[Solved] Help with source of artifacts

You can leave out the flat frames to see if the integration now looks more normal. And if not, do the same with the darks and maybe some other calibration files. Step by step until the effect goes away.

At the very least, taking flat frames pointing at a laptop screen is not a good way as it's not a full spectrum it sends out nor necessarily very even. You may want to look at a flat panel instead.

Cheers, Vincent

Felix

I had the same problem when I used a Sigma Art 85mm lens with a Canon EOS 6D mk2. I redid the flats multiple times and couldn't solve it. Eventually, in desperation, I moved the lens to an ASI2600MC and the problem went away. I'm not sure if this helps you or not, but my guess is that something is happening in the camera software that is changing the offset in a non-linear way, causing APP to calibrate incorrectly.  Of course this is just a guess and I could be completely wrong. Vincent's suggestion that the laptop might not be an ideal light source is also worth pursuing, as I can't remember if I changed my light source when I changed cameras.

Vincent

Welcome back - hope you're doing well.

JC

Thank you very much for your quick replies, Vincent and JC.

Without the flat frames (just using dark frames), the integration does look more normal, but also the nebula is barely visible, so I am not sure if this fully excludes other sources of artifacts (I'm including a picture -- not sure how much I could improve this with more advanced post processing)

I have ordered a proper flat panel, and will see if that improves the situation.

A somewhat related question: How careful do I need to be when shooting the flats?

- How much does the sensor temperature matter, and could it account for such strong artifacts?  I shot my first set of flats right after I took these pictures, but then the later ones when the camera had room temperature.  The artifacts are similar so I don't think that's the main source of the problem.

- I took the camera off the mount and brought it inside to take the flats.  I assume that's less than ideal, but could that create such strong artifacts?

- How much does the exact maintenance of the focus position matter when taking the flats?  The focus point was certainly similar, but probably not identical.

- I shot the light frames with a lens hood (should I be doing that?), but the flat frames without a hood.  Is this right?

I suspect that in the place where I shoot these pictures there is quite a bit of light pollution (some adjacent houses with windows, etc.).  APP has a light pollution filter tool, but this runs after the integration of the frames, so since the camera had been moving for 2-3 hours, the light pollution in all the individual frames is probably at differing angles.  Is it possible that this circumstance could introduce artifacts like this, and is there some process that would allow removing the light pollution on the light frames before they get stacked? (Sorry, I have no idea how exactly the light pollution removal algorithm works, so I don't know if this is even a valid question.)

Thanks again for your help,

Felix

Posted by: @connor231

↑

Vincent

Welcome back - hope you're doing well.

 

JC

Thanks for your advice and yes I'm doing much better thanks for asking. 🙂

Posted by: @edx

↑

Without the flat frames (just using dark frames), the integration does look more normal, but also the nebula is barely visible, so I am not sure if this fully excludes other sources of artifacts (I'm including a picture -- not sure how much I could improve this with more advanced post processing)

Yes that makes sense as it's now not normalized and corrected. You might try now to do it again with the darks and bias and see if that makes it a bit better again. I can't see clear artefacts in this integration so it may well be that the flats are not working well here.

- How much does the sensor temperature matter, and could it account for such strong artifacts?  I shot my first set of flats right after I took these pictures, but then the later ones when the camera had room temperature.  The artifacts are similar so I don't think that's the main source of the problem.

As far as I know it shouldn't matter much as long as the flats are corrected with calibration files of the same temperature. I would however advice to take all calibration frames at the same temperature as the lights, just so you don't have other differences in the data.

- I took the camera off the mount and brought it inside to take the flats.  I assume that's less than ideal, but could that create such strong artifacts?

 

That depends, the focus should be the same as the lights you shot outside, so ideally you don't touch the setup much for flats. I'm not sure that would create these effects though. But it may point to having to improve your workflow a bit.

- How much does the exact maintenance of the focus position matter when taking the flats?  The focus point was certainly similar, but probably not identical.

Ideally at the same focus, because you want the exact same characteristics in your flats as in your lights for the illumination of the sensor.

- I shot the light frames with a lens hood (should I be doing that?), but the flat frames without a hood.  Is this right?

 

No, that may influence the illumination as well I think (I never did it without so not 100% sure there). Rule of thumb is to use the exact same setup as you took your lights with. Shooting flats certainly is a bit tricky so you may need to experiment a bit.

Illumination is very important, you don't want to look at the histogram of your camera as it's not lineair. A flat needs have its peak to be around the middle of a linear histogram, f you try this with the built-in screen of a Dslr, you'll end up illuminating too short usually. You then run the risk that the peak is cut off and you don't have enough information to process with.

