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

Fixed Pattern Noise getting worse and worse

  • Please log in to reply
144 replies to this topic

#1 joelin

joelin

    Surveyor 1

  • *****
  • topic starter
  • Posts: 1653
  • Joined: 14 Jan 2008
  • Loc: Saratoga, CA

Posted 12 December 2018 - 02:33 AM

Okay so I started my astroimaging journey with a Nikon D600, 180mm f/2.8 lens stopped down to f/5.6 and imaged the Pleiades with a cheap iOptron Skytracker (this is obviously unguided). I got a pretty good result. Later on I tried the Sony A7S, AT60ED f/6 and Orion Sirius. I started seeing some ugly noise because it was unguided and I think fixed pattern noise started to show. More recently, I used a cooled ASI1600MC, Hyperstar on a C8 and Orion Sirius guided but no dithering. The fixed pattern noise became horrendous.

 

See below for a comparison of the same area between the Nikon D600 unguided (left), Sony A7S unguided (center), ASI1600MC guided (right).

 

In all three cases I stacked using Deep Sky Stacker and did a screen transfer function stretch in Pixinsight, nothing else. 

 

Why did the fixed pattern noise go from non-existent to somewhat annoying to extremely horrendous?!? In all cases I had at least 30 minutes of exposure in fairly dark skies. In all 3 cases I did not dither. First two were unguided. Third was guided but no dither (ASIAir doesn't support this yet)

 

dVHTwSE.jpg


Edited by joelin, 12 December 2018 - 02:35 AM.


#2 einarin

einarin

    Surveyor 1

  • *****
  • Posts: 1596
  • Joined: 23 Dec 2016

Posted 12 December 2018 - 03:32 AM

Well, you should dither and use temperature matched darks.


  • Walleye likes this

#3 joelin

joelin

    Surveyor 1

  • *****
  • topic starter
  • Posts: 1653
  • Joined: 14 Jan 2008
  • Loc: Saratoga, CA

Posted 12 December 2018 - 03:34 AM

Well, you should dither and use temperature matched darks.

well i know that

 

but can you explain why my older setup with the D600 (which did not use temperature matched anything) has a better result than the ASI1600MC which does have temperature matched darks



#4 einarin

einarin

    Surveyor 1

  • *****
  • Posts: 1596
  • Joined: 23 Dec 2016

Posted 12 December 2018 - 03:50 AM

Well, I'm guessing here but maybe the poorer tracking helped with D600 image (poor mans dithering).



#5 joelin

joelin

    Surveyor 1

  • *****
  • topic starter
  • Posts: 1653
  • Joined: 14 Jan 2008
  • Loc: Saratoga, CA

Posted 12 December 2018 - 04:01 AM

usually the poor tracking will result in more FPN right? after all it will result in some star drift in a specific direction which is not random....

 

I used the same unguided mount for the A7S and the ASi1600 so why was one of them much better noise wise. 



#6 happylimpet

happylimpet

    Gemini

  • *****
  • Posts: 3434
  • Joined: 29 Sep 2013
  • Loc: Southampton, UK

Posted 12 December 2018 - 04:43 AM

Not that this directly answers your question, but i used to get very similar looking FPN with my ASI1600MM, which I reduced greatly recently by doing kappa-sigma stacking on both my darks and lights.

 

Previously I was using median stacking for my darks, and that left far worse walking noise - my images looked like yours on the right.



#7 ChrisWhite

ChrisWhite

    Skylab

  • *****
  • Posts: 4002
  • Joined: 28 Feb 2015
  • Loc: Colchester, VT

Posted 12 December 2018 - 06:26 AM

The 1600 is a very low noise camera, and you generally use shorter exposures so the fixed pattern noise shows up more clearly than your other higher noise cameras where these read noise patterns are buried beneath thermal noise and longer exposure sky signals.

You MUST dither frequently and aggressively to avoid this kind of noise with these cooled cmos Astro cameras.

#8 entilza

entilza

    Soyuz

  • *****
  • Posts: 3619
  • Joined: 06 Oct 2014
  • Loc: Oakville, ON, Canada

Posted 12 December 2018 - 10:35 AM

Hey Joel. To help understand try this.

- Use the blink tool.
- Load your light frames.
- Cycle your images in sequence. (Make sure its sequential ie 0001->000N - Another topic lol)
- Note the drift direction

Do this for your old images and new images, compare the drift in each image.

How are you polar aligning?

