Advice needed - star masking

[Moonstruck]

Hi all,
I have a question about background star masking. My images (nebulas in particular) generally have been turning out pretty well using Startools, with the exception of the background stars, which often are bloated and/or have bright halos around them, from increasing the intensity too much during processing. I am still in the process of getting up to speed with Startools, and there is a lot I still don't know. I have been increasing the intensity (stretching) my images a lot in order to have the nebulas stand out more, which creates fantastic nebula images, but in the process it causes the background stars to become way too prominent. I am certain there is a way in Startools to increase the intensity of the nebulas, but keep the stars at their original intensity, and that is really something I'd like to learn. Can that be done using the masking function? If so, can someone provide simple instructions on how to use it? Also, can the masking include the stars behind and within the nebula, or can only stars outside the nebula only be masked? Is there a way to maybe make a background layer of stars, and keep that at the original intensity, but stretch the nebula layer?

I am by no means tech savvy, so if you can provide some simple instructions, or point me to a user-friendly tutorial or video I'd really appreciate it!

Below is an example of the elephant's trunk nebula which I processed a while ago, and have posted on this forum before. I appreciate the input everyone gave me on that earlier, but I thought I would use it as an example again, as it is relevant to the subject of star masking. I posted both raw and processed images below so you can see again what I am referring to, particularly with regard to the one giant bloated star near the nebula... just an example of something I would like to fix in future processing. Thanks in advance.

Autosave raw data:
https://drive.google.com/file/d/18dr8CL ... sp=sharing

My previously processed image:
https://drive.google.com/file/d/1_eJQx2 ... sp=sharing

[admin]

Hi,

There are a few things at play here.

Some observations;

• The stars do not appear bloated at all, with good, well-defined cores. Star bloat is typified by the brightness of the core bleeding into the rest of the stellar profile. As a result 2 stars in the (apparent) close vicinity around HD 206267 are resolved and not hidden in any bloating.
• The halos aren't too bad, but the large "halo" around HD 206267 is likely caused by poor atmospheric transparency at the time of recording. Sometimes poor filter coatings can have the same sort of effect.
• I'm not entirely sure how this was stretched, but the stretching appears to very focused on dynamic range around IC1396, causing the extended nebulosity and dust to be largely invisible. This is, of course, a choice that is wholly up to you. However, confining the dynamic range to just that area will neglect allocating dynamic range to the rest of the image. As a result, stars (which tend to span the entire dynamic range), will stand out against a mostly invisible background, rather than being "embedded" in the rest of the detail that is present in the complex. Again, this is a choice that's as valid as any, but from your post, it sounds like this was not a choice.

E.g., a simple image-wide AutoDev, yields this;

Autosave(8)_ad_1024.jpg (217.5 KiB) Viewed 2002 times

Star location and prevalence tends to correlate with larger scale structures; where there are emissions there are stars. Where there is dust, stars are obscured. This is incidentally, the premise behind the Super Structure module as well and why the things like the Isolate or DimSmall presets in that module work so well.

Keeping the two balanced (super structure and star prevalence) will result in a much more pleasant (and informative!) image to look at.

Use local dynamic range manipulation (e.g. Contrast and possibly HDR if the image requires it) if you feel too much of an object is lost in the highlights (or even shadows). E.g. a default application of the Contrast module brings out IC1396 by locally re-allocating the dynamic range;

Autosave(8)_contrast_1024.jpg (221.08 KiB) Viewed 2002 times

Then, after SVDecon and Color module, you should get something like this;

Autosave(8)_contrast_svdeconcolor_1024.jpg (311.21 KiB) Viewed 2002 times

Notice the now visible Ha emission to the left, and notice how the distribution of stars feel like they "make sense", given the larger structures and different emissions that are now visible.

For your particular case, the Shrink module has a special "Unglow" feature that actively suppresses halos around "smaller" stars.;

Autosave(8)_ad_1024_halos.jpg (258.02 KiB) Viewed 2002 times

Autosave(8)_ad_1024_unglow.jpg (251.78 KiB) Viewed 2002 times

As mentioned above, you can further exploit the relationship/correlation between star location/brightness/temperature and physical processes/dust/emissions by using the Super Structure module;

Autosave(8)_ad_1024_dimsmall.jpg (207.67 KiB) Viewed 2002 times

Autosave(8)_contrast_svdeconcolor_shrink_dimsmall_1024.jpg (247.09 KiB) Viewed 2002 times

Now, stars don't feel "out of place", but appear like a natural part of their surroundings; because they are!
No special masking techniques required, no star removal, relayering or any other uncouth/selective processing shenanigans.

