Back to forums

Artefacts in stacked image.

ZWO ASI585MC Pro Processing Calibration Deep Sky Artifacts
16 replies•524 views
MFarq avatar
MFarq avatar
Evening all. I'm having trouble removing dust donuts from my stacked images. See below.
This is the unprocessed stack calibrated with flat frames which I checked in APT for the correct ADU.
Setup is a Celestron C8 with an ASI 585MC camera.
Any suggestions as to what might be going on and  what can be done about it greatly appreciated. Cheers. Mark.



Noah Tingey avatar
Noah Tingey avatar
I suspect that you did not take flat frames correctly. A checklist of troubleshooting steps off the top of my head:
  • Was your telescope in the same focus position for the flats as it was for the lights?
  • Did you take apart your imaging train in between taking lights & taking flats?
  • Were your flats taken with the same gain & offset?
Well Written Helpful Insightful Concise
MFarq avatar
MFarq avatar
Thanks Noah, much appreciated. 
The focus was unchanged between the lights and flats and the imaging train was left intact between imaging the lights and the flats. 
The gain for both the flats and lights was the same. 
I'm not sure what you mean by offset?
Well Written Respectful
Rob Kiefer avatar
Rob Kiefer avatar
Looks like that something did not work out with your flats. Check your single flat frames and the masterFlat, compare them if they show the same pattern. As well check the calibrated light frames if the flat correction worked. Maybe check out these two videos from Adam Block. They might help.
PixInsight: Flats (from Adam Block's Office Hours)
EZ Selective Rejection in PixInsight
Helpful
MFarq avatar
MFarq avatar
Thanks Rob, will have a look.
Willem Jan Drijfhout avatar
Willem Jan Drijfhout avatar
First thought is indeed like others mention, a problem with the flats. But if you checked all that and  it isn't it, maybe we should look for less likely options.
Upon closer inspection, some of the donuts appear to be not fully concentric, but with the dark dot in the middle trending towards the center of the frame. Can there be some kind of reflection in play? How does the pattern look before and after the meridian flip, do you see it in other parts of the sky, is it the same in all frames throughout the night, etc.?
Helpful Insightful Respectful Engaging Supportive
Konstantin Dzuin avatar
Konstantin Dzuin avatar
Just a guess, but did you properly calibrate the flats with bias or darks and lights with darks only?
MFarq avatar
MFarq avatar
Willem Jan Drijfhout:
First thought is indeed like others mention, a problem with the flats. But if you checked all that and  it isn't it, maybe we should look for less likely options.
Upon closer inspection, some of the donuts appear to be not fully concentric, but with the dark dot in the middle trending towards the center of the frame. Can there be some kind of reflection in play? How does the pattern look before and after the meridian flip, do you see it in other parts of the sky, is it the same in all frames throughout the night, etc.?

Thanks for the suggestions Willem. I did wonder if it could be something other than the flats as they are effective on my other scope/camera combinations.
Well Written Respectful
MFarq avatar
MFarq avatar
Willem Jan Drijfhout:
First thought is indeed like others mention, a problem with the flats. But if you checked all that and  it isn't it, maybe we should look for less likely options.
Upon closer inspection, some of the donuts appear to be not fully concentric, but with the dark dot in the middle trending towards the center of the frame. Can there be some kind of reflection in play? How does the pattern look before and after the meridian flip, do you see it in other parts of the sky, is it the same in all frames throughout the night, etc.?

Thanks for the suggestions Willem. I did wonder if it could be something other than the flats as they are effective on my other scope/camera combinations.
Konstantin Dzuin:
Just a guess, but did you properly calibrate the flats with bias or darks and lights with darks only?

Hi Konstantin. I usually just calibrate the light frames with flats.
Konstantin Dzuin avatar
Konstantin Dzuin avatar
Hi Konstantin. I usually just calibrate the light frames with flats.


In order to get the master flat image, you stack calibrated individual flat frames, usually by matching exposure dark frames, or just bias frames.
To calibrate light subframe with the flat frame, you need to subtract the read and dark current noise components from the original frame, to get correct result from the flat field master image division. Dark frames contain both dark current and read noise component, so it should be either/or correction, either dark master frame, or bias master frame.

If either of these calibrations were not done correctly, you'll face over- or under-correction in your calibrated individual light frames.
The error is cumulative, and will be very prominent on the stack of light subframes.
Helpful
Rick Krejci avatar
Rick Krejci avatar
Hi Konstantin. I usually just calibrate the light frames with flats.


Like Konstantin said, this is your problem.    With that camera, you should be fine to just take about 30-50 bias frames, create a master bias and use that for flat calibration and light calibration if you aren't taking darks.  THEN use the bias-subtracted flats to flat calibrate the bias-subtracted lights   

The bias is essentially an additive offset/pedestal to both the light and the bias so the flat frame won't properly divide out if you don't subtract that offset from both.    Most SW suites like Pixinsight will handle that for you automatically.

You can also use time/temperature matched Darks for both lights and flats which will slightly more accurately remove the Bias + Dark signals, but modern cameras like yours have little dark signal compared to bias and light, so the difference is minor...just biases are "good enough" to get you going in the right direction.
Helpful
Noah Tingey avatar
Noah Tingey avatar
Thanks Noah, much appreciated. 
The focus was unchanged between the lights and flats and the imaging train was left intact between imaging the lights and the flats. 
The gain for both the flats and lights was the same. 
I'm not sure what you mean by offset?

Sounds like others have likely figured out your problem, but just to explain the concept of the offset setting:

Raising the offset value effectively shifts your histogram to the right by a small amount before the pixel values are recorded. This ensures that the pixels that would be darkened by the noise are never clipped to black. A clipped noise profile will mess with your integrated results.

Ideally you raise the offset value such that zero pixels are clipped to black. So you'd take your camera and have it loop short exposures with the lens cap on while you adjust the offset value such that the minimum pixel ADU is always greater than 0. Do note that the offset values of all of your frames (including calibration frames) need to be the same. Just like with gain.

If you use an ASIAIR, then you cannot adjust the offset value.
Well Written Helpful Insightful
Konstantin Dzuin avatar
Konstantin Dzuin avatar
@MFarq,

Did you manage to try different approach to calibration? Any updates on your problem?
MFarq avatar
MFarq avatar
Konstantin Dzuin:
@MFarq,

Did you manage to try different approach to calibration? Any updates on your problem?

Hi Konstantin.

I got some new data last night, so will let you know if I've managed to sort my Flats out. Cheers. Mark
MFarq avatar
MFarq avatar
Konstantin Dzuin:
@MFarq,

Did you manage to try different approach to calibration? Any updates on your problem?

Hi Konstantin.

I got some new data last night, so will let you know if I've managed to sort my Flats out. Cheers. Mark

It did help with much reduced artefacts, but there are still some.
MFarq avatar
MFarq avatar
MFarq
Konstantin Dzuin:
@MFarq,

Did you manage to try different approach to calibration? Any updates on your problem?

Hi Konstantin.

I got some new data last night, so will let you know if I've managed to sort my Flats out. Cheers. Mark

It did help with much reduced artefacts, but there are still some.

MFarq avatar
MFarq avatar
Hi all. Issue sorted. I took some bias frames and stacked along with the lights and flats.  No more dust donuts. Only had this issue with my C8 and F4 imaging newt. Stacked light frames and flats are completely artefact free with my refractors. Hanks for all the help. You live and learn :-)