Stacked SII image shows wave pattern noise at 1200s exposure

Acquisition Deep Sky M27 Messier Objects Nebula
23 replies•227 views
jujumaster avatar
jujumaster avatar

I stacked 14 SII subs each at 1200s. I took these images in Bortle 8 skies and the moon was down.

In the calibrated and registered folders, I see decent images with random background noise. In the stacked image, I get a clear pattern in the background noise that almost look like waves of dark/light oriented at about 20 degrees from the top of the image. This effect did not seem to show up in the 600s version of the same image. Does anyone know what this noise is and how I remove it?

M27 comparison.png

The attached image shows an example of the calibrated and registered image at the top, and then across the bottom is the stacked image.

thanks in advance for any insights,

Ian

Well written Respectful Engaging
Tony Gondola avatar
Tony Gondola avatar

It looks for all the world like walking noise.

jujumaster avatar
jujumaster avatar

Thanks for the response. I’ll look up walking noise and figure out how to get rid of it. Is it typical that it only shows up in narrow band?

Well written Respectful
Iosif Bodnariu avatar
Iosif Bodnariu avatar

And why did you consider 1200 seconds better than 300 seconds?

Tony Gondola avatar
Tony Gondola avatar

jujumaster · Jun 17, 2026, 03:23 PM

Thanks for the response. I’ll look up walking noise and figure out how to get rid of it. Is it typical that it only shows up in narrow band?

The way you get rid of it is to properly dither your frames. Given the long length of your sub exposures you’ll want to do this for every frame.

jujumaster avatar
jujumaster avatar

Thank you Tony! I appreciate the quick responses here. I’ll post a follow-up just to show how the dithering worked out.

Thanks,

Ian

Well written Respectful Supportive
Tony Gondola avatar
Tony Gondola avatar

jujumaster · Jun 17, 2026, 03:51 PM

Thank you Tony! I appreciate the quick responses here. I’ll post a follow-up just to show how the dithering worked out.

Thanks,

Ian

Great, hope that works out for you!

andrea tasselli avatar
andrea tasselli avatar

jujumaster · Jun 17, 2026, 03:23 PM

Thanks for the response. I’ll look up walking noise and figure out how to get rid of it. Is it typical that it only shows up in narrow band?

It shows more in NB because there is less signal there, more so in SII. I never dither and never had it. This said, I don’t do 1200s integration for this and other good reasons. One way to ameliorate the situation is to carry out cosmetic correction with a 1200s dark frame master. Won’t remove all of it but you’ll get a better result.

Concise
Tony Gondola avatar
Tony Gondola avatar

Also, Cosmic Clarity now has a walking noise removal tool that works pretty well.

TiffsAndAstro avatar
TiffsAndAstro avatar

Consider dropping your sub time to 300s or even lower.

This will have the benefit of allowing more dithered frames.

It definitely looks like walking noise.

jujumaster avatar
jujumaster avatar

I should have mentioned that I am using a FLI 16200 CCD camera. I’m not sure I can get a SII signal at 300s, but I’ll give it a try. I think 300s would do better with CMOS cameras.

I just looked up Cosmic Clarity and will try out there walking noise removal tool. That may help me recover some historical data that I just assumed was bad.

Tony Gondola avatar
Tony Gondola avatar

jujumaster · Jun 17, 2026, 06:24 PM

I should have mentioned that I am using a FLI 16200 CCD camera. I’m not sure I can get a SII signal at 300s, but I’ll give it a try. I think 300s would do better with CMOS cameras.

I just looked up Cosmic Clarity and will try out there walking noise removal tool. That may help me recover some historical data that I just assumed was bad.

For A CCD sensor an exposure time of 1200 sec. would not be unusual, especially in narrowband under dark skies.

Well written Concise
TiffsAndAstro avatar
TiffsAndAstro avatar

Ajujumaster · Jun 17, 2026, 06:24 PM

I should have mentioned that I am using a FLI 16200 CCD camera. I’m not sure I can get a SII signal at 300s, but I’ll give it a try. I think 300s would do better with CMOS cameras.

I just looked up Cosmic Clarity and will try out there walking noise removal tool. That may help me recover some historical data that I just assumed was bad.

