StarXTerminator struggles with saturated stars in NGC 6960 processing - Forums

I’m running into a pretty specific issue while processing a starless version of NGC 6960 (Western Veil) with StarXTerminator, and I’m curious if anyone has found a better workflow for this kind of target.

The problem is that some large, saturated stars are not being fully recognized as stars, so parts of them remain in the starless image, while the stars image ends up with incomplete or partially “eaten” stars.

The problematic stars are mainly those with:

large halos
saturated cores
nebula filaments crossing very close to the star

My setup:

Sky-Watcher Evostar 72ED
ZWO ASI585MC Pro
PixInsight workflow
StarXTerminator tested both on linear and non-linear data
Large Overlap enabled

I already tried:

running StarX on linear data
downsampling before StarX
BlurXTerminator before StarX
Large Overlap
pre star-reduction
double-pass StarX workflows

…but the behavior stays very similar.

What I’m noticing is that StarX seems to interpret part of the stellar halo as actual Veil nebulosity.

At this point I’m considering a different approach:

creating a manual mask only for the residual stars
applying local sharpening/star reduction only on the problematic stars
or rebuilding/fixing those stars directly in the stars image

Has anyone experienced the same issue on filamentary nebulae like the Veil / Cygnus Loop?

I’d especially like to know:

if there’s a more reliable workflow
whether full starless separation is simply not ideal for these targets
or if local manual correction is basically the best solution

I’ll attach a crop showing the issue.

Apparently removing stars from the sky is still harder than removing wrinkles from social media selfies.

📷 masterLight_BIN-1_3840x2160_EXPOSURE-120.00s_FILTER-NoFilter_RGB_autocrop.jpg📷 NGC_6960.jpg

I know it’s not ideal, but if nothing else works, you can always try to clone stamp these away. If not done very carefully though, it can later on create horrible looking artifacts.

make one with starnet++ and one with starxterminator.

Blink both and copy the better areas to one picture.

You will see, that especially starxterminator looses some faint areas because no software is perfect. So take the best from both for one picture.

If there are halos in both … try what Obbe J has told.

I had already thought about the Clone Stamp, but if I'm not mistaken this allows me to "delete" the star and not "clone" it in the "starry" image that will be used to recompose everything, in this way the final image will have fewer stars, am I wrong?

Obviously this was only done with a screengrab from here, and I apologise in advance as I don’t like doing this without consent, but;
This isn’t so bad, basically only one problematic star partially left behind. Correct only, followerd by SXT using large overlap. But the permanent fix to your issue is to root out what causes the problems in the first place. Several things look off in your image.

1. Your backfocus is off, this may not be a dealbreaker to you - but processing isn’t going to fix the edges, it’s going to attempt to mask them - but they will not look as good.

2. You didn’t mention what kind of filter you are using, the 585MC only has an AR cut window, so you will at the very least need a UV/IR cut filter in front of it. Not using one will result in star bloat and questionable focus.

If you did use a filter then maybe that filter is no good.

3. You don’t mention gain, but HCG mode kicks in at 200. This is pretty high and you should use the histogram to adjust your exposures accordingly to avoid saturating the stars if this is what’s happening in your case. There’s basically no reason to push exposure lengths too far. If you still want to use longer exposure times then you should be reducing gain but at the possible cost of added noise.

I'd say it handles them pretty well.

With my 72ed and 533 it seemed a lot worse.

As said above, you can clone stamp them out, but that might make thos stars look worse overall. Might not :)

Also you could try giving the image a bit of stretch before removing the stars. This might help.

Might not :)

To be fair this target has pretty horrendously bright stars, second only to Alnitak near horse head/flame nebula in my limited experience :(

This is an ill advised 4 panel mosaic taken not long after getting a mount and filter. Maybe you smaller pixels don't help, but I'm not sure about that :) maybe helpful comparison.

https://telescopius.com/pictures/view/199619/deep_sky/sh-2-103/veil-nebula-4-panels-80mins-per-panel/by-tiffsandastro

I don't think you have anything to worry about.

