Flats undercorrecting vignetting - Forums

Topic starter

16 days ago · Edited: 16 days ago

Hello all!

I'm facing issues trying to calibrate my light frames with flats, they seem to be undercorrecting the vignetting on the frame:

The equipment I'm using is the following:
- Modificated Sky-Watcher Quattro 250P OTA
- Sharpstar 1x GPU Coma corrector (very similar, if not a rebrand of the Sky-Watcher GPU CC)
- ASI2600MM with 36mm unmounted Antlia LRGB filters (Luminance filter used in this instance)
- My bortle is 5/6 but that shouldn't matter because all my other telescopes calibrate correctly

I tried a few things already:
- I first suspected of bad darks, so I did them again (60 images), both 180s darks for the lights and 0.1s flat darks for the flats.
- I tried making longer exposure flats - 1s and 5s. Both gave the same result as the 0.1s flats on the image above.
- I stacked the uncalibrated frames and it seems just as it should, with lots of vignetting:

Calibrating with flats really help but they do not correct 100% of the vignetting, which I find very strange.
- I'm sure the gain and offset settings of the lights match the darks and flats;
- I'm using the same driver to capture all data (including calibration) - which is the native ZWO camera driver in NINA (not the ASCOM camera driver).

I haven't tried other filters yet, but will I do it as soon as the sky clears up on a moonless night.

Suggestions?

Well Written Helpful Engaging

Oscar

16 days ago · Edited: 16 days ago

I think that's at least 90% fixable with GradientCorrection

Mauricio Christiano de Souza

Topic starter

16 days ago

Oscar:
I think that's at least 90% fixable with GradientCorrection

It can be, but on the long term I wanted to fix the root cause to not have to deal with it.

This also is possibly making me loose signal on the edges, which is not very good on larger dim targets.

Dan Watt

16 days ago

Few things worth trying:

Triple check that gain and offset is the same all around. This had me stumped for awhile once until I realized my mistake.

Try bias frames instead of darks altogether. Your workflow should be fine, but hey, see what happens.

I've read that some CMOS sensors go into a different sort of readout mode with exposures under .2 sec (something related to their origins in consumer photo/video cameras) so see if that makes a difference for you.

How are you taking your flats? Sky flats, flat panel, t-shirt, etc?

Helpful Engaging Supportive

Kristof Vandebeek

16 days ago

I have a nearly identical setup (Skywatcher 250PDS, TSGPU CC with a lot of vignetting, ASI533MM camera and Astronomik filters) and my flats are correcting the vignetting without any problems.

I use NINA for capturing 'sky flats' and I use the flat wizard with the telescope pointing to zenith, the histogram is set to 30% and the flats are captured with the same gain and offset that I use for the lights. I calibrate my flats with bias frames, not with dark flats.

If skies are not clear in the morning I still use sky flats, but I use two 30cmx30cm opaque plexiglass plates that are resting on the front of my OTA, this gives a sufficiently spread light for flats.

Helpful Respectful Supportive

Mauricio Christiano de Souza

Topic starter

16 days ago

Dan Watt:
Few things worth trying:

Triple check that gain and offset is the same all around. This had me stumped for awhile once until I realized my mistake.

Try bias frames instead of darks altogether. Your workflow should be fine, but hey, see what happens.

I've read that some CMOS sensors go into a different sort of readout mode with exposures under .2 sec (something related to their origins in consumer photo/video cameras) so see if that makes a difference for you.

How are you taking your flats? Sky flats, flat panel, t-shirt, etc?

I'll try bias frames.

I'm taking my flats with a Wanderer Cover V4 Pro EC Flat Panel, which doesn't have a linear brightness behavior, making it a bit hard to find the ideal exposure for LRGB filters.

Well Written

Mauricio Christiano de Souza

Topic starter

16 days ago

Kristof Vandebeek:
I have a nearly identical setup (Skywatcher 250PDS, TSGPU CC with a lot of vignetting, ASI533MM camera and Astronomik filters) and my flats are correcting the vignetting without any problems.

I use NINA for capturing 'sky flats' and I use the flat wizard with the telescope pointing to zenith, the histogram is set to 30% and the flats are captured with the same gain and offset that I use for the lights. I calibrate my flats with bias frames, not with dark flats.

If skies are not clear in the morning I still use sky flats, but I use two 30cmx30cm opaque plexiglass plates that are resting on the front of my OTA, this gives a sufficiently spread light for flats.