Apologies that's a very nice camera, but you're right it needs longer exposures :(

Maybe take new darks frames, if your dark frame library is a little dated

Helpful Respectful Supportive
Lynn K avatar
Lynn K avatar

The Kodak KAF16200 CCD should not have any walking noise. It may be the drivers. I had issues with Starlight Xpress drivers.

I have two mono cameras with the KAF16200 CCD chip (58% QE). The Starlight Xpress SX-46 and Moravian G3-16200 MC MKII. I also have two camras with the Sony IMX571 CMOS chip. The mono QHY268M Pro & OSC ZWO ASI2600MC. The reason I mention that is the CMOS is not all that more sensitive in SII than the CCD.. The larger KAF16200 6nm pixels makes up for that a lot. However, you will loose resolution, depending on focal length and your skies.

As far as SII the KAF16200 is at about 43% EQ and the IMX571 is at 50% QE. The IMX571 is better, but not all that much, and with 3.76nm Pix compared to 6nm, it may be a draw. The SII signal can be very weak in some emission nebula, and it may appear the KAF16200 is insufficient without a direct comparison to a CMOS camera.

Where the CMOS IMX571 has an advantage is in OIII at 90% QE. The CMOS low read noise, I think, may be of little advantage with 300sec or more exposures.

In my opinion, CCD has a bad reputation based on poor QE of Kodak CCD chips. Especially the popular KAF8300. The SONY CCD chips were superior.

Just my opinion based on my experience.

Lynn K

Well written Helpful Insightful Respectful Engaging Supportive
Tony Gondola avatar
Tony Gondola avatar

Lynn K · Jun 17, 2026, 09:22 PM

The Kodak KAF16200 CCD should not have any walking noise. It may be the drivers.

How can that be the case? The way I understand it is, if there is any drift at the pixel level between subs, the fixed pattern noise of the sensor will render as walking noise.

jujumaster avatar
jujumaster avatar

Well, I can definitely see the noise walking when I blink my calibrated but unregistered frames. How in the world does that happen when you are autoguiding the whole time? I use an OAG with a 4s repeating exposure.

TiffsAndAstro avatar
TiffsAndAstro avatar

jujumaster · Jun 17, 2026, 10:42 PM

Well, I can definitely see the noise walking when I blink my calibrated but unregistered frames. How in the world does that happen when you are autoguiding the whole time? I use an OAG with a 4s repeating exposure.

maybe if some pixels have died or gone hot since the darks were taken?

it is a few years old.

Tony Gondola avatar
Tony Gondola avatar

jujumaster · Jun 17, 2026, 10:42 PM

Well, I can definitely see the noise walking when I blink my calibrated but unregistered frames. How in the world does that happen when you are autoguiding the whole time? I use an OAG with a 4s repeating exposure.

Autoguiding is not perfect and frame to frame shifts still happen. That’s why you can see it when you blink your frames.

andrea tasselli avatar
andrea tasselli avatar

Tony Gondola · Jun 17, 2026, 09:38 PM

Lynn K · Jun 17, 2026, 09:22 PM

The Kodak KAF16200 CCD should not have any walking noise. It may be the drivers.

How can that be the case? The way I understand it is, if there is any drift at the pixel level between subs, the fixed pattern noise of the sensor will render as walking noise.

No.

Tony Gondola avatar
Tony Gondola avatar

andrea tasselli · Jun 17, 2026, 11:57 PM

Tony Gondola · Jun 17, 2026, 09:38 PM

Lynn K · Jun 17, 2026, 09:22 PM

The Kodak KAF16200 CCD should not have any walking noise. It may be the drivers.

How can that be the case? The way I understand it is, if there is any drift at the pixel level between subs, the fixed pattern noise of the sensor will render as walking noise.

No.

and?

Lynn K avatar
Lynn K avatar

Tony Gondola · Jun 17, 2026, 09:38 PM

Lynn K · Jun 17, 2026, 09:22 PM

The Kodak KAF16200 CCD should not have any walking noise. It may be the drivers.

How can that be the case? The way I understand it is, if there is any drift at the pixel level between subs, the fixed pattern noise of the sensor will render as walking noise.

