Page 1 of 1

Help with starless image

Posted: Tue Jun 14, 2022 7:02 am
by Rowland
Hi. If anyone is happy to help? I have followed the unofficial starless instructions to a point- v1.8.

This is the Heal result after blob masking many stars with halos.

I've tried various masks, but how do I remove the grey spots left by the stars.

It probably has something to do with mask settings, but this is the best result so far after numerous attempts.

The image has a few more steps to go, but I'm lost getting a uniform background.
starlessST.jpg
starlessST.jpg (319.41 KiB) Viewed 3045 times

Re: Help with starless image

Posted: Tue Jun 14, 2022 12:28 pm
by hixx
Hi Rowland,
try to play around with 'new must be darker than old' setting and the 'New must be darker than....%' slider. These will help to reduce the brightness of the grey blobs and will search source pixels with brightness matching the nebula.

Good luck,
Jochen

Re: Help with starless image

Posted: Wed Jun 15, 2022 9:12 am
by Rowland
Hi Jochen.

Thank you. It's a tough one. But improvement with the "must be darker than new" setting, Yes and 99%. Still not enough to be happy with the result.

It would be ideal to mask the grey left by the stars but I can't get a mask to do that automatically and there are thousands upon thousands of stars.

The best result might be to remove the brighter stars and push the rest into the background. The grow blob function works quite well removing the grey over the red.

Re: Help with starless image

Posted: Wed Jun 15, 2022 11:49 am
by hixx
Hi,
another way to improve might be enlarging the Heal area in HEAL module a bit. Also make sure to "Grow bob in MASK 2 or 3 times after "Auto-Stars" selection. Then reduce new darker than to 30-60% or so. This is walking a thin line which is very much depending on the object surrounding the stars.
Cheers,
Jochen

Re: Help with starless image

Posted: Fri Jun 17, 2022 2:56 am
by admin
Hi Rowland,

Great to see you back on the scene, and I hope you're well!

For the removal of stars for aesthetic reasons, you may have more success with the AltStars preset in the Auto Mask generator (only available with Tracking on however!). This new feature (new in version 8) uses a star detection algorithm that is much better at detecting stars with full stellar profile included (change Star Aberrations and Star Size as needed).

E.g. single-click mask;
StarTools_2835.jpg
StarTools_2835.jpg (201.48 KiB) Viewed 2967 times
Hope this helps!

Re: Help with starless image

Posted: Sat Jun 25, 2022 4:41 am
by Rowland
Hi Ivo. Yes, it's been a while, and things seem busier on the forum with lots of great images. Too many to comment on them all. ST keeps getting better. Great work.

Unfortunately, I can't devote time to imaging and processing as before, but I am determined to give the starless image a go.

It's probably a mistake to think this image has potential—a single Heal iteration with most work in Shrink and Superstructure. There is room for improvement, and it's not so detailed but an interesting exercise.
r_r_r_pp_light_stacked.jpg
r_r_r_pp_light_stacked.jpg (499.51 KiB) Viewed 2767 times
Thanks again for your help.

Re: Help with starless image

Posted: Thu Jun 30, 2022 5:30 pm
by Rowland
I think this is a more sensible result with the stars pushed back. Widefield starless images are beyond my skills atm. The nebula is not overwhelmed, and the result is more natural without the distortion of star fields as in previous efforts.
ec.jpg
ec.jpg (346.68 KiB) Viewed 2739 times

Re: Help with starless image

Posted: Sat Jul 02, 2022 6:04 am
by hixx
Hi Rowland,
I think You might be able to make more mileage using a narrowbandfilter and NBA. Or You could use a duoband for Luminance - this would push stars back in capturing. It also would produce less problems with bleeding stars, SVDecon artifacts etc., minimizing the need of "overprocessing" in rder to bring out the nebula.

cheers,
Jochen

Re: Help with starless image

Posted: Sat Jul 02, 2022 10:52 pm
by Rowland
Thanks Jochen. That is valuable insight. I have spent hours on this image at various times. And now that you mention it, there is perhaps a workaround with the present data. I'll report back, eventually.