Donut Stars in HaRGB Synthetic Luminance

[brasspoodle]

Hi,

Looking around for solutions to combine Ha and RGB and avoid getting donut like stars (dense white core from the Ha and a donut of colour around it from the RGB). I assume the tighter Ha stars mixed with the RGB cause this. It may be an acquisition issue (i.e. shorter RGB subs, longer Ha) although light pollution and an average portable mount limit this. Currently have mixed 300s Ha (5nm) with 60s RGB subs. I've attached the Synthetic Luminance to show the result. Any advice on a StarTools workflow that would help with this would be great.

[admin]

There may be a few ways to tackle this, but much depends on the characteristics of the data.
Is is just one channel that is particularly "bloated"?

The Magic module may help, or some clever masking may alleviate the problem when compositing or when creating the synthetic luminance.

Of course, the problem would go away entirely when using just the Ha for the luminance...

[brasspoodle]

Upon further investigation it seems the halos occur when Autodev is applied to any of the data I have (RGB single or stacked subs and Ha single or stacked subs). Given this occurs on a single sub as well as the combined stacking I assume it is not a focus issue. Maybe there is a setting I need to adjust in autodev? Some of the files are here: https://www.dropbox.com/sh/6jy5jdvzmftf ... v0i1a?dl=0

As an update to this I've tried stretching more extremely in Nebulosity and the same halos appear so somehow these arise in taking the image with my Skywatcher f7.5 ED80 refractor, Astrodon filters and ZWO ASI 1600MM-Pro. Not sure where in this chain of things I can change something or are these just to be expected with this setup? If I stretch the image a lot less then I can avoid the problem.

Con 80 minutes in Ha.jpg (198.11 KiB) Viewed 6464 times

Actually found similar halos in my DSLR images but less pronounced. I might just be getting cleaner halos rather something that has come about through the change in camera/wheel setup?

Apologies for millions of false assumptions about things. I'm very new to processing Mono RGB and Ha images so likely I've missed the obvious. Really appreciate the advice and happy to provide more files/data if needed although the issue seems

[admin]

Many thanks for uploading - that was very helpful.
Since the issue is not present in the Ha data, this rules out optics-related secondary reflections as a cause.

One (likely) explanation I can think of is an issue with your LRGB filters;

http://www.astronomik.com/en/news/astro ... o_problem/
https://www.cloudynights.com/topic/3876 ... he-answer/

This is obviously an optical train issue and not a software/processing issue.

I don't think you uploaded your green channel data, but as I alluded to before, there may be some stop-gap processing tricks we could use to avoid the colored halos somewhat.

[brasspoodle]

Many thanks. I could reproduce the halos with the Ha data though as soon as I stretch it in Autodev? The stretch I attached above was very mild. This made me think it wasn't a filter issue but perhaps a focal reducer reflection issue? I've also noticed these halos in my old DSLR data (which also used the focal reducer), but slightly less defined. So, yes, definitely not a processing issue!

As for dealing with the halos I've tried shrinking stars in the Magic module by 5 pixels in the non-linear image (so no noise reduction etc.) and have found I need to get the star mask pretty spot on for this to work well, although the issue remains and is just less obvious. I imagine this might work better at an earlier stage in the processing, rather than a final fix? Attached the original image and the star shrunk image. Wonder what else you'd recommend?

Stars as they were:

Version 2 (1).jpg (276.92 KiB) Viewed 6458 times

Stars shrunk in Magic (still a messy experiment):

Version 2.jpg (294.42 KiB) Viewed 6458 times

[admin]

If you can truly replicate the issue in the Ha data, then we're looking at some sort of secondary reflection issue indeed. Another possibility is poor seeing or haze.

That said, to me, the star "bloating" looked much less severe in the Ha data, to the point of being unobtrusive (IMO). Your optical train's diffraction pattern will always cause point lights to spread their energy around their local neighborhood, so don't expect perfect point lights.

In the other channels, the halos are indeed excessive (blue being worst by far).

How does it look when you use solely Ha as luminance, using R G and B only for color?