Hello,
I am still seeing dust motes/donut shaped spots even after calibrating with flats. I tried few exposures with flats but no improvement. I use dark, bias and flat calibration frames and default wbpp settings. from the attached screenshot, you can see the smaller donut is almost cleaned up but the larger donut at bottom right is appearing in the final stack.
Is there anything I can try to correct this?wbpp_flat_error.JPG
andrea tasselli · Sep 30, 2025, 06:16 PM
The camera/filter has shifted between the flat exposures and the light exposure. Redo the flats, if you can.
i did that too.. didn’t help. guess I am out of options for this time
Arun H · Sep 30, 2025, 06:45 PM
certainly try what Andrea suggests.
Also:
Take flats frequently, every session if possible, especially for RGB, and especially for light polluted sites
Make sure your filter wheel is recalibrated at each power up before taking lights and flats
If above does not fix, your only recourse may be what I had to do - clean my filters.
thank you @Arun H . I do agree with Andreas feedback. there definitely has been a slight shift. close inspection of dust motes does tell that. Will consider these advices for next session. definitely take flats without disturbing the image train from the mount. thank you both
Amit:
i did that too.. didn’t help. guess I am out of options for this time
andrea tasselli · Sep 30, 2025, 07:15 PM
Amit:
i did that too.. didn’t help. guess I am out of options for this time
*Not really. If you are desperate to get that data there is an additional option, which is to shift/rotate the flat so that it matches the lights. It is rather time-consuming and requires knowledge of either Iris or Siril. I can do that too.
@andrea tasselli , you always amaze me with you knowledge from your post responses. I have siril but no idea how to match the flat
andrea tasselli · Sep 30, 2025, 07:15 PM
Amit:
i did that too.. didn’t help. guess I am out of options for this time
*Not really. If you are desperate to get that data there is an additional option, which is to shift/rotate the flat so that it matches the lights. It is rather time-consuming and requires knowledge of either Iris or Siril. I can do that too.
Another question - On my attached image, the donut edge/circumference appears thicker (kind of 3d) compared to the one in the flat frame, which has a sharp edge. Is it because my imagetrain (mainly the secondary mirror of edgehd8) shifted a little after meridian flip and the dust mote shifted in the lights frame itself?
Amit:andrea tasselli · Sep 30, 2025, 07:15 PMAmit:
i did that too.. didn’t help. guess I am out of options for this time
*Not really. If you are desperate to get that data there is an additional option, which is to shift/rotate the flat so that it matches the lights. It is rather time-consuming and requires knowledge of either Iris or Siril. I can do that too.
Another question - On my attached image, the donut edge/circumference appears thicker (kind of 3d) compared to the one in the flat frame, which has a sharp edge. Is it because my imagetrain (mainly the secondary mirror of edgehd8) shifted a little after meridian flip and the dust mote shifted in the lights frame itself?
Amit:andrea tasselli · Sep 30, 2025, 07:15 PMAmit:
i did that too.. didn’t help. guess I am out of options for this time
*Not really. If you are desperate to get that data there is an additional option, which is to shift/rotate the flat so that it matches the lights. It is rather time-consuming and requires knowledge of either Iris or Siril. I can do that too.
@andrea tasselli , you always amaze me with you knowledge from your post responses. I have siril but no idea how to match the flat
I have the EdgeHD 9.25. Dust on the corrector plate, primary or secondary won’t make a donut. I don’t even think dust on the flattener lens (in the exit tube) will be visible, but it doesn’t hurt to blow it off anyway.
Probably the dust mote for the big donut is on the filter. The dust mote for the small donut is on the sensor window. My guess, as other have said, for the ‘embossed’ look of the big donut is because the filter was not in the same position for the lights as for the flats, for some reason. It then gets over-corrected on one side, and under-corrected on the other side.
I had an issue like your small donut not too long ago. I tried to fix it by making a mask to match the donut and adjust the image background there. Limited success.
The real solution for me was to blow the dust off the sensor window and filters, and then use a flashlight, a magnifying glass and a Q-tip to locate and remove all remaining visible dust motes (that could cast a shadow on the sensor).
The imaging train is sealed for us EdgeHD users, so it stays clean. We do not need to treat dust as a fact of life.
Bill Dirks · Oct 1, 2025, 12:05 AM
I have the EdgeHD 9.25. Dust on the corrector plate, primary or secondary won’t make a donut. I don’t even think dust on the flattener lens (in the exit tube) will be visible, but it doesn’t hurt to blow it off anyway.
Probably the dust mote for the big donut is on the filter. The dust mote for the small donut is on the sensor window. My guess, as other have said, for the ‘embossed’ look of the big donut is because the filter was not in the same position for the lights as for the flats, for some reason. It then gets over-corrected on one side, and under-corrected on the other side.
I had an issue like your small donut not too long ago. I tried to fix it by making a mask to match the donut and adjust the image background there. Limited success.
The real solution for me was to blow the dust off the sensor window and filters, and then use a flashlight, a magnifying glass and a Q-tip to locate and remove all remaining visible dust motes (that could cast a shadow on the sensor).
The imaging train is sealed for us EdgeHD users, so it stays clean. We do not need to treat dust as a fact of life.
thanks Bill for the explanation.
