Red Blotchy Noise AFTER noise reduction

[Jockinireland]

Jim @Juno16 and I were discussing a problem with red blodchy noise in our images after noise reduction. i am using an ASI200mc and Jim the 533.

What I was finding was that after noise reduction (I routinely use EZ-Denoise in PI but it happens with other linear NR also), when I then stretched the image to non linear i would be fighting red blotch noise during the rest of the process and the backround was just not coming right. Ether I had this red blotchy noise or, if I did further non linear NR, the bg ended up "lumpy" and just not good.

Then I read a comment saying that this was happening with others with the new generation of ASI cameras and they suggested that there was actually no need/point in doing non-linear NR as the noise was so low with these camera's.

So, I did an experiment by taking an image which had been through e-z decon, duplicating it and putting the copy through ez-denoise then put both through MaskedStretch. This is a tight zoom of the result;

on the left is no linear NR - right is with ez-noise then both with MS using the parapeters and same preview location for bg reference.

Now the left hand one is fairly noisy but its only 2 hours (40 x 180s) of integration from a not great sky. I think if it was 100 subs that would be much better. I have had the blotchy bg on much longer integration also.

I have processed the "no linear noise reduction" one to completion including a gentle TGVD and its come out reasonably well - still a little noisy but no red blotch or lumpy bg..

its here if you wish to look


Anyway, just thought I'd share in case anyone is interested.

[STEVE333]

I think all of us at one time or another have had problems with a lumpy or plastic looking background. I could almost always trace the problem back to being too aggressive in noise reduction at the linear stage.

So, for your case, when doing the liner noise reduction try reducing the amount of reduction and/or number of iterations. I use MLT for the noise reduction in the linear phase with the settings shown below.

Hope this helps,

Steve

[Juno16]

Bravo David @Jockinireland for bringing this up! Also, excellent sleuthing work!

Yes, the EZ Denoise does state that it is to be used with a linear image and I also have been using it. Wow, very interesting results you showed!

I plan to give my last M101 image another shot in PI this week and thanks to you and Steve @STEVE333 , I now have some ammo to counter the red splotchy background noise (great and accurate description).

Thanks a bunch! Thank you also Steve for your nicely detailed illustration displaying your MLT settings and experience!

Thanks guys!

[Jockinireland]

thanks for that input @STEVE333 I had tried MLT following Waren Kellers plan (similar to yours) and was still finding the bg to have red blotchy patches. I'm going to try again tonight and try reducing the amount/iterations and see what that gives.

Good luck Jim @Juno16 Do let us know how it goes.

[STEVE333]

thanks for that input @STEVE333 I had tried MLT following Waren Kellers plan (similar to yours) and was still finding the bg to have red blotchy patches. I'm going to try again tonight and try reducing the amount/iterations and see what that gives.

Hope it helps. It took me a while to realize that "noise reduction" doesn't mean "noise elimination". I was trying to make my images noise free which never worked. Good luck with your processing.

Steve

[Jockinireland]

@STEVE333 - So I went back and tried doing MLT and using many different parameters, but no matter what I tried, when I stretched I was still finding the nasty red blotchy appearance. But thanks for you help anyway - it was valuable to learn how to properly use MLT at linear rather than rely purely on EZ-denoise.

@Juno16 Jim, how did you get on?

[Juno16]

@STEVE333 - So I went back and tried doing MLT and using many
@Juno16 Jim, how did you get on?

Hi David,

Been too busy to try it out yet. Have to go out of town today for a medical procedure so it might be a few days before I get to test it.
I think you are on to something with EZ Denoise and I am looking forward to running through the new process with my M101 data. I’m determined to get a full process in PI.
I looked back at my M51 image and it it really effected by the background red blotches, where M63 is not.
Created a new master light of M101 last night with light frames that I ran through the Subframe Selector Process. Pretty cool stuff.
Have you worked with the Subframe Selector Process? I wonder if the output can be used with the WBPP script since the WBPP already has a Subframe Weighing section?
I’m going to try !

[Jockinireland]

Hi Jim, Good luck with your procedure. I had surgery on my larynx last week (removing a benign growth) so I hear ya!

Yes my red bloches are not on everything either - I think it may be more prevalent when the sky is not particularly clear - high haze for me - but i'm not sure.

Yes I started to use SFS last week - on my Leo triplet, because I wanted to be sure that I was not using poor images which were contributing to this issue. Using that I culled out 10 from my original 50 subs (which was quite severe) then put it through WBPP. It did not make any real difference compared to WBPP on all 50 in relation to the bg. I am not really sure how the sub frame weighting actually works in WBPP - It does not seem to actually reject or not integrate the worst images.

[STEVE333]

David - Are you using dithering with your guiding? Sometimes (especially with DSLR cameras) one can get a "walking noise" when dithering isn't used. Maybe the same could happen with a CMOS camera.

Just a thought.

Steve

[Jockinireland]

Hi Steve, Yes these were dithered at quite a high rate (worked out using your own calculation method on APT forum )

It does not look like walking noise. Its more like "colour noise" but it only becomes apparent after noise reduction.

[STEVE333]

There is a PI tutorial in Light Vortex Astronomy where he deals with color noise in the background. The link below is to the tutorial

https://www.lightvortexastronomy.com/tu ... l#Section4

Section 4. of the tutorial: Reducing background colour and contrast is the portion you might be interested in.

Cheers,

Steve

[Jockinireland]

thanks @STEVE333 I'll take alook at that.

[Juno16]

Hi David and Steve, @Jockinireland @STEVE333

I had some time to look at the red background blotchiness this morning.

Thanks David for pointing me to the Masked Stretch process! I found it to be perfect in this situation (red blotchy background noise). However, the main culprit (for my image) was the fact that I ran EZ Denoise on the linear image (as you had mentioned David!).

I started with the linear image and below (STF stretch). Below are my images before and after EZ Denoise and a Masked Stretch.

No EZ denoise

EZ Denoise with masked stretch

Very blatant example!!!!!


I masked the galaxy (GAME), inverted the mask and increased the RGB/K contrast with an s-curve and the background was quite acceptable (to me).

Still some low level background noise, but nowhere near the blotchy issues. I like this! See zoomed in image below.

I will have to try the masked stretch on old and future images, but I learned two things from this for sure. No EZ Denoise in the linear phase and the Masked Stretch process.

Thanks David and Steve!

[Jockinireland]

Hi @Juno16 Jim, yes thats EXACTLY the outcome I'm getting. Even if i try to do aminor MRT noise reduction at linear i'm still getting it.

I need to look properly into GAME script - I had a quick read but gave up because it just seemed too difficult for me!

[Juno16]

Hi David @Jockinireland ,

That is really good news to me. No more red splotchyness!

Man, I have so much to learn!

You should definitely install the GAME script. It is very intuitive and lets you create multiple masks (ellipses) when there are multiple targets in the image.

Thanks for your troubleshooting on this issue. It has been very helpful to me and much appreciated!