New star de-emphasis technique

[alan4908]

I noticed that Adam Block has published a written tutorial on a new Pixinsight star de-emphasis technique - see https://adamblockstudios.com/articles/star_demph 

It targets star halos rather than stars themselves, so it doesn't use an erosion filter. It doesn't work well on bright stars so you have to exclude them from your processing. The technique is also claimed to introduce fewer artifacts into the finished result than traditional erosion based techniques.

So, with a desire to increase my PI processing skills, I had a go with it on one of my own dense star field images and I was quite impressed.

Alan

Original

De-emphasised stars

Blink GIF

[Xiga]

Yeah I heard about this recently myself. It does look interesting.

Being a PS guy though, i'd love it if some smartypants was able to work out the mathematics of what is going on behind the scenes with this, and then reverse engineer a PS method that does likewise. Not pointing any fingers @sharkmelley ? ?

ps - In the example above though, i actually prefer the original version. Even though it's busier, i like that the star colours are more visible.

[souls33k3r]

Very nicely done mate.

[JamesF]

1 hour ago, Xiga said:

ps - In the example above though, i actually prefer the original version. Even though it's busier, i like that the star colours are more visible.

For me, the second version makes the fainter nebulosity around the edges of the PN easier to see at first which I like, but if I go back to the original then I can see it's also there.  On the other hand the mind-bogglingly dense star field of the original creates quite an impact too.  I'm not sure I could say that I prefer one over the other.

James

[davew]

Alan - Image No. 1 for me. If I wanted a star eater I'd buy a camera with it built in ? I think your image is just the job as you originally did it but it's all personal taste of course.

2 hours ago, Xiga said:

Yeah I heard about this recently myself. It does look interesting.

Being a PS guy though, i'd love it if some smartypants was able to work out the mathematics of what is going on behind the scenes with this, and then reverse engineer a PS method that does likewise. Not pointing any fingers @sharkmelley ? ?

ps - In the example above though, i actually prefer the original version. Even though it's busier, i like that the star colours are more visible.

There are a few actions you can purchase at a reasonable price -

Noel Carboni and Annie's actions. I have used Carboni's actions as I don't have PI and found them excellent, especially Local Contrast Enhance, Space Noise Reduction and of course, Make Stars Smaller.

Dave.

[Xiga]

1 hour ago, davew said:

Alan - Image No. 1 for me. If I wanted a star eater I'd buy a camera with it built in ? I think your image is just the job as you originally did it but it's all personal taste of course.

There are a few actions you can purchase at a reasonable price -

Noel Carboni and Annie's actions. I have used Carboni's actions as I don't have PI and found them excellent, especially Local Contrast Enhance, Space Noise Reduction and of course, Make Stars Smaller.

Dave.

Sorry Dave I should have said, I do already use both of these methods (and also the minimum filter as well) but I'm always on the lookout for new methods to test out ?

Annie's action can impact nebulosity (if ran enough times), Carboni's one darkens the background slightly, and the minimum filter (my preferred choice) can be a bit harsh if you're not careful, so it's always good to have alternatives to play about with ?

[davew]

No Problem Ciaran. I normally tried to run the actions twice on a copy layer. Then I could fine tune the result by stating with a layer opacity of 50% and then increase or decrease to taste.

There seem to be as many ways of doing things as there are people doing them

Dave.

[alan4908]

6 hours ago, Xiga said:

Yeah I heard about this recently myself. It does look interesting.

Being a PS guy though, i'd love it if some smartypants was able to work out the mathematics of what is going on behind the scenes with this, and then reverse engineer a PS method that does likewise. Not pointing any fingers @sharkmelley ? ?

ps - In the example above though, i actually prefer the original version. Even though it's busier, i like that the star colours are more visible.

Yes, I agree it's an interesting technique.

I don't see why you could not do a similar thing in PS (although I haven't actually tried).  As far as I can see all you need to do is end up with a contour mask which targets the star halos but protects the stars and a substitute pixel image which is used for the halo de-emphasis - I presume this could be generated via a context aware fill (rather than the PI MMT operation). A context aware fill is likely to give a more accurate representation of the background replacement pixels since it would automatically mirror the background noise.

Personally, I'm not sure I prefer the second image or not - I really was just using the image to demonstrate the technique and to improve my PI skills 

Alan

4 hours ago, souls33k3r said:

Very nicely done mate.

Thanks ! 

Alan

[alan4908]

15 hours ago, davew said:

Alan - Image No. 1 for me. If I wanted a star eater I'd buy a camera with it built in ? I think your image is just the job as you originally did it but it's all personal taste of course.

Thanks for the comment Dave. 

As I mentioned above, I was really just using one of my finalized imaged from my album (Deep Sky III if you are interested) to demonstrate the technique, rather than attempt to improve the image.

However, I think it is quite interesting and gives a different effect to that obtained by star reduction.  Previously, I have used PS based techniques such as the PS minimum filter and the PS plug in star shrink.

As to the end result, I totally agree - it is all a matter of personal taste. 

Alan