Interesting. I will try to take sky flats and see if the result improves.

Well Written Respectful

Francesco Meschia

16 days ago

In order to diagnose with data, can you please share:

1 uncalibrated light
the master dark you use to calibrate the light
the flats (not the master flat) you used to calibrate the light
at least one of the corresponding flat darks (not the master flat dark)

Thanks!

Mauricio Christiano de Souza

Topic starter

16 days ago

Francesco Meschia:
In order to diagnose with data, can you please share:

1 uncalibrated light
the master dark you use to calibrate the light
the flats (not the master flat) you used to calibrate the light
at least one of the corresponding flat darks (not the master flat dark)

Thanks!

Here is the Google Drive link to the raw data:

https://drive.google.com/drive/folders/1SdXIRUSlMVACbbmVzpQL4-MiW2pzCv9S?usp=sharing

I uploaded 30 Lights, 30 Flats, 60 Flat Darks and the 180s Master Dark.

If you need something else or find out a problem, please tell me!
Many thanks in advance.

Well Written Respectful

Francesco Meschia

16 days ago

Fantastic! Let me take a look and get back to you.

Well Written Respectful

Francesco Meschia

16 days ago

Great. The good news is: I performed a manual calibration with PixelMath, and the flats corrected perfectly.

Now that we know this, we need to look elsewhere for the root cause. The most likely place to look at is the processing pipeline. Did you use WBPP? If so, can you upload the log file produced by WPBB.

Well Written Helpful Insightful Concise

Mauricio Christiano de Souza

Topic starter

16 days ago

Francesco Meschia:
Great. The good news is: I performed a manual calibration with PixelMath, and the flats corrected perfectly.

Now that we know this, we need to look elsewhere for the root cause. The most likely place to look at is the processing pipeline. Did you use WBPP? If so, can you upload the log file produced by WPBB.

That is very good news! Did you stack it though? This vignetting appears only after stacking from what I could tell.

Yes, I used WBPP. I just uploaded a new folder called WBPP logs. Refresh the Google Drive page and it should be already there.

Francesco Meschia

16 days ago

No I didn’t stack, because Google Drive doesn’t want to let me download the whole folder, and it forced me to download individual files. I will try to give more time to see if it can actually prepare the archive and download it.

Jonathan Saine

15 days ago

Fix the root cause of the vignetting in the first place, as opposed to trying to fix it ex post facto.

Well Written

Dan Watt

15 days ago

Jonathan Saine:
Fix the root cause of the vignetting in the first place, as opposed to trying to fix it ex post facto.

And how might they do that? This amount of vignetting is normal when using an aps-c sized sensor on a newt designed with a 2" focuser and flats are a perfectly fine way to handle it. You'll notice the 30% or whatever less SNR in the far corners when going hard with processing but the solution to get rid of most of the vignetting would cost more then the scope itself.

Concise

Francesco Meschia

15 days ago

I downloaded and stacked the entire folder. There is a residual gradient, but I don’t think it’s due to vignetting. Let me explain.

First of all, if you use on the master light the same STF you use on a single uncalibrated light, there magnitude of the gradient looks much less. In other words, it’s the more aggressive STF made possible by the low noise of the master light that highlights the lack of uniformity of the background.

Here’s an example. This is an individual uncalibrated light with a rather aggressive STF applied to it:

📷 orig.jpgthis is the integrated master light with the same STF applied to it:

📷 integrated.jpgAt this level of stretching, there’s basically no vignetting left.

If I compare the contour plots of the two images, it becomes clearer that the residual pattern present in the background of the master light is quite different from the one visible in the uncalibrated light, not only in intensity but also in “shape” (centering).

Uncalibrated light:

📷 Image18_contourPlot.jpgmaster light:

📷 integration2_contourPlot.jpgSo I don’t think it’s residual uncorrected vignetting: could it be a light pollution gradient?

Helpful Insightful Engaging

Mauricio Christiano de Souza

Topic starter

15 days ago

Jonathan Saine:
Fix the root cause of the vignetting in the first place, as opposed to trying to fix it ex post facto.

I wish I could, but as Dan stated below, that is not financially viable. It would require redesigning the whole telescope, which at that point would be worth to just buy a brand new much more expensive astrograph, and that is not an option...

