Jump to content

  •  

CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.

Photo

SharpCap green blotches / squares

  • Please log in to reply
11 replies to this topic

#1 Lucian

Lucian

    Vostok 1

  • *****
  • topic starter
  • Posts: 148
  • Joined: 02 Nov 2021
  • Loc: San Francisco CA

Posted 17 January 2025 - 03:40 PM

I don't have much experience in planetary but am trying to slowly add it to my night. Lots of great planets right now!

 

Most of the time (all?) when I try to use Autostakkert to make a stack of relatively few images, I get large green patches. As I add more images, the patches slowly reduce until they are gone. This happens with any target but is more pronounced with lunar. But its pretty annoying because I have to really examine each image closely to see what they look like and then re-try with more images. To be fair, the worst is with less than 10 images or less than about 5%. But sometimes those lower numbers of images show good results. Of course I checked the SER files and no issues there. 

 

My camera is a DSO camera, a ZWO ASI294MC Pro. Its what I have right now and does OK. I also like the larger sensor to help find the planets before reducing the size for shooting. Still working on how to do that more reliably...

 

I have tried using different numbers of APs from 600+ to 260 to 10. All different green patches but all about the same number/size of them. 

 

I have done some searches and don't see anything like what I am seeing. Thanks for any thoughts.

 

Attached are images at 5, 10, and 20%. 

Attached Thumbnails

  • 5 Percent.jpeg
  • 10 Percent.jpeg
  • 20 Percent.jpeg

Edited by Lucian, 18 January 2025 - 01:42 AM.


#2 StargazerLuigi

StargazerLuigi

    Vanguard

  • -----
  • Posts: 2,059
  • Joined: 19 Apr 2014
  • Loc: New London Twp, Pennsylvania, USA

Posted 17 January 2025 - 04:53 PM

If I am understanding this correctly, you are seeing the green patches during capture in SharpCap, but those patches are not present in the SER file. Maybe SharpCap is having a hard time keeping up with automatic debayering? 

 

What kind of frame rates are you getting while capturing? Is your capture PC relatively recent?



#3 Lucian

Lucian

    Vostok 1

  • *****
  • topic starter
  • Posts: 148
  • Joined: 02 Nov 2021
  • Loc: San Francisco CA

Posted 17 January 2025 - 07:12 PM

I feel so bad. I’m pretty new to this and accidentally wrote SharpCap instead of Autostakkert. SharpCap makes nice SER files. When I process them in Autostakkert they come out with these green blotches.

I will update the original post to make clear that mistake.

#4 Borodog

Borodog

    Hubble

  • *****
  • Posts: 12,512
  • Joined: 26 Oct 2020
  • Loc: St. Augustine, FL

Posted 17 January 2025 - 07:34 PM

You are not stacking enough frames.

#5 Lucian

Lucian

    Vostok 1

  • *****
  • topic starter
  • Posts: 148
  • Joined: 02 Nov 2021
  • Loc: San Francisco CA

Posted 18 January 2025 - 01:40 AM

So everyone has this if they stack too few frames? Is there a number that Autostakkert needs? Or a ratio that would be a guideline?



#6 Borodog

Borodog

    Hubble

  • *****
  • Posts: 12,512
  • Joined: 26 Oct 2020
  • Loc: St. Augustine, FL

Posted 18 January 2025 - 09:20 AM

Autostakkert does not explicitly debayer the frames. It debayers by using a 1:1 drizzle. So if you try to stack an extremely small number of frames, there is simply not enough data to fill in the holes. Remember, most of the data in a OSC frame is missing. Only 50% of the green pixels and 25% each of the red and blue pixels are there. If you are going to stack an extremely small number of frames, you will want to use PIPP to preprocess the SER file and debayer it using the HQ Linear method first. But you probably shouldn't do that. This isn't deep sky imaging with exposures so long that the seeing is all averaged out, and every frame is a duplicate of every other frame plus some shift and noise. The exposures are so short that the seeing distorts the frames. We stack a large number of frames to average these effects out. Near critical sampling with an OSC camera we typically want to stack a minimum of hundreds of frames. I would say an absolute minimum of at least a hundred frames, but even that is small.  


  • dcaponeii likes this

#7 dcaponeii

dcaponeii

    Hubble

  • *****
  • Posts: 13,574
  • Joined: 01 Sep 2019
  • Loc: Waxahachie, TX

Posted 18 January 2025 - 10:41 AM

Autostakkert does not explicitly debayer the frames. It debayers by using a 1:1 drizzle. So if you try to stack an extremely small number of frames, there is simply not enough data to fill in the holes. Remember, most of the data in a OSC frame is missing. Only 50% of the green pixels and 25% each of the red and blue pixels are there. If you are going to stack an extremely small number of frames, you will want to use PIPP to preprocess the SER file and debayer it using the HQ Linear method first. But you probably shouldn't do that. This isn't deep sky imaging with exposures so long that the seeing is all averaged out, and every frame is a duplicate of every other frame plus some shift and noise. The exposures are so short that the seeing distorts the frames. We stack a large number of frames to average these effects out. Near critical sampling with an OSC camera we typically want to stack a minimum of hundreds of frames. I would say an absolute minimum of at least a hundred frames, but even that is small.  

