StarNet2 and StarXTerminator comparison

[drjolo]

I am currently considering a StarXTerminator purchase and decided to make a small comparison on some frames I captured last year. And here are the results. 

Each comparison contains six images. The first row is the wider view, and the second row is a 1:1 crop of detail. Starting from the left there is a raw stack, then StarXTerminator outcome and StarNet2 outcome. Both tools were installed as plugins to PixInsight and worked at default settings.

M31 fragment imaged with WO ED80 scope and QHY247C color camera:

Rosette nebula in Ha band imaged with 90mm APO and QHY268M camera:

Alnitak area in Ha band imaged with 90mm APO and QHY268M camera:

Tadpoles imaged with 90mm APO and QHY268M camera with Baader RGB filters:

 

All the samples above were processed in the linear phase before any stretching was applied. The last one is after stretching.

Lion nebula fragment with 90mm APO and QHY268M camera through Baader HSO filters:

 

And just for fun Cocoon nebula with some stars removed (StarNet2 was used here). Quite a dramatic view

[The Lazy Astronomer]

Edited to correct an error pointed out by vlaiv

If I'm reading the order of the images correctly, then it looks to me, on this selection of images at least, that Starnet is removing less non stellar detail than StarX, but StarX has dealt with Alnitak better, in that it has been completely removed (albeit with artifacts, but completely removed nonetheless).

Edited February 24, 2023 by The Lazy Astronomer

[vlaiv]

38 minutes ago, The Lazy Astronomer said:

If I'm reading the order of the images correctly, then it looks to me, on this selection of images at least, that Starnet is removing less non stellar detail than StarX, and has dealt with Alnitak better, in that it has been completely removed (albeit with artifacts, but completely removed nonetheless).

I think that StarX removed Alnitak completely, while Starnet preserves non stellar detail better.

@drjolo

What is your view on this after conducting the comparison? Which one did you find better?

[drjolo]

It is a really hard call for me. There are a few places where some small stars or star remnants were left by StarNet2 and were removed completely by StarX. On the other hand in a few places, the nebulosity detail seems to be a bit better preserved by StarNet2, but the differences are really super small. 

One thing I forgot to mention is that StarNet2 is much faster than StarX, about 5-6 times faster. It is not a big issue for a single image processing, but a few days ago I decided to process with StarXTerminator 200 comet images to remove stars before stacking. It took almost 30 hours. It is easy to calculate how much time we can save using StarNet2 for that task. The PC is an i7 with 16GB RAM, subframes were 24Mpx large. 

There are still two weeks of StarX trial left, so maybe I will make up my mind somehow  

[The Lazy Astronomer]

1 hour ago, vlaiv said:

I think that StarX removed Alnitak completely, while Starnet preserves non stellar detail better

Woops! That's unconscious bias happening right there - I see what I want to see 😁

Edited original post

Edited February 24, 2023 by The Lazy Astronomer