I suspect that in the place where I shoot these pictures there is quite a bit of light pollution (some adjacent houses with windows, etc.).  APP has a light pollution filter tool, but this runs after the integration of the frames, so since the camera had been moving for 2-3 hours, the light pollution in all the individual frames is probably at differing angles.  Is it possible that this circumstance could introduce artifacts like this, and is there some process that would allow removing the light pollution on the light frames before they get stacked? (Sorry, I have no idea how exactly the light pollution removal algorithm works, so I don't know if this is even a valid question.)

If you end up having different light sources across the image, then that may be quite challenging indeed. APP corrects for a light pollution gradient, but not for different points of light sources. However you may still get a decent picture with a bit more work, but it certainly doesn't help. I see a ring effect in your integration which I wouldn't assume is from various light sources, but that's unclear at the moment. It might also be reflection internally or something like that. If you tested a few things here we can also have a look at the data if you want.

Hi @edx and @vincent-mod,

Seeing the intiial result in the first post of this topic, leads me to think that the issue is definitely in the flats show using the laptop screen. Taking flats correctly is probably the most important and most rewarding step in the data calibration process. Depending on the optics and the focal lengths, making good flats can be hard. Flat panels are there to help you with this, so please let us know if the flat panel gives better results.

Mabula

Vincent, thank you so much for this detailed reply!

It took me a while to respond here, because the flat panel I ordered only arrived today.  The image definitely looks different, but the artifacts are still there (it seems a little better to me, though?).  Of course, I have used the camera in the meantime and also changed the lens etc. so there are very many potential sources of errors.

I still took the flat frames without a hood -- the light panel is probably to small to get good flats with a lens hood on.  Perhaps I will try taking the light frames without a hood during my next session.  Taking the flats directly at the end of the session (which is much easier now that I have a flat panel) will probably also help.

One last thing I noticed (would be good to get your input on) -- when loading the flats (or also the master flat) into the image viewer, I noticed that in the dark areas the values of the R and B channel are 0 for quite a bit, only G increases.  So I am guessing that I am indeed underexposing the flats and perhaps this contributes to the problem?  I will give this another try.

I will report back after my next session, in case I figure out where the artifacts come from.

Thank you all so much, this is a great help.

- Felix

Yes make sure you at least have a nice peak and distribution of that peak in the flat within APP (on the top right for the histogram). If you have no good way to judge when your flat is exposed properly (because for instance you only have the DSLR lcd to view the histogram on) then try to expose with a peak all the way to the right on the DSLR LCD. Ideally a flat exposes for about 1-3 seconds, so you may have to bring down the brightness on the panel to match that. Then also take flat-darks at that exposure (which can be done inside, but make sure it's 100% dark, maybe putting the camera with a cap on and inside a box). Also, are you using any software to control the shots? Usually that software has a nice way of shooting flats as well and looking at a proper histogram.

Looking at the above integration, I can see dust still in the result, which means the flats didn't properly match the lights. So there still is something that needs to be done. You can repeat the flats, with the above advise as well and see. After that, if it still doesn't work, we will ask for that data so we can have a better understanding of what may be going on.

@edx Hi Felix and Vincent @vincent-mod,

Indeed, like Vincent indicates 😉 make sure that you expose your flats long enough to have the R,G,B channels all have proper histograms. Then the results should be much better.

Mabula

Thank you both again for your help.  I will need to try NGC 1499 again at some point, perhaps from a different location.  I think there is something about the light pollution in those light frames (in the setting that I am shooting in) that makes this particular target particularly challenging.  With the flat panel, I am getting good results on a different target, without artifacts like the one above.

Aside: When I was looking at the RGB values of the flat frames, they were stretched, so it is probably not surprising that they looked clipped.  When I turn off stretching, I am getting more reasonable values.  As you can tell, I am still very much learning how this all works.  It is still true that I need to expose the flat frames longer -- it looks like about 8 seconds in my particular case.

- Felix

Hi Felix @edx,

You are most welcome.

I will need to try NGC 1499 again at some point, perhaps from a different location. I think there is something about the light pollution in those light frames (in the setting that I am shooting in) that makes this particular target particularly challenging. With the flat panel, I am getting good results on a different target, without artifacts like the one above.

Yes, light pollution/sky conditions can play a role. The more you shoot flats and the more experience that you get with creating good flats, the better the result will be 😉

5-10 seconds per flat can be very normal, it will always work much better when compared to flats that are only fractions of a second.

Mabula