Hi Tony. I have used CCD cameras from 2005 till now, and it is my ubderstanding CCD camera do not have fixed pattern noise. Fix pattern noise in CMOS cameras are cause by the indivigual Pixel Analog to Digital converters. Google A.I explaines it better than I can. Pattern Noise is caused by “patial, time-constant non-uniformities across an image. Unlike random noise, it appears exactly the same way in every frame. It is caused by hardware inconsistencies rather than random quantum fluctuations. “ In this case the hardware referred to is the Analog to Digital converters.

CCD has no such pattern because has only has ONE Analog to Digital coverter. The the chip data is downloaded roll by roll, then converted to digital, rather than pixel by pixel in CMOS.

Which is why the downloads can be slower and not good for video.

Again, Google A.I. explaines it better. “ In traditional CCDs, a single amplifier reads the entire sensor. CMOS sensors, however, feature active amplifiers for every single pixel or column of pixels. Because no two micro-amplifiers are 100% identical, these slight gain and offset variations create a visible, constant grid-like or vertical stripe pattern in the final image. “

Lynn K.

Tony Gondola avatar
Tony Gondola avatar

That’s interesting. So no CCD camera can exhibit walking noise? That would imply that even though it’s a singe amplifying circuit, there are absolutely no differences pixel to pixel. I’m not questioning you or your experience and I’ve never used a CCD, it’s just an interesting data point if that’s the case.

Well written Respectful Engaging
Lynn K avatar
Lynn K avatar

Hi Tony. I'm not implying that CCD chips or perfect with no pixel variation. Wouldn't that be nice, no more bias or darks ever needed. CCD pixels vary as muvh as CMOS. All chips, CCD or CMOS will have pixel vaiation. That is the Bias. The pixels are bias. Some are better at registering photons into electron and others are worst. Variation is Bias may be fairly consistent, but I not sure. But adding Dark current makes It random and will not repeat itself identically the same in every sub. The dark current also varies. That is true of both CCD and CMOS.

Where the CMOS design varies is in having indivigual amplifiers (analog to digital converter) on each pixel as oppose to a single Amplifier/Converter in the CCD design. These CMOS amplifiers/converters are not random. Thay are very consistant and will create an almost idenical patter in each sub. So, it is necessary to dither so the same pixels do not keep aligning up. Probably tracking inperfection keep them form perfectly aligning, but they do it well enough to creat a pattern.

It not the incosistancy of pixels that creat pattern noise. It's the consistency of the CMOS pixel amplifiers that create a pattern. And that pattern will be repeated in every sub.

Dithering also has benifits in CCD image acquisition. But unlike CMOS, its primary goal is not to eliminate patter noise. In CCD and CMOS dithering stops brighter and darker pixel from consistantly aligning up during stacking. By pixels not alway aligning up on themselves, pixel variations are averaged out. With CMOS there is the added benifit of eliminating the pixel amplifier pattern noise.

I hope my discription is not too confusing.

Lynn K.

Tony Gondola avatar
Tony Gondola avatar

Lynn K · Jun 18, 2026, 02:46 AM

Hi Tony. I'm not implying that CCD chips or perfect with no pixel variation. Wouldn't that be nice, no more bias or darksever needed. CCD pixels vary as muvh as CMOS. All chips, CCD or CMOS will have pixel vaiation. That is the Bias. The pixels are bias. Some are better at registering photons into electron and others are worst. Variation is Bias may be fairly consistent, but I not sure But adding Dark current makes It random and will not repeat itself identically the same in every sub. The dark current also varies. That is true of both CCD and CMOS.

Where the CMOS design varies is in having indivigual amplifiers (analog to digital converter) on each pixel as oppose to a single Amplifier/Converter in the CCD design. These CMOS amplifiers/converters are not random. Thay are very consistant and will create an almost idenical patter in each sub. So, it is necessary to dither so the same pixels do not keep aligning up. Probably tracking inperfection keep them form perfectly aligning, but they do it well enough to creat a pattern.

It not the incosistancy of pixels that creat pattern noise. It's the consistency of the CMOS pixel amplifiers that create a pattern. And that pattern will be repeated in every sub.

Dithering also has benifits in CCD image acquisition. But unlike CMOS, its primary goal is not to eliminate patter noise. In CCD and CMOS dithering stops brighter and darker pixel from consistantly aligning up during stacking. By pixels not alway aligning up on themselves, pixel variations are averaged out.

I hope my discription is not too confusing.

Lynn K.

Not at all, makes perfect sense. Thanks for getting into the details.

Respectful Supportive