Finally, if you really wanted to, you could use the Filter module's "Reject" Filter Mode to reduce the brightness of the blue halos;

Autosave(8)_contrast_svdeconcolor_shrink_dimsmall_reject_1024.jpg (247.21 KiB) Viewed 2002 times

(no mask needed)

Season everything to taste obviously.

Hope this helps!

EDIT: I should also say that it is important to compare your images to other images taken in the same spectrum; IC1396 is a very popular narrowband object, and as such star prevalence will be much different/less in such narrowband composites.

[Moonstruck]

Thanks for the detailed reply and examples! Much appreciated. I will have to go through it in detail. You are correct that I narrowly focused the ROI on the elephant's trunk. I did that on purpose to try to make it stand out. But in retrospect, I should have included the surrounding nebulosity more, as it really adds to the picture!

Just a general question though, for other images. Can the masking function be used to keep the stars at a lower intensity (i.e., "mask" them), while stretching the more nebulous regions to make them stand out? And if so, are there instructions or perhaps a tutorial on how to do that? I saw the mask icon, and didn't know if it could be used for that, or if there are other reasons one would use it.

[Mike in Rancho]

Hi John!

When you see things like "masked stretching," that is generally going to be an other software technique (because they have no choice) and I believe is primarily meant for preserving star colors.

What data are you having star troubles with?

After setting your global stretch in AutoDev, optimized to your chosen settings, later on (usually post Color) you can try out various things in the Shrink module, which does use star masks.

Likewise, as Ivo noted, SS (dimsmall) can also be useful for maintaining the super structures while pushing back stars and background. Be sure to set your Airy disk parameter to suit! To the left for wide field, higher for narrow field. My personal rule of thumb is the default 50% works well for 600mm focal length and APS-C (without major cropping), and I adjust from there.

If you left your stretch weaker, you may be able to come back in with a second SS (brighten and/or saturate) to help enhance the target against the stars and background.

Finally, back to your original AutoDev again, if you can you might consider including a bright star within any ROI, so that the global stretch has that star to work with in setting the overall dynamic range. That may help keep bright stars under control if they are getting away from you (classic example of this is 52 Cyg in the Veil).

Second finally ( ), remember that you can go back to AutoDev at any time in tracking to re-do global stretch, although if you have used various enhancing modules since then it may not be 100% optimal (but generally isn't too bad at all and the differences would be minor).

[Moonstruck]

Thanks Mike! Good to hear from you. I’m not really having any specific issues with any particular image lately. My mount is in the shop now and unfortunately as a result I haven’t been doing any imaging lately. I left my rig out in the rain and also ruined my laptop! just asking in general terms for when I am imaging/processing again. My previous images using StarTools have been a huge improvement over what I was doing before, but looking back on them they all seem to have overly bright stars. That could have been because I had blown out the cores from imaging too long, or from over-stretching my images during processing. I’m trying to get better handle on the processing side, since I think that is where I need the most improvement!

[Startrek]

To answer your question “I am certain there is a way in Startools to increase the intensity of the nebulas, but keep the stars at their original intensity, and that is really something I'd like to learn”

You can use the Superstructure module furthe along in your process workflow ( towards the end either before color or after color )

Here’s how I do it ……

*****To enhance Nebula but keep background Star field the same original brightness

Create Star Mask - Mask > Auto > Filter Sensitivity use 1 , then Invert so stars are not sampled ( green background , stars are white ) shrink mask if some smaller stars are still green
Gamma : Try 0.20 to 0.50
Brightness, Color : Process Both for colour and Brightness for Mono or Synthetic Luminance
Saturation : 50%
Detail Preservation : Linear
Brightness Mask
Detail Preservation Radius : Default
Compositing Algorithm : Power of Inverse
Strength : 50% to 80% ( lower than 50% is too bright , higher than 80 % just increases dynamic range or brightness to selective areas of Nebula )
Brightness Retention : Off
Airy Disk Radius : default 50%
Mask Fuzz : default 1.0 pixels

Hope the above is helpful

Clear Skies

Martin

[Moonstruck]

Thanks Martin, I'll give that a try.