Dither will solve this for you

Edited by entilza, 12 December 2018 - 10:36 AM.


#9 entilza

entilza

    Soyuz

  • *****
  • Posts: 3619
  • Joined: 06 Oct 2014
  • Loc: Oakville, ON, Canada

Posted 12 December 2018 - 10:54 AM

To try to give you some more tips now,

Since your ASI air can't dither you need to get a very good polar alignment.

ZWO is making a polar align tool I believe for the ASI Air... So you have to wait lol.

If you have a laptop, you can use sharpcap to polar align before that would be what I recommend.

Finally and this is extreme and will test your patience, but you may not have a choice...Manually dither... move your frame randomly every few minutes. Just slightly (an average star size distance)

Edited by entilza, 12 December 2018 - 10:54 AM.


#10 joelin

joelin

    Surveyor 1

  • *****
  • topic starter
  • Posts: 1653
  • Joined: 14 Jan 2008
  • Loc: Saratoga, CA

Posted 12 December 2018 - 01:12 PM

The 1600 is a very low noise camera, and you generally use shorter exposures so the fixed pattern noise shows up more clearly than your other higher noise cameras where these read noise patterns are buried beneath thermal noise and longer exposure sky signals.

You MUST dither frequently and aggressively to avoid this kind of noise with these cooled cmos Astro cameras.

actually with the A7S and D600, I was also doing 30 second exposures (the ASI1600 was at 60 seconds) at a high enough ISO for the skyfog to clear the left edge of the histogram...I'm assuming the DSLRs have more noise than the ASI1600..and the D600 has more noise than the A7S...yet with the D600 I saw the least amount of pattern noise...

 

so the most noisy cameras have the least amount of pattern noise?? so how do you explain that?

 

and for the less noisy ASI1600 cameras I did longer exposures than the more noisy cameras with shorter exposures... and now the result is more fixed pattern noise on the ASI1600 ? i can't explain that

 

also, for all 3 samples I showed.. I did not dither


Edited by joelin, 12 December 2018 - 01:14 PM.


#11 joelin

joelin

    Surveyor 1

  • *****
  • topic starter
  • Posts: 1653
  • Joined: 14 Jan 2008
  • Loc: Saratoga, CA

Posted 12 December 2018 - 01:15 PM

To try to give you some more tips now,

Since your ASI air can't dither you need to get a very good polar alignment.

ZWO is making a polar align tool I believe for the ASI Air... So you have to wait lol.

If you have a laptop, you can use sharpcap to polar align before that would be what I recommend.

Finally and this is extreme and will test your patience, but you may not have a choice...Manually dither... move your frame randomly every few minutes. Just slightly (an average star size distance)

i know I need to dither..that is what everyone says.....but my actual question is why even noisier cameras had less pattern noise without dithering...in fact the overall result of the noisier camera without dithering just looks better than the cooled ASI1600


Edited by joelin, 12 December 2018 - 01:15 PM.


#12 joelin

joelin

    Surveyor 1

  • *****
  • topic starter
  • Posts: 1653
  • Joined: 14 Jan 2008
  • Loc: Saratoga, CA

Posted 12 December 2018 - 01:17 PM

Hey Joel. To help understand try this.

- Use the blink tool.
- Load your light frames.
- Cycle your images in sequence. (Make sure its sequential ie 0001->000N - Another topic lol)
- Note the drift direction

Do this for your old images and new images, compare the drift in each image.

How are you polar aligning?

Dither will solve this for you

I'm hoping there won't be any drift since I was guiding with the ASIAir the entire time...the ASIAir doesn't dither yet but the guiding should prevent any different...

 

I will test this out tonight. 



#13 entilza

entilza

    Soyuz

  • *****
  • Posts: 3619
  • Joined: 06 Oct 2014
  • Loc: Oakville, ON, Canada

Posted 12 December 2018 - 01:22 PM

You can drift even when guiding.. That's what's important to realize and the blink tool will show you this.

#14 Peter in Reno

Peter in Reno

    Voyager 1

  • *****
  • Posts: 10158
  • Joined: 15 Jul 2008
  • Loc: Reno, NV

Posted 12 December 2018 - 03:23 PM

Personally I find ASI1600 image much better than the first two images. The first two images show badly bloated stars and ASI1600 image is much sharper.

 

I assume the first two images were taken with unmodified DSLR cameras and they typically contain several filters in front of the sensors to purposely soften and blur the images (to help make the images look better for daytime photography) which causes big star bloats and mask out some noise. There are no filters in front of ASI1600 sensor so it looks much sharper.