TiffsAndAstro · May 18, 2026, 11:30 AM
I'd say it handles them pretty well.
With my 72ed and 533 it seemed a lot worse.
As said above, you can clone stamp them out, but that might make thos stars look worse overall. Might not :)
Also you could try giving the image a bit of stretch before removing the stars. This might help.
Might not :)
To be fair this target has pretty horrendously bright stars, second only to Alnitak near horse head/flame nebula in my limited experience :(
This is an ill advised 4 panel mosaic taken not long after getting a mount and filter. Maybe you smaller pixels don't help, but I'm not sure about that :) maybe helpful comparison.
https://telescopius.com/pictures/view/199619/deep_sky/sh-2-103/veil-nebula-4-panels-80mins-per-panel/by-tiffsandastro
I don't think you have anything to worry about.

Congratulations! It seems like a great thing to me

Jan Erik Vallestad · May 18, 2026, 11:26 AM
Obviously this was only done with a screengrab from here, and I apologise in advance as I don’t like doing this without consent, but;
This isn’t so bad, basically only one problematic star partially left behind. Correct only, followerd by SXT using large overlap. But the permanent fix to your issue is to root out what causes the problems in the first place. Several things look off in your image.

1. Your backfocus is off, this may not be a dealbreaker to you - but processing isn’t going to fix the edges, it’s going to attempt to mask them - but they will not look as good.

2. You didn’t mention what kind of filter you are using, the 585MC only has an AR cut window, so you will at the very least need a UV/IR cut filter in front of it. Not using one will result in star bloat and questionable focus.

If you did use a filter then maybe that filter is no good.

3. You don’t mention gain, but HCG mode kicks in at 200. This is pretty high and you should use the histogram to adjust your exposures accordingly to avoid saturating the stars if this is what’s happening in your case. There’s basically no reason to push exposure lengths too far. If you still want to use longer exposure times then you should be reducing gain but at the possible cost of added noise.

1. Why do you tell me the backfocus isn't correct? What do you mean? Sorry, but I'm quite a novice; I focused with the Bathinov mask and it was practically perfect.

2. You're right, I didn't specify the equipment clearly; the camera is an ASI585MC Pro (set at 0°C) with an Optolong l-eNhance filter.

3. The gain is indeed very high; I set it to 250 on purpose because this session was just a test; this session was definitely not intended to produce a good image, but rather to experiment... I only had two hours to spare and wanted to try out the new camera (previously I had the uncooled version) and the new filter. I decided to attempt the photo session, knowing that I had little time, so I turned up the gain too high.

The first thing to realize is that stars do not always have to be removed from an image. The easiest fix is no fix at all but to celebrate and portray what is actually there. It’s not wrong to do that just because it’s not thr “style” at present. Case in point, I usually do not remove stars in detailed galaxy images because it it can be very destructive to the tiny details when the software can’t distinguish between stars and not stars.

If you insist on removing the stars then the first steps I would take would be change your shooting technique to minimize the impact of the bright stars in the field. First thing is gain. With the ZWO585 the sweet spot is at a gain of 200. That’s where HCG mode kicks in giving you the most dynamic range in combination with low read noise. There’s really very little reason to shoot this sensor with any other setting. Set your gain to 205-210 to be sure your in HCG and forget it.

Another thing that will help is short exposure times. You can go as low as 15 sec. with this camera and not loose any detail in the faint areas. Keeping subs short will reduce or eliminate star clipping and bloating and should help a bit with extended star halos.

Another thing that can help with extended halos is to make sure your optics are extremely clean and dust free. Also be on the lookout for high hazes that you’ll never see visually but will scatter light in your images. The best way to prevent that is to make sure you don’t have any subs with an abnormally small star count. There are different ways to do this so let me know if you don’t know how to do it.

If you do these things you might find that the star removal software is better behaved. giving you the result you want without the artifacts.

In astrophotography, you can read everything and the opposite of everything: some say to keep the gain low to avoid activating HGC, others to reach a long enough exposure time to have a good separation from the sky background, but not to saturate the stars (mount permitting). For a novice, all of this is very "disturbing" 😄. However, cleaning the lenses is something I always put off doing!

There’s a lot of conflicting advice out there for a lot of reasons. Some of it is just old and incorrect, some is just misunderstanding. Most of it is well meaning and well founded but the problem is, what applies in one situation, under certain skies with certain hardware won’t apply in another. The best way to work with that is to use a given set of advice as a launching point into your own research into the issue. You won’t be able to really separate good advice from bad until you really understand what’s behind it.

Let’s take a look at what I said about gain settings as an example. You have to understand what read noise, dynamic range. really mean. The next thing in your case is to look at what the available performance data for your sensor really means. Here’s what it looks like for the ZWO585MC:

📷 585mc performace curves.webpOnce you understand what these graphs are tell you, you’ll know what settings are best for what you’re trying to accomplish. The bottom line is, this hobby is very complicated. There’s certainly no easy button, even with “smart” telescopes. The good news is there’s plenty of people here to help!

