After much frustration waiting for the clouds and rain to stop here in the UK over the last few weeks I decided to take another look at the data I captured back in August 2023 of the Crescent Nebula. I was really very happy with myself when I first processed the data of 4 nights but since then I have pushed the boat out and gone PixInsights for my processing, also some of the AI based tools to take the place of deconvolution and noise removal.
All the data remains the same but in short there were some 20 hours of exposures through my l-enhance Optolong filter (now sold and upgraded to the l-extreme). I won’t go through the rest of the capture and processing other than it was Siril and Photoshop based. Take a look at the Blog for that session if you want more details.
This blog will take comparative view to see what difference the new tools and hopefully some more refined skills makes!
The old processed images (one cropped in more on the Crescent,
that I was very pleased with:-
The new image processed using PixInsight and Russ Crowman’s AI based BlurXTerminator / NoiseXTerminator / StarXTerminator:-
I'd say the reprocessing is an improvement, the main differences:-
I would say that the Pre-Processing in PI, although slow, provides better results, subjectively say 10% better.
The main differences I see is the BlurXTerminator which really does sharpen the image and reduce the stars more effectively that the deconvolution using PSF previously performed in Siril.
The NoiseXTerminator is also very good (a trial of this was run on the previous version and I did mention that I would by buying!)
Next, is that I have discovered Masking, this has really allowed me to focus the final improvement to bring out the blues in the Crescent and the Reds in the gaseous Hydrogen clouds separately. Once your get round some of the idiosyncrasies of PixInsight there is obviously a large amount of power here.
I leave here then with a little awe at how the same data processed using new improved tooling and techniques can provide a significantly improved image. Hope you agree!
Kommentare