 

It's not a good way to compare unguided images with guided images. Unguided images can also make the stars to bloat just as big and mask out the "noise" you are describing.

 

Personally I don't think the noise is FPN. What you are seeing is what some people calls it "mottling" due to deBayer process when converting from mono to color from OSC camera. If this was taken with mono camera with filters, you should see much less mottling since there's no deBayering process.

 

There's no such a thing as perfect guiding and not get any drifts. There will always be drifts during long exposure imaging due to mis-polar alignment, differential flexure, mirror flops, etc.

 

Like others said, dither aggressively and your problem will be mostly solved.

 

I think you are trying to find the least path of resistance at the lowest possible cost in this hobby and nothing wrong with that but it can be difficult.

 

Peter


Edited by Peter in Reno, 12 December 2018 - 03:27 PM.

  • rgsalinger likes this

#15 fmeschia

fmeschia

    Viking 1

  • *****
  • Posts: 792
  • Joined: 20 May 2016
  • Loc: Mountain View, CA

Posted 12 December 2018 - 04:26 PM

The ASI1600 image has a lot of walking noise. How was the frame calibrated?

Francesco



#16 Jon Rista

Jon Rista

    ISS

  • *****
  • Posts: 22320
  • Joined: 10 Jan 2014
  • Loc: Colorado

Posted 12 December 2018 - 05:32 PM

If you cannot dither yet with ASIAir, then the best solution would be to dial in your polar alignment as best as you possibly can. And you would really need it to be very good to eliminate the walking noise you see in the ASI1600 image. Also note that with guiding, you can experience differential flexure, so even once you do dial in your polar alignment, diff flex may still result in some uncorrected drift that will still give rise to the walking noise.

 

To combat the remaining walking noise...you may need to use longer exposures, to get more of the signal above the noise floor and bury the pattern. The more deeply you can bury the pattern, the less it will matter. You will also want to create the best dark and flat masters you can to ideally model the FPN and remove it. Both darks and flats are necessary to fully correct FPN, as FPN comes from both dark signal and light signal. So make sure you are calibrating with both.

 

I agree with Peter, your ASI1600 image shows the potential of the CMOS camera and its small pixels...it is a much sharper image. I do think your subs are probably too shallow, though, which is why the DFPN is being more of an issue.


Edited by Jon Rista, 12 December 2018 - 05:33 PM.


#17 Peter in Reno

Peter in Reno

    Voyager 1

  • *****
  • Posts: 10158
  • Joined: 15 Jul 2008
  • Loc: Reno, NV

Posted 12 December 2018 - 05:36 PM

The ASI1600 image has a lot of walking noise. How was the frame calibrated?

Francesco

Walking Noise is not caused by camera alone but caused by slight drift from frame to frame due to mis-polar alignment or differential flexure. I've never heard of any camera itself to cause "walking noise".

 

If dithering was more aggressive, then walking noise would be either reduced or eliminated.

 

Peter



#18 fmeschia

fmeschia

    Viking 1

  • *****
  • Posts: 792
  • Joined: 20 May 2016
  • Loc: Mountain View, CA

Posted 12 December 2018 - 05:39 PM

Walking Noise is not caused by camera alone but caused by slight drift from frame to frame due to mis-polar alignment or differential flexure. I've never heard of any camera itself to cause "walking noise".

 

If dithering was more aggressive, then walking noise would be either reduced or eliminated.

 

Peter

Never said it was caused by camera. Walking noise is the result of residual FPN plus spatial correlation across frames (which could be due to drifting as you said). As the OP said, he can’t do dithering, so spatial correlation is unavoidable for him. The remaining option would be to reduce residual FPN with calibration.


Edited by fmeschia, 12 December 2018 - 05:40 PM.

  • jdupton likes this

#19 Peter in Reno

Peter in Reno

    Voyager 1

  • *****
  • Posts: 10158
  • Joined: 15 Jul 2008
  • Loc: Reno, NV

Posted 12 December 2018 - 05:45 PM

Never said it was caused by camera. Walking noise is the result of residual FPN plus spatial correlation across frames (which could be due to drifting as you said). As the OP said, he can’t do dithering, so spatial correlation is unavoidable for him. The remaining option would be to reduce residual FPN with calibration.

OK, I thought you meant the camera. Going back to your previous post, you said "image" when I thought you said "camera" so I got confused.