You can get the flats perfectly clean, no visible donuts even when stretching the contrast.
And that’s a great target! Good luck.
Hello
I have exactly the same problems and here is a small summary of my experience:
small donuts: dust on the sensor window
big donuts: dust on the filters
experience shows that the filters dust move always and quite a lot because at each filter move the filter is never back at exactly the same position . Unfortunately the sensor window dusts move too sometimes, but much less. Shifts of the dust shadow on the sensor: 10 to 20 pixels for dust on the sensor window but usually several hundred pixels for dusts on the filters (I have 5.2 micron pixels)
the best solution:
clean everything but even there in my case dusts come back after a while even if I keep closed the optical chain
make flats very often and for each filter
avoid imaging when there is a lot of background light (moon for instance)
the brute force solution: at treatment make a starless image and a star image (using for instance StarXterminator if you have pixinsight), then erase the defaults using clonestamp and then reassemble the corrected starless with the stars. Of course, this is not always possible but when it works it is probably the easiest and fastest way to proceed
other solution: keep a library of flats and choose the flat closest to your donuts position or shift/rotate your flats to reproduce the real dust positions but this is extremely time consuming and the result is never perfect
to conclude 1) make often flats 2) clean filters and camera window 3) avoid light polluted nights
regards
JP
I would add if the dust mote is shifting around the frame, as in its position is shifting on the sensor, then I don’t think rotating the flats will matter. Taking flats often helps, but if dust motes move around the only thing I can think of is to pull the stars out of the image and use an image processing program like Photoshop or Affinity to clone out the problem areas if possible. Which won’t work if they are sitting on top of important details.
Jean-Pierre Lees:
to conclude 1) make often flats 2) clean filters and camera window 3) avoid light polluted nights
thank you all for sharing your experiences and advice. i am just using a filter drawer and I do see dust on the filter and a very slight play (meaning the filter could have been screwed a tiny bit more). I will keep these things in mind for for next shoot. thank you🙏
Well, I do this in Photoshop - separate the stars and then eliminate those residual mots, rings and all that stuff - is easy to do in PS
Amit · Oct 1, 2025 at 03:05 PM
thank you all for sharing your experiences and advice. i am just using a filter drawer and I do see dust on the filter and a very slight play (meaning the filter could have been screwed a tiny bit more). I will keep these things in mind for for next shoot. thank you🙏
That’s the problem with filter drawers, every time you take it out you’re inviting more dust. A filter wheel keeps the setup self-contained to some degree.
Which camera are u using? Have u tried processing without calibration frames? Have you tried turning off local normalization? With my OSC 533, calibration frames never helped. My mono 2600 does pretty good changing the settings i mentioned without calibration frames. Im new at this, Frank from Seti Astro has great tutorials.
Good Luck.
The flats scenario you have shown in the original post, is what I called the “embossed mote” or the “white mote” these occur when the flat corrects for a defect that is no longer present. This is different than a “dark mote” or “black mote” as I call them, these are defects that were present, that the flat did not correct for at all.
The only way to solve the black mote, is to have flats with the same mote in them. The only way to solve the white mote, is to use flats that did NOT have the mote in the flat, that was not present in the lights.
This is why we take flats every single night, and we use automation software (Minerva) to calibrate data daily with the fresh flats taken. Even with this level of consideration for flats, we still run into situations where the mote moves once or more over the course of the night, and ruins the lights.
In this case, you have a few options.
Hope the one or two frames with uncorrectable white or dark motes, are buried in the pixel stack and not visible. This is actually easier than one would think it is.
Obtain a set of flats to correct either case properly.
Use clone stamp to remove the offending motes. Or use PS for similar means.
Reshoot the lights, and throw away the bad ones.
Motes are a pain at times, and is a good reason to keep the camera, filters, and window clean as best you can do. Using a nightly regimen of sky flats can really reduce this issue very close to zero.
BlackStarsAstro · Oct 4, 2025, 12:32 AM
Which camera are u using? Have u tried processing without calibration frames? Have you tried turning off local normalization? With my OSC 533, calibration frames never helped. My mono 2600 does pretty good changing the settings i mentioned without calibration frames. Im new at this, Frank from Seti Astro has great tutorials.
Good Luck.
I use both the 533mm and mc pro cameras. I started taking calibration frames very early in my imaging as I started noticing the dust motes/Vignetting. taking calibrations especially flats has always helped clean my images. this was strangely the first instance of it. Based on all the feedbacks I received, I am convinced that it most likely was caused due to filter shift/movement. I noticed a small play in the filter holder which I have tightened. Just waiting for the next clear sky to image and check.
I also have a roll off roof which keeps my optics evenly dusty.😀
Hi Amit
I had the same problem last year: in case you use an electronic filter wheel, this has a tiny difference whether turning left or tight. default is to do that to minimize the travel time to the next filter. Whenever you don’t use the filters in its order in the wheel you get this effect.
Drove me crazy as would sometime happen and sometime not :-)
By simple checking the box “unidirectional” it only ever goes .
In case you are using a filter drawer I coul imagine that the filters might shift a little when removed and reinjected. You might want to check for too much space a wiggling …
Arny