Red Glow after WBPP Stacking

[mventi1]

Hi,
I am using the following configuration:
Scope – Celestron Edge 8HD
Image train – MC533 Pro, Optilong L-Pro and Celestron .7 reducer

I keep getting this red glow within the image stacking using PixInSight. I have done this with several sets of shooting different objects. I am using the WBPP within PIX and I just started seeing this about 3 weeks ago. When I stack with ASI Deepstack or SIRIL I do not get this red glow. I have tried stacking in PIX with dark, flat, darkflat and lights. I have tried plane lights and dark and lights and the same result. Any ideas please.

I have attached screenshots from each.

 

[cloudbait]

Hi,
I am using the following configuration:
Scope – Celestron Edge 8HD
Image train – MC533 Pro, Optilong L-Pro and Celestron .7 reducer

I keep getting this red glow within the image stacking using PixInSight. I have done this with several sets of shooting different objects. I am using the WBPP within PIX and I just started seeing this about 3 weeks ago. When I stack with ASI Deepstack or SIRIL I do not get this red glow. I have tried stacking in PIX with dark, flat, darkflat and lights. I have tried plane lights and dark and lights and the same result. Any ideas please.

I have attached screenshots from each.

No attachments. But screenshots of images are usually not helpful. What we need are XISF files, uploaded to a cloud service with shared links provided here. Your master frames (lights and calibration) out of WBPP would be a good start. Maybe your full WBPP log, as well.

 

[mventi1]

No attachments. But screenshots of images are usually not helpful. What we need are XISF files, uploaded to a cloud service with shared links provided here. Your master frames (lights and calibration) out of WBPP would be a good start. Maybe your full WBPP log, as well.

CB,
Thanks for your prompt response. Can you guide me how to do that? I am new here and fairly new with PI.
Thanks,

 

[cloudbait]

CB,
Thanks for your prompt response. Can you guide me how to do that? I am new here and fairly new with PI.
Thanks,

You just need to upload the images to a cloud service (Google Drive, Dropbox, Send Anywhere, etc). All those services let you flag files as public and will provide a shareable link to them, which you can post here.

 

[mventi1]

You just need to upload the images to a cloud service (Google Drive, Dropbox, Send Anywhere, etc). All those services let you flag files as public and will provide a shareable link to them, which you can post here.

Copy, thanks CB will do that and reach back out.

 

[mventi1]

CB,
I uploaded the light FITS files and the calibration frames I used along with the PIX output files and logs. Hope this is what you needed.

PixInSight - Google Drive

drive.google.com

Thanks,
Mike

 

[cloudbait]

CB,
I uploaded the light FITS files and the calibration frames I used along with the PIX output files and logs. Hope this is what you needed.

PixInSight - Google Drive

drive.google.com

Thanks,
Mike

That's not the log file. It will be in the log folder in your WBPP output path. That said, just look at your red master. It's completely messed up. Without more information, I'd guess that it is the result of badly failed alignment, and most likely because there are so many hot pixels. Are you using CosmeticCorrection in WBPP?

 

[tom886]

So, I was using your files and run them in WBPP.
Couldn't duplicate your result.

I used Cosmetic Correction and set the Pedestal to Automatic.
Above is the master and only auto stretched.

Edit: I attached my log file, maybe you can compare with yours and see what's different to your configuration

Cheers
Tom

 

[mventi1]

That's not the log file. It will be in the log folder in your WBPP output path. That said, just look at your red master. It's completely messed up. Without more information, I'd guess that it is the result of badly failed alignment, and most likely because there are so many hot pixels. Are you using CosmeticCorrection in WBPP?

CB,
My alignment seems spot on. My other software is looking good in the stacking. For the CosCor, no I am not. I have seen tutorials on it but am still learning what it is. The logs I uploaded and shared are from the area you mentioned.
Mike

 

[mventi1]

So, I was using your files and run them in WBPP.
Couldn't duplicate your result.
View attachment 22944
I used Cosmetic Correction and set the Pedestal to Automatic.
Above is the master and only auto stretched.

Edit: I attached my log file, maybe you can compare with yours and see what's different to your configuration

Cheers
Tom

Tom,
I did not use CosCor but will take your advice and settings and see what happens. As mentioned in original post this just surfaced as an issue. It was not there in previous runs. Also, can you please provide more detail on what the settings were for CosCor and where the Pedestal settings are to set "Automatic".
Thanks,
Mike

 

[cloudbait]

CB,
My alignment seems spot on. My other software is looking good in the stacking. For the CosCor, no I am not. I have seen tutorials on it but am still learning what it is. The logs I uploaded and shared are from the area you mentioned.
Mike

Based on the completely broken red channel, and Tom's experience, I'd say that the failure to use CosmeticCorrection is the cause of the problem you're seeing.

