Suggestions for controlling relative star brightness in Pixinsight please.

[Ouroboros]

What’s the best way to reduce the brightness of small stars whilst keeping the brightness of the bright ones?  I find Curves doesn’t do this very well.

Is there a script that leaves the main, bright stars unaffected, but attenuates the small stars?

[Budgie1]

Have a look at this script:

 

[Ouroboros]

Thanks, @Budgie1.  I’ve got that. I tend to use blur exterminator now. 

But isn’t Star Reduction doing the exact opposite of what I’m saying. It allows you to protect  small stars. What I want is, having reduced stars in blur exterminator, to protect the brightest stars whilst reducing the many small stars.

Trying to do this with levels or curves doesn’t work because it tends to give bright stars a cut off look - as the screen shot in this much zoomed in image (left) shows.

Edited January 13 by Ouroboros

[Budgie1]

Have a look at MorphologicalTransformation.

Using the settings below, it maintains the brighter stars (with their coloured halos) but reduces the smaller ones. I think you increase the "Amount" setting to reduce further, but I've not really played with it that much.

[Ouroboros]

Thanks, @Budgie1.   I’ll look into that.  I also had a reply from Adam Block on the Pixinsight Forum drawing my attention to the star de-emphasiser script. Apparently this can be used to preferentially de-emphasise smaller stars. I have yet to look into it. I’ll report back. I might be some time.   

[ollypenrice]

Do you use StarXterminator?  It's a wonderful tool for controlling star size, and manipulating the Curve on the star-only layer does give control over fainter and brighter stars independently. I don't do this in PI but in Ps, but it ought t be the same.

Olly

[Ouroboros]

6 hours ago, ollypenrice said:

Do you use StarXterminator?  It's a wonderful tool for controlling star size, and manipulating the Curve on the star-only layer does give control over fainter and brighter stars independently. I don't do this in PI but in Ps, but it ought t be the same.

Olly

I do use StarXTerminator. Unless I'm somehow missing it, I cannot see anything that allows fainter stars to be processed in a different way to brighter stars. 

This is the user interface I see. It's more of less press button and go.  Maybe how it works in PS is different. 

[ollypenrice]

It seems that you can generate a star image, which I take to mean 'stars only.' That's what it gives me in Ps and that image can be manipulated in Curves before being finally applied over the background. Pulling down the bottom of the Curve more than the top will bring the fainter stars below the brightness of the background, so they become invisible.

Olly

[Ouroboros]

4 hours ago, ollypenrice said:

It seems that you can generate a star image, which I take to mean 'stars only.' That's what it gives me in Ps and that image can be manipulated in Curves before being finally applied over the background. Pulling down the bottom of the Curve more than the top will bring the fainter stars below the brightness of the background, so they become invisible.

Olly

Yes, that’s what I have done hitherto, but it leads to unsatisfactory results IMO. In fact to the artefacts I referred to in the left-hand picture above ie round bullet shaped bright stars with hard edges.

The reason is obvious of course. Curves works on intensity. The Curves process  is completely unselective of whether it attenuates the faint outer edge of bright stars or the fainter stars themselves.

What we want is a process which says leave those bright stars completely alone and kill or at least de-emphasise the fainter stars. 

[ollypenrice]

40 minutes ago, Ouroboros said:

40 minutes ago, Ouroboros said:

The reason is obvious of course. Curves works on intensity. The Curves process  is completely unselective of whether it attenuates the faint outer edge of bright stars or the fainter stars themselves.

Yes, that's true. I do, quite often, selectively process a handful of bright stars which haven't responded well to the overall stretch. This is easy with the selection tools available in Ps but I don't know if PI has an equivalent.

Olly

[Ouroboros]

1 minute ago, ollypenrice said:

Yes, that's true. I do, quite often, selectively process a handful of bright stars which haven't responded well to the overall stretch. This is easy with the selection tools available in Ps but I don't know if PI has an equivalent.