 

Peter


Edited by Peter in Reno, 12 December 2018 - 05:45 PM.


#20 happylimpet

happylimpet

    Gemini

  • *****
  • Posts: 3434
  • Joined: 29 Sep 2013
  • Loc: Southampton, UK

Posted 12 December 2018 - 08:06 PM

Might also be worth playing around with your stacking rejection parameters. I used to just do average stacks, and got bad walking noise sometimes, but switching to kappa-sigma has almost eliminated it. As we speak im going back and re-processing some images with bad walking noise. These are ASI1600MM images by the way.

 

comp.jpg


  • tjones84 likes this

#21 joelin

joelin

    Surveyor 1

  • *****
  • topic starter
  • Posts: 1653
  • Joined: 14 Jan 2008
  • Loc: Saratoga, CA

Posted 13 December 2018 - 02:11 AM

Personally I find ASI1600 image much better than the first two images. The first two images show badly bloated stars and ASI1600 image is much sharper.

 

I assume the first two images were taken with unmodified DSLR cameras and they typically contain several filters in front of the sensors to purposely soften and blur the images (to help make the images look better for daytime photography) which causes big star bloats and mask out some noise. There are no filters in front of ASI1600 sensor so it looks much sharper.

 

It's not a good way to compare unguided images with guided images. Unguided images can also make the stars to bloat just as big and mask out the "noise" you are describing.

 

Personally I don't think the noise is FPN. What you are seeing is what some people calls it "mottling" due to deBayer process when converting from mono to color from OSC camera. If this was taken with mono camera with filters, you should see much less mottling since there's no deBayering process.

 

There's no such a thing as perfect guiding and not get any drifts. There will always be drifts during long exposure imaging due to mis-polar alignment, differential flexure, mirror flops, etc.

 

Like others said, dither aggressively and your problem will be mostly solved.

 

I think you are trying to find the least path of resistance at the lowest possible cost in this hobby and nothing wrong with that but it can be difficult.

 

Peter

the stars are bloated because of different image scales...the first image was at a 180mm focal lenght, the second was at 360mm focal length but with large pixels, the last image was at 425mm focal length with small pixels...that would cause the star sizes to change



#22 joelin

joelin

    Surveyor 1

  • *****
  • topic starter
  • Posts: 1653
  • Joined: 14 Jan 2008
  • Loc: Saratoga, CA

Posted 13 December 2018 - 02:13 AM

Walking Noise is not caused by camera alone but caused by slight drift from frame to frame due to mis-polar alignment or differential flexure. I've never heard of any camera itself to cause "walking noise".

 

If dithering was more aggressive, then walking noise would be either reduced or eliminated.

 

Peter

im guiding with the ASI1600 so there shouldn't be much drift between frames correct? any drift would be compensated as the guider would see it and push the star back....



#23 joelin

joelin

    Surveyor 1

  • *****
  • topic starter
  • Posts: 1653
  • Joined: 14 Jan 2008
  • Loc: Saratoga, CA

Posted 13 December 2018 - 02:14 AM

Never said it was caused by camera. Walking noise is the result of residual FPN plus spatial correlation across frames (which could be due to drifting as you said). As the OP said, he can’t do dithering, so spatial correlation is unavoidable for him. The remaining option would be to reduce residual FPN with calibration.

if im guiding then there shouldn't be any drift right? i do calibrate with darks, bias and flats...since this is a cooled CMOS I can get the exact temperatures...I do about 30 of each calibration frame using the default settings in DSS



#24 fmeschia

fmeschia

    Viking 1

  • *****
  • Posts: 792
  • Joined: 20 May 2016
  • Loc: Mountain View, CA

Posted 13 December 2018 - 02:15 AM

if im guiding then there shouldn't be any drift right? i do calibrate with darks, bias and flats...since this is a cooled CMOS I can get the exact temperatures...I do about 30 of each calibration frame using the default settings in DSS

Well you have the original light frames, it should be easy for you to compare the first and the last one to see if there’s any drift.



#25 pfile

pfile

    Aurora

  • -----
  • Posts: 4777
  • Joined: 14 Jun 2009

Posted 13 December 2018 - 02:30 AM

if you are using a separate guidescope then drift is possible - differential flexure. if you are using an OAG DF is pretty unlikely but could still happen depending on how the guide camera connects to the guide port and whether or not you have cables tugging on the guide camera somehow.

 

rob




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