 

[mventi1]

That's not the log file. It will be in the log folder in your WBPP output path. That said, just look at your red master. It's completely messed up. Without more information, I'd guess that it is the result of badly failed alignment, and most likely because there are so many hot pixels. Are you using CosmeticCorrection in WBPP?

CB,
I am confused then. I

Based on the completely broken red channel, and Tom's experience, I'd say that the failure to use CosmeticCorrection is the cause of the problem you're seeing.

Tom/CB,
Thank you both for the input. I will retry with Tom's suggestion. I really appreciate the inputs.
Mike

 

[cloudbait]

CB,
I am confused then. I

Tom/CB,
Thank you both for the input. I will retry with Tom's suggestion. I really appreciate the inputs.
Mike

I only saw what you had in the log files folder (which is a process icon script). What's in the root is fine. And looking at it, the multiple RANSAC failures also point to uncorrected hot pixels.

 

[mventi1]

I only saw what you had in the log files folder (which is a process icon script). What's in the root is fine. And looking at it, the multiple RANSAC failures also point to uncorrected hot pixels.

CB,
Can you please elaborate more and educate me further on what that means? I have seen the red pixels and from my understanding the darks should correct for this. What else am I missing?
Thanks,
Mike

 

[tom886]

What else am I missing?

The Cosmetic Correction.

Cheers
Tom

 

[cloudbait]

CB,
Can you please elaborate more and educate me further on what that means? I have seen the red pixels and from my understanding the darks should correct for this. What else am I missing?
Thanks,
Mike

Darks mitigate hot pixels, but do not always completely remove them. In most cases (and certainly yours) CosmeticCorrection needs to be part of the preprocessing workflow. Because you have a lot of uncorrected hot pixels, your red frames did not align properly, so the integrated master is a muddled mess (which is where the red "glow" in your channel combined master is coming from).

UPDATE: Had a little more time this morning, so I downloaded your full set and ran it through WBPP. With all the default settings, and no CC, I got the same sort of muddled alignment mess you did. Simply adding CC with the common hot sigma=3 setting resulted in a perfect result, all three channels aligned and integrated into a very nice image. Below, with and without CC, no other differences.

 

[fredvanner]

I get the same results - perfect alignment if CC is used. I note that your darks are about two and a half years old; I suggest that you capture some new darks.

 

[cloudbait]

I get the same results - perfect alignment if CC is used. I note that your darks are about two and a half years old; I suggest that you capture some new darks.

I'm not sure that this sensor normally has much of a problem with hot pixels. The darks here are interesting- both the very short ones and the long ones have virtually identical backgrounds, suggesting there's a large offset being added somewhere and almost no dark current. But the long ones are loaded with hot pixels. Maybe normal for this sensor, maybe not. And indeed, it's possible that a new set of darks might provide better correction and avoid this registration problem. I'd still use CC, though. No reason not to.

 

[mventi1]

I'm not sure that this sensor normally has much of a problem with hot pixels. The darks here are interesting- both the very short ones and the long ones have virtually identical backgrounds, suggesting there's a large offset being added somewhere and almost no dark current. But the long ones are loaded with hot pixels. Maybe normal for this sensor, maybe not. And indeed, it's possible that a new set of darks might provide better correction and avoid this registration problem. I'd still use CC, though. No reason not to.

CB - There are the normal same time darks as lights then the dark_flat that is the shorter exposure.
Fred - Greetings and thanks for joining in. Do darks really have a life span on them? I was not aware of that.
Mike

 

[cloudbait]

CB - There are the normal same time darks as lights then the dark_flat that is the shorter exposure.
Fred - Greetings and thanks for joining in. Do darks really have a life span on them? I was not aware of that.
Mike

There's no such thing as a "dark_flat". You have darks here with two exposure times, long and short. The equal backgrounds suggest there's almost no dark current from your sensor (which is good). But the growth of the hot pixels with the long exposure is interesting, and may or may not be normal for that sensor.

The stability of darks with time depends entirely on the sensor. With some cameras they may be stable for years, with some even a few months might be pushing it. You can take some new ones and compare them with the old ones to see how stable your sensor actually is.