Igor Fulvi · May 18, 2026, 11:02 AM
I had already thought about the Clone Stamp, but if I'm not mistaken this allows me to "delete" the star and not "clone" it in the "starry" image that will be used to recompose everything, in this way the final image will have fewer stars, am I wrong?

The way you could use Clone Stamp in this case is by using it on the starless version.
You can get rid of the halos by setting your reference (CTRL + LMB) right next to the halo (so the gradients would be about the same) and then removing the halo.
This is a process that takes some time to figure out, you have to tweak the opacity and softness a bit.

Here is a quick clone stamp I did on the screenshot you provided.
It can sometimes be quite difficult to perfectly remove the halo without ruining any of the nebulosity, but sometimes you have to live with a little halo.

You can always try to make the stars a bit more prominent in your final image, so that the halo’s will be less noticeable.

📷 image.png📷 image.png📷 image.png

Tony Gondola · May 18, 2026, 05:34 PM
There’s a lot of conflicting advice out there for a lot of reasons. Some of it is just old and incorrect, some is just misunderstanding. Most of it is well meaning and well founded but the problem is, what applies in one situation, under certain skies with certain hardware won’t apply in another. The best way to work with that is to use a given set of advice as a launching point into your own research into the issue. You won’t be able to really separate good advice from bad until you really understand what’s behind it.
Let’s take a look at what I said about gain settings as an example. You have to understand what read noise, dynamic range. really mean. The next thing in your case is to look at what the available performance data for your sensor really means. Here’s what it looks like for the ZWO585MC:
📷 585mc performace curves.webpOnce you understand what these graphs are tell you, you’ll know what settings are best for what you’re trying to accomplish. The bottom line is, this hobby is very complicated. There’s certainly no easy button, even with “smart” telescopes. The good news is there’s plenty of people here to help!

You're absolutely right, we need to approach this issue rationally. From the graph you posted, which I've looked at a thousand times without ever stopping to look at, it's clear what you're saying. The only thing I don't understand is DR(stops). What does it refer to?

Igor Fulvi · May 18, 2026, 06:20 PM
Tony Gondola · May 18, 2026, 05:34 PM
There’s a lot of conflicting advice out there for a lot of reasons. Some of it is just old and incorrect, some is just misunderstanding. Most of it is well meaning and well founded but the problem is, what applies in one situation, under certain skies with certain hardware won’t apply in another. The best way to work with that is to use a given set of advice as a launching point into your own research into the issue. You won’t be able to really separate good advice from bad until you really understand what’s behind it.
Let’s take a look at what I said about gain settings as an example. You have to understand what read noise, dynamic range. really mean. The next thing in your case is to look at what the available performance data for your sensor really means. Here’s what it looks like for the ZWO585MC:
📷 585mc performace curves.webpOnce you understand what these graphs are tell you, you’ll know what settings are best for what you’re trying to accomplish. The bottom line is, this hobby is very complicated. There’s certainly no easy button, even with “smart” telescopes. The good news is there’s plenty of people here to help!
You're absolutely right, we need to approach this issue rationally. From the graph you posted, which I've looked at a thousand times without ever stopping to look at, it's clear what you're saying. The only thing I don't understand is DR(stops). What does it refer to?

Stops is the way the graph is expressing dynamic range. Notice that it says “DR(stops). That means it’s showing the available dynamic range in F stops. F stops are used in conventional photography to express differences in brightness so it’s something very intuitive for anyone with a photographic background. The main thing to remember is that the greater the DR(stops) is, the greater the range of brightness you can record before clipping the bright parts of the image, think stars here because they are usually the brightest things in the image.

I've studied those graphs (just a little) and understand the meaning of DR, and I can agree with the 200 gain, and I understand why you suggest short exposures: because the full well is much lower and therefore the stars can easily be saturated. However, in situations like the Veil with stars dominating, perhaps it would be better to have ample electron well availability and therefore extend the exposure times with better granularity. Again, it's always a question of equipment and also of the subjects to be framed.

You could go with a gain of 0 to gain a half stop of dynamic range but look at what happens to the read noise, it skyrockets. That’s why a gain of 200 is the sweet spot.

Extending the sub exposure times will not help.