I stack with between 100 and 150 frames and don't have any issues so I think Mike's on the right track with his stacking recommendations.  I also see that kind of splotching in the background (not on the disk) if there is a background haze or with a bright background which I think is due to AS!4 having no features to align the AP's with.
 



#8 Lucian

Lucian

    Vostok 1

  • *****
  • topic starter
  • Posts: 148
  • Joined: 02 Nov 2021
  • Loc: San Francisco CA

Posted 19 January 2025 - 11:57 PM

Mike, thanks for the interesting info. Just to be clear, when you say "...we typically want to stack a minimum of hundreds of frames" you are referring to the frames selected for stacking, right? My understanding is that my particular SER file often has 2500 frames over a two minute video and I choose something like 5% in AS to get 125 frames to stack. So, 5% is getting to the bottom limit of that scenario, right?

 

To my eye, the fewer frames stacked the better the output image, assuming no green patches. I thought this was because it was only stacking the few best images. But something I have been realizing is that I can get a lot more images if I push up the gain a bit (shorter exposures) and so that may shift the benefit to more images to stack. I had been running with very low gain settings because it seemed like 100ms was ridiculously short already. I am now realizing I should be going much shorter. 

 

Thank you!



#9 Borodog

Borodog

    Hubble

  • *****
  • Posts: 12,512
  • Joined: 26 Oct 2020
  • Loc: St. Augustine, FL

Posted 20 January 2025 - 08:58 AM

Yes, 125 frames is near the minimum number of frames I would stack with an OSC camera anywhere near critical sampling. If you are under-sampled you can get away with fewer; if you are under-sampled enough a single frame will often look just fine. Mono cameras can get away with fewer frames even at critical sampling because they have no Bayer matrix. 

 

Stacking a very small number of frames from an OSC camera simply does not provide enough data to produce a good image. A reasonable number of frames are required to (a) fill in all of the data that is missing due to the Bayer matrix, (b) increase the SNR of the image, and ( c ) average out the relative motions within the image due to the seeing. Averaging out the relative motions in the image due to seeing will necessarily blur the image compared to a single frame. But that's ok, because we will have increased the SNR enough that it can stand deconvolution/sharpening to unblur it. Without using denoising at the end I would usually want to stack at least a thousand frames. The way that I like to shoot, full disk mosaics with high megapixel cameras, that can be prohibitive, and I end up using denoising to mitigate the limited number of frames in the stack. 

 

You should be looking at exposures that are 10ms, give or take, not 100ms. A 100ms exposure is an eternity at critical sampling, and will virtually never yield a sharp image unless you happen to be taking it from above the Earth's atmosphere. 


Edited by Borodog, 20 January 2025 - 08:59 AM.


#10 Lucian

Lucian

    Vostok 1

  • *****
  • topic starter
  • Posts: 148
  • Joined: 02 Nov 2021
  • Loc: San Francisco CA

Posted 20 January 2025 - 11:14 AM

All good advice. Thanks. I am helping my son's astronomy class with their project to calculate the length of a day on Jupiter by capturing the SER files for them to do some simple math on. I'll use these approaches. His personal project for the class is calculating the rotational speed of Andromeda. Sadly, I cannot help him with that :(



#11 Borodog

Borodog

    Hubble

  • *****
  • Posts: 12,512
  • Joined: 26 Oct 2020
  • Loc: St. Augustine, FL

Posted 20 January 2025 - 11:30 AM

All good advice. Thanks. I am helping my son's astronomy class with their project to calculate the length of a day on Jupiter by capturing the SER files for them to do some simple math on. I'll use these approaches. His personal project for the class is calculating the rotational speed of Andromeda. Sadly, I cannot help him with that frown.gif

 

He will need some pretty advanced equipment for that. 



#12 astrolexi

astrolexi

    Vostok 1

  • -----
  • Posts: 169
  • Joined: 30 Apr 2018
  • Loc: Saulheim, Germany

Posted 24 January 2025 - 02:52 PM

I got similar blotches when I processed my ser files captured with a Altair GPCAM3 224C USB 3.0 camera in Autostakkert!4.

 

Using Autostakkert!3 eliminated the blotches. Maybe a driver thing...

 

Just my two cents.

 

Best wishes and good luck

Klaus




CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.


Recent Topics






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics