

I didn't know how to clean up a decent image I'd captured of M13.  I did the starmask then tried to work on the haze in the middle of the glob and I should have just discarded that data altogether.  It was a mess after I got through with it!  Using the starmask alone sounds like an interesting option.  I'll try that method.  Thanks!

I am new to imaging also. On the video I saw about processing globulars he separated the stars from the haze and discarded the starless image. Then he proceeded to process the star mask image as the final image.

Edited by rlmxracer, 19 June 2025 - 08:38 AM.

[jmillsbss]

I've found that doing some limited stretches on the whole image, then running StarNet gives me better results. Using the GHS transform on the "half stretched" starless / starmask layers before recombining them works best for me.

I found that if I did not do this then I got all manner of strange artifacts (square patterns in the background and scatterings of random coloured pixels) around brighter stars.

V.

Well my trouble has been mainly with galaxies, where I worry the starmask is pulling a lot of galaxy detail in as misshapen stars. It seems bad on M101 and m51 particularly. Chunky stuff in the galaxy arms i guess.

Then with nebulae, if I don't pull out the stars pretty early, they'll come back in my combined image with miscalibrated colors, often with a ring of color. Not CA. I was using a Newt for all my galaxy work. More of a blue or purple limb, but the full circumference of the stars. If I do a photometric color calibration and do the starmask before any stretching, I have decent clean stars when I go to work on them prior to recombination. Seems to me that galaxies and nebulae require radically different workflows. No?

[rlmxracer]

What is the fix for when a couple of bright stars are not removed from the starless image?

[jmillsbss]

What is the fix for when a couple of bright stars are not removed from the starless image?

Well I don't know what the parameter is called but when you open the little gear tool in Starnet Star Removal, you can change the settings to ignore or include stars of a certain brightness or shape.  You just have to play with the settings for each image. Once you've adjusted the settings you hit the little selector icon (the 3 little yellow stars) and it will do another grab with your new settings.  I'm sure there's someone that can explain it more specifically and a little clearer.

[rj144]

Well I don't know what the parameter is called but when you open the little gear tool in Starnet Star Removal, you can change the settings to ignore or include stars of a certain brightness or shape.  You just have to play with the settings for each image. Once you've adjusted the settings you hit the little selector icon (the 3 little yellow stars) and it will do another grab with your new settings.  I'm sure there's someone that can explain it more specifically and a little clearer.

Hmmm?  There are no options to protect stars that I've seen.

 

 

Where is that option?

Edited by rj144, 19 June 2025 - 04:04 PM.

[rlmxracer]

Hmmm?  There are no options to protect stars that I've seen.

 

 

Where is that option?

I believe it is on the bottom right corner of the dropdown of the Star Processing tab.

[jmillsbss]

I believe it is on the bottom right corner of the dropdown of the Star Processing tab.

That's what I think.  Not looking at it.  When the stars are all selected (when it has them circled in orange), there are options in that window.

[rj144]

That's what I think.  Not looking at it.  When the stars are all selected (when it has them circled in orange), there are options in that window.

The PSF star detection?  That has nothing to do with Starnet.

Edited by rj144, 19 June 2025 - 06:09 PM.

[rj144]

I believe it is on the bottom right corner of the dropdown of the Star Processing tab.

Thanks, but that does not affect Starnet.

[wkight]

The Seti Astro Suite (SAS)  has a few tools that might be helpful.   In particular, Blemish Blaster can be used to remove leftover bright star remains from the starless image.   Also, the Seti Astro Suite supports Starnet, but I believe it uses a different stretch prior to star removal; sometimes I find this works better, other times not so much.   Lastly,  SAS has an AI based star removal tool called darkstar that works well; try both and see how it works in your case.   

 

https://www.setiastro.com

 

 

[jmillsbss]

The Seti Astro Suite (SAS) has a few tools that might be helpful. In particular, Blemish Blaster can be used to remove leftover bright star remains from the starless image. Also, the Seti Astro Suite supports Starnet, but I believe it uses a different stretch prior to star removal; sometimes I find this works better, other times not so much. Lastly, SAS has an AI based star removal tool called darkstar that works well; try both and see how it works in your case.

https://www.setiastro.com

It's as good a time as any!

[OzViking]

What is the fix for when a couple of bright stars are not removed from the starless image?

I use the 'patch' tool in Affinity Photo to remove these. There is enough data in the star mask layer for the star to appear in the final image, so removing the halo / bright patch works well.

 

V.

[wkight]

Just want to add that if you image a Star Cluster, you also can do the star removal, discard the starless image and switch to the star image and just stretch it. Do not recombine with starless. This will get you a cleaner cluster image.

I just gave this a try on a project I did with M3 and it came out great.  I think the difference is quite noticeable. Thanks for the suggestion!

[Buzz999]

I've found that doing some limited stretches on the whole image, then running StarNet gives me better results. Using the GHS transform on the "half stretched" starless / starmask layers before recombining them works best for me.

 

I found that if I did not do this then I got all manner of strange artifacts (square patterns in the background and scatterings of random coloured pixels) around brighter stars. 

 

V.

If you do this, I still suggest you do a star removal first even if you end up stretching before removing stars again.This is to enable you to use your unstretched starmask when you recombine, giving you maximum stretch control over the stars you are reintroducing in your image recomposition. Otherwise the starting point for your stars during recomposition will be a stretched starmask, and you may be limited in how much you can alter the outcome at that point.

 

In a similar vein, if you make use of luminance, sometimes it is useful to use a luminance star mask in your final colour recomposition. Sometimes coloured stars can compete with the starless detail, and using the starmask from your luminance gives clean white stars, which may suit better.

[jmillsbss]

If you do this, I still suggest you do a star removal first even if you end up stretching before removing stars again.This is to enable you to use your unstretched starmask when you recombine, giving you maximum stretch control over the stars you are reintroducing in your image recomposition. Otherwise the starting point for your stars during recomposition will be a stretched starmask, and you may be limited in how much you can alter the outcome at that point.

In a similar vein, if you make use of luminance, sometimes it is useful to use a luminance star mask in your final colour recomposition. Sometimes coloured stars can compete with the starless detail, and using the starmask from your luminance gives clean white stars, which may suit better.

This seems reasonable to me. I've tried it both ways and do find I'm limited with star mods post recombination. Thank you!