Dan Watt:
And how might they do that? This amount of vignetting is normal when using an aps-c sized sensor on a newt designed with a 2" focuser and flats are a perfectly fine way to handle it. You'll notice the 30% or whatever less SNR in the far corners when going hard with processing but the solution to get rid of most of the vignetting would cost more then the scope itself.

Mauricio Christiano de Souza

Topic starter

15 days ago

Francesco Meschia:
So I don’t think it’s residual uncorrected vignetting: could it be a light pollution gradient?

I thought this, but I was doubting it is light pollution because of the vignetting like pattern I'm getting. Anyway, this is a fair explanation.

I wish more experienced people had their inputs though. Maybe someone is able to explain why it is looking like this (vignetting effect)...

I also had an 8" f/4 astrograph and the light pollution gradients were much more smooth and in very specific directions rather than radial. My Epsilon and Rokinon don't produce gradients like this as well.

I still want to try sky flats and check how it looks on the other filters (RGB).

Lucas Gonçalves Miranda

13 days ago

Did you fix it? I’ve got the same problem

David Foust

12 days ago

Is it possible the focuser tube could be sagging a miniscule amount when you take your flats? Could the misalignment of the tube during flats vs lights explain how the similar but shifted vignetting pattern isn't aligning between the flats and lights?

I never had that issue on my newt, but I also used a 533mc pro with my newt, which didn't vignette whatsoever because of the small sensor size, so I've never had to handle correcting vignetting much.

Engaging

Mauricio Christiano de Souza

Topic starter

12 days ago

Lucas Gonçalves Miranda:
Did you fix it? I’ve got the same problem

Not yet. I'm waiting for the right time to make sky flats and see what the results are.

Well Written Respectful

Mauricio Christiano de Souza

Topic starter

12 days ago

David Foust:
Is it possible the focuser tube could be sagging a miniscule amount when you take your flats? Could the misalignment of the tube during flats vs lights explain how the similar but shifted vignetting pattern isn't aligning between the flats and lights?

I never had that issue on my newt, but I also used a 533mc pro with my newt, which didn't vignette whatsoever because of the small sensor size, so I've never had to handle correcting vignetting much.

That is possible but I would say it is unlikely because my focuser has 8 bearings supporting the drawtube + the rack and pinion. For that to happen at least a few bearings would need to be loose... I had this same model of focuser on 3 other telescopes and I found it very reliable at all instances... Also, I guess that much of a difference by the vignetting shifting would be visible while blinking frames.

Image circle shouldn't be a problem on this telescope, although I'm using an ASI2600. On the previous OTA, an 8" f/4, the vignetting was easily 100% corrected with flats. Coma corrector is the same as well. The only thing that has changed is the flat panel and the OTA.

Helpful Insightful Respectful Engaging

David Foust

12 days ago

Mauricio Christiano de Souza:
That is possible but I would say it is unlikely because my focuser has 8 bearings supporting the drawtube + the rack and pinion. For that to happen at least a few bearings would need to be loose... I had this same model of focuser on 3 other telescopes and I found it very reliable at all instances... Also, I guess that much of a difference by the vignetting shifting would be visible while blinking frames.

That makes sense; it was mostly a guess 😂 That's the only other thing I could think of, aside from some sort of light leak from the focuser or behind the primary, or somewhere else on the flats that isn't present on the lights, causing an imbalance on the flats themselves and resulting in poor correction.

I feel if it were a light pollution gradient, the gradient would be more linear instead of a radial pattern, and it would also be evident on the light frames... I've had to deal with some strong LP gradients before and I've never seen a radial LP gradient.

Helpful Insightful Respectful Engaging

Mauricio Christiano de Souza

Topic starter

12 days ago

Ok everybody. Great news:

Sky flats solved the issue completely. It seems like there is some difference between the natural night light entering the OTA and the light from the flat panel, leading to a mediocre correction of the vignetting.

Here is the data calibrated with the sky flats:

The gradients on the background are now much closer to what I'm used to from light pollution and it is easier to identify the dim dust in the field.

I appreciate the people that contributed on this topic. Hopefully this is useful for someone in the future that is having the same problem.

If I discover a way to use the flat panel successfully instead of sky flats, I will come back to tell my findings.

Well Written Helpful Insightful Respectful

David Foust

11 days ago

Glad to see you found a solution! Perhaps the flat panel is unevenly illuminated? I've definitely heard of that issue being the culprit for bad flats for other folks. Perhaps there's something up with the light panel?

Well Written Respectful