Jump to content

SkySurveyBanner.jpg.21855908fce40597655603b6c9af720d.jpg

Ways of dealing with Oiii noise in narrowband mono recomposition?


Recommended Posts

I've recently taken up the dark art of mono narrowband, and most excellent it is too, but brings its own sets of hills to climb, not least being the invariably noisy Oiii stack

I can stack, calibrate and register the separate stacks, and linear match the Ha, Sii and Oiii to get three balanced masters. I understand the basics of pixel math and processing, but I'm struggling with dealing with Oiii noise.

On my recent Jellyfish Nebula this process resulted in a swamp of noise in the Oiii linear stack, due to the weak Oiii, which is then carried into the pixelmath SHO combination. The Ha and Sii on the other hand were pretty good, not too far apart, so I assume the linear match is pulling up the Oiii (and noise) to match these two 

Apart from the obvious of tripling my Oiii integration, is there any mileage in approaching this in a different way, namely stretching each stack separately (and after Starnet 2 removal) rather than linear matching? And trying to deal with the noise in Oiii this way?

 

 

Link to comment
Share on other sites

Your image was quite impressive. I found it a bit green though, I usually do a green noise removal and you suddenly are left with that golden yellow blue hubble look. Or you can align the histogram peaks individually to get a similar result.

O3 is noisy unfortunately. I usually pre process each mono layer separately, stretch a little, make starless, then stretch again and manually align them back to their own star layer for orientation. I do all this manually though, you have much better control over the aesthetics making adjustments in real time within PS/GIMP. If the layers are particularly noisy I'll run a very mild Topaz on them prior to combining them back. 

Personally I find very smooth images "unnatural", at least if there's some noise in the image it looks like it has been captured by an actual camera by someone. Same with star distortions.

Edited by Elp
Hubble not bubble
Link to comment
Share on other sites

14 hours ago, Elp said:

Your image was quite impressive. I found it a bit green though, I usually do a green noise removal and you suddenly are left with that golden yellow blue hubble look. Or you can align the histogram peaks individually to get a similar result.

O3 is noisy unfortunately. I usually pre process each mono layer separately, stretch a little, make starless, then stretch again and manually align them back to their own star layer for orientation. I do all this manually though, you have much better control over the aesthetics making adjustments in real time within PS/GIMP. If the layers are particularly noisy I'll run a very mild Topaz on them prior to combining them back. 

Personally I find very smooth images "unnatural", at least if there's some noise in the image it looks like it has been captured by an actual camera by someone. Same with star distortions.

Thanks Elp. I tried the green noise removal in Siril but it drains all the colour, areas appear almost monochrome. I'll try playing around with some different combinations in Siril Pxlmath. The Ha mapped to green in SHO is the cause I think. Ill try blending some into the other channels.

I'll also try your suggestion of stretching each band separately. I don't have any denoise software so I'll try to get more Oiii (if we have any more clear nights...)

Link to comment
Share on other sites

I would rather say that OIII is faint, rather than noisy. :grin:

Russel Croman's Noise Xterminator is an order of magnitude better on astrophotos than any other NR routine I've tried. Because our RASA 8 data invites extreme stretching, I do often lift it well above the noise floor but StarXt sorts it out in an invisible or, in extreme cases, almost invisible manner. Here's a close crop pushed beyond its limit. Before Noise Xt...

Before.jpg.9623aad01e5fdefc31f6fb19d47d5dc0.jpg

After noise Xt...

After.jpg.85ec332eeb4694b42a74facb44a0c3cf.jpg

Had I wanted to erase the brighter parts of the noise reduced image, feeling that the NR had damaged them, this would have been a ten second operation in Photoshop Layers but, really, are they damaged? Here I had the 'Preserve details' option in StarXt set to minimum. Alternatively, the noise reduced image could now be further sharpened.

Olly

 

Edited by ollypenrice
Clarification
  • Thanks 1
Link to comment
Share on other sites

1 hour ago, 900SL said:

denoise software

You can try the Gmic plugin for gimp as it's free, Iain's Denoise operation smooths quite well but you have to dial it back until it's quite subtle.

  • Like 1
Link to comment
Share on other sites

22 hours ago, ollypenrice said:

I would rather say that OIII is faint, rather than noisy. :grin:

Russel Croman's Noise Xterminator is an order of magnitude better on astrophotos than any other NR routine I've tried. Because our RASA 8 data invites extreme stretching, I do often lift it well above the noise floor but StarXt sorts it out in an invisible or, in extreme cases, almost invisible manner. Here's a close crop pushed beyond its limit. Before Noise Xt...

Before.jpg.9623aad01e5fdefc31f6fb19d47d5dc0.jpg

After noise Xt...

After.jpg.85ec332eeb4694b42a74facb44a0c3cf.jpg

Had I wanted to erase the brighter parts of the noise reduced image, feeling that the NR had damaged them, this would have been a ten second operation in Photoshop Layers but, really, are they damaged? Here I had the 'Preserve details' option in StarXt set to minimum. Alternatively, the noise reduced image could now be further sharpened.

Olly

 

Thanks Ollie. I had a play around with the noise reduction in Siril, it does improve the background although not on the level of your software. It's also a bit hit and miss, you need to get an initial stretch just so otherwise you get a lot of unnatural artifacts. It's all good practice and a learning experience:)

Link to comment
Share on other sites

21 hours ago, Elp said:

You can try the Gmic plugin for gimp as it's free, Iain's Denoise operation smooths quite well but you have to dial it back until it's quite subtle.

Thanks Elp. Is there a stand alone denoise app, or a decent astro denoise module fir Affinity Photo?

Link to comment
Share on other sites

I've only tried affinity briefly so not sure. Just reading apparently you can install Gmic on affinity.

Topaz is standalone and sometimes they have promos, try it first to see, I bought it as I knew it'd be useful to denoise my daytime raw files too rather than buy a standalone program for astro specific processing.

Link to comment
Share on other sites

12 hours ago, 900SL said:

Is there a stand alone denoise app,

There is Astro De Noise PY, which is a standalone software and free. It can leave images a bit “plasticky”, so I usually dial the settings back a tad.

Link to comment
Share on other sites

Posted (edited)
On 09/03/2024 at 21:50, WolfieGlos said:

There is Astro De Noise PY, which is a standalone software and free. It can leave images a bit “plasticky”, so I usually dial the settings back a tad.

Thanks Wolfie. I've just downloaded the NX software trial as per @ollypenrice recommendation as it is a direct plugin to Affinity. A first test on a recent M51 has me impressed. I'll try this on the Oiii stack. Now to figure out where you use this in the Siril/Affinity workflow...

This is a zoomed in crop on M51 with the 533MC/90mm F6. It was pretty noisy beforehand, NX seems to have preserved the detail

Before and after (and some colour adjustments)

20240313-M54BAFF.thumb.jpg.f67ffbe3914330fc7adeeb2355088161.jpg

M51F.thumb.jpg.696a74f82590a70f4a1924ca192c5475.jpg

Edited by 900SL
  • Like 3
Link to comment
Share on other sites

I denoise in a Photoshop top layer with the original underneath. This lets me erase, wholly or partially, any areas adversely affected by the NR. Often it's only needed on the faint stuff so it's easy to select and erase the bright stuff or just erase it by eye.

I sharpen in the same way, though I sharpen the bottom layer and selectively erase the unsharpened top layer.

Olly

  • Thanks 1
Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...

Important Information

We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue. By using this site, you agree to our Terms of Use.