Olly

The only process I am aware of is the star de-emphasis script which brings together in one place a step of processes developed and described by Adam Block. I have not yet investigated it in detail, but as far as I can see it works by using the MLT process to differentiate between objects (in this case stars) on the basis of their size. Bright stars are bigger than small stars. So big stars can be masked and protected from being de-emphasised, whilst smaller stars less so.

Even so the star de-emphasised script appears a bit clunky from my current position of ignorance. I am a bit surprised that in these days of AI driven processes someone hasn’t come up with a script or process to make this a simple one-step tool ie leave the top  10, 20, 30% of stars (user set) untouched and de-emphasise the rest. 

[Ouroboros]

Having now installed and played with the star de-emphasis script written by Maxim Valenko it does indeed appear to do what I want. ie De-emphasise the small stars while leaving the bright ones alone. Result!

Thanks for suggestions and thoughts. 

[Budgie1]

2 hours ago, Ouroboros said:

Having now installed and played with the star de-emphasis script written by Maxim Valenko it does indeed appear to do what I want. ie De-emphasise the small stars while leaving the bright ones alone. Result!

Thanks for suggestions and thoughts. 

I've just tried the star de-emphasis script on a wide-field image taken with my Samyang 135mm and it works very well.

This example is just ABE, STF stretch, create the starless version and then apply the star de-emphasis script, jst as a tester. No stars are gone and the larger stars & their halos remained at original size.

Thanks for finding this script, it'll came in handy.

[Ouroboros]

13 hours ago, Budgie1 said:

I've just tried the star de-emphasis script on a wide-field image taken with my Samyang 135mm and it works very well.

This example is just ABE, STF stretch, create the starless version and then apply the star de-emphasis script, jst as a tester. No stars are gone and the larger stars & their halos remained at original size.

Thanks for finding this script, it'll came in handy.

Good. Glad you found it useful. I now understand  there’s a more recent way to do this using BlurXTerminator.  Maybe someone knows how this is done? I am familiar with and use BXT, but how is used to only de-emphasise smaller stars? 

[ngc1535]

On 15/01/2024 at 05:55, Ouroboros said:

 

Even so the star de-emphasised script appears a bit clunky from my current position of ignorance. I am a bit surprised that in these days of AI driven processes someone hasn’t come up with a script or process to make this a simple one-step tool ie leave the top  10, 20, 30% of stars (user set) untouched and de-emphasise the rest. 

I appreciate you guys mentioning my method for star de-emphasis. Yes, I came up with it back in 2018... well before BXT. 

In my most recent video (NGC 1491, Fundamentals) I leverage BXT to do this job. If you understand my original methodology... you can probably figure out what I did. However, being a member of my site will save you from guessing.

-adam

 

[Ouroboros]

13 hours ago, ngc1535 said:

I appreciate you guys mentioning my method for star de-emphasis. Yes, I came up with it back in 2018... well before BXT. 

In my most recent video (NGC 1491, Fundamentals) I leverage BXT to do this job. If you understand my original methodology... you can probably figure out what I did. However, being a member of my site will save you from guessing.

-adam

 

Hello, Adam. Mmm! What challenge!  I’ll have to get thinking. I keep meaning to sign up for some of your video series. Trouble is I keep getting distracted by other stuff. 

[wimvb]

When you use morphological transformation to reduce stars, you can to a certain degree, control which stars to transform. Create a star mask that uses few layers (scale) and adjust the noise threshold. Use morphological transformation with morph selection.

https://pixinsight.com/forum/index.php?threads/really-need-removal-of-tiny-stars.8833/

[Ouroboros]

Thanks @wimvb.  Looks like there are several ways now to do it. Cheers.  

[bosun21]

Thanks to this thread I now fully understand why people call astrophotography " The dark arts".

[wimvb]

1 hour ago, bosun21 said:

Thanks to this thread I now fully understand why people call astrophotography " The dark arts".

It can also function as an explanation of the saying "there's more than one way to skin a cat."