Luminance Noise Pattern / WBPP Errors - Forums

Topic starter

Jul 17, 2025

The past few months I've been noticing a pattern in my luminance data that I'm hoping the forum here might be able to help me diagnose. The pattern looks to me like walking noise, however I've been using the same dithering settings for over a year with no issues until recently and so I'm doubting that walking noise is the culprit. This pattern also doesn't show up in any of my R,G,B or narrowband data.

The examples I have here are master lights from 600s sub frames. My scope is a Wiliam Optics FLT91 and the camera is a 2600mm. I'm dithering at 5 pixels every 5 frames. I'm shooting from the dark skies at Starfront Observatories and the data was all taken on moonless nights. You'll have to ignore the pinched optics from my telescope as that's another issue entirely!

I've processed the data using both PixInsight and APP but have had the same results with both. I've also tried to process with just flats and bias (and no darks) but the pattern still shows. When I'm running WBPP in PixInsight I'm also getting the following error message when the image is going through the astrometric solution process:

I'm at a loss for what this could be and I'm open to any/all suggestions. Thanks!

Well Written Engaging

Jerry Gerber

Frequent contributor

Jul 17, 2025 · Edited: Jul 17, 2025

Hi Jack,

I'm not seeing anything but the normal noise that occurs in any integrated master light before processing with Noise-X.

I must be misunderstanding the issue you're having.

Are you doing anything different lately from how you've been pre-processing or processing before the problem you're having occurred?

Jerry

Well Written Respectful Concise

Tony Gondola

Frequent contributor

Jul 17, 2025

These are very aggressive stretches, especially the first one. I agree with Jerry, things get pretty messy when you dig that deep. Is this something that's suddenly changed and wasn't there before?

Well Written

andrea tasselli

Frequent contributor

Jul 17, 2025

What is your image scale? I guess it's too high for the telescope you're using. That is just shot noise.

danieldh206

Jul 17, 2025

With 600-second subs, suggest you dither after every image. Dithering is one of the most effective ways to reduce noise in a stacked image.

Well Written Concise

Jack Groves

Topic starter

Jul 17, 2025

andrea tasselli:
What is your image scale? I guess it's too high for the telescope you're using. That is just shot noise.

*** Hi Andrea, I’m at 1.44” so it should be well sampled. I see your train of thought but it doesn’t explain why my prior images were fine but this issue is now happening . ***

andrea tasselli

Frequent contributor

Jul 17, 2025

Just to get a grasp of the magnitude of the issue are those images stacks of how many seconds/minutes/hours? As a means of comparison I'm shooting with a 6" at 1.63"/px and still get shot noise even with 10 hours' worth of imaging time.

Engaging

andrea tasselli

Frequent contributor

Jul 17, 2025

I do not mean to be pessimistic (and crossing fingers it won't be anything like this) but the only time I had similar issues (increased noise/fixed pattern developing when there was none before) is before my IMX533 based camera died on me.

Jack Groves

Topic starter

Jul 17, 2025

andrea tasselli:
Just to get a grasp of the magnitude of the issue are those images stacks of how many seconds/minutes/hours? As a means of comparison I'm shooting with a 6" at 1.63"/px and still get shot noise even with 10 hours' worth of imaging time.

The shot with the M58 crop is 41 hours of integration, and NGC 1333 is about 8 hours iirc. There’s definitely still some shot noise, but my main issue is the fixed pattern noise that I’m seeing.

also - re your point above about the camera dying, I’m hoping that’s not the case with my IMX571! I’ll add that the cooler wasn’t running particularly hard when I imaged the data

andrea tasselli

Frequent contributor

Jul 17, 2025

I'd check your shooting parameters, signally gain settings. If RANSAC can't find a fix is because the image is exceedingly shallow as when you run at very high gain settings. This said with 40 hours of integration from a prime dark site you shouldn't see any of the fixed pattern noise but to be sure I'd have to look at a single raw shot.

Adam Block

Jul 18, 2025

How old are your dark frames? More than 3 months?

-adam

Jack Groves

Topic starter

Jul 18, 2025

Adam Block:
How old are your dark frames? More than 3 months?

-adam

Hi Adam - the darks were just under a month old when I captured the data.

Well Written Respectful

Adam Block

Jul 18, 2025

Jack Groves:
Adam Block:
How old are your dark frames? More than 3 months?

-adam

Hi Adam - the darks were just under a month old when I captured the data.

This effect tends to happen when there is a small mismatch between darks and lights (for whatever reason). Warm/Hot pixels do not cleanly subtract and you have "correlated noise" fixed pattern effect. This is been in issue in the past for some remote astro-data companies. It is hard to keep up with updating calibration data for the best fit of the current sensor characteristics- especially when running a pipeline.

Please see the attached picture... in addition to giving you my opinion I will also show an example from my experience and suggest a solution.- hope you agree this is similar. The red lines indicate the artifact which filled the masters (and more easily seen in the noisiest areas). It is quite subtle. What fixed the issue was new contemporaneous dark frames.

Helpful Insightful Respectful Engaging

Jack Groves

Topic starter

Jul 18, 2025

Adam Block:
Jack Groves:
Adam Block:
How old are your dark frames? More than 3 months?

-adam

Hi Adam - the darks were just under a month old when I captured the data.

This effect tends to happen when there is a small mismatch between darks and lights (for whatever reason). Warm/Hot pixels do not cleanly subtract and you have "correlated noise" fixed pattern effect. This is been in issue in the past for some remote astro-data companies. It is hard to keep up with updating calibration data for the best fit of the current sensor characteristics- especially when running a pipeline.

Please see the attached picture... in addition to giving you my opinion I will also show an example from my experience and suggest a solution.- hope you agree this is similar. The red lines indicate the artifact which filled the masters (and more easily seen in the noisiest areas). It is quite subtle. What fixed the issue was new contemporaneous dark frames.

Hi Adam - this is absolutely the type of noise pattern I’m seeing! I’ll attempt to calibrate with a new set of matching darks and see if there is any improvement.

As far as the RANSAC error goes is that something that could be related or it is likely another issue altogether?

Well Written Respectful

Adam Block

Jul 19, 2025

Different issue entirely.
-adam