Neowise SiriL processing help/pointers request

PeterdH · July 23, 2020, 3:08pm

Hi there,

this is my first post here. Be gentle

Like many other photographers I’ve been busy trying to take pictures of the great comet Neowise. Taking the pictures was fun, but the post processing has been quite frustrating. First and foremost I assume it is ME who messes things up. But with me there are many other photographers having the same issue with Neowise, especially in SiriL processed images, so maybe it’s a general misunderstanding on how to process these images properly in SiriL. So if you experts could give me a few pointers on what I do wrong, or what I should improve, that would be GREATLY appreciated. Here we go.

I am fairly new to SiriL. I have read the guides and watched many videos, but this specific issue I have not encountered.

The problem is there is an ring artifact with a very dark center which becomes visible after BG subtraction… No it doesn’t matter which lens I use or at which focal length (I have not tried stacking <70mm) or on which day I take the images (low/high moisture), if I use 10 subs, or 100… it always appears. When I try similar or deeper stacks on star fields, I do not get it… so if it is a flat field problem, then why I don’t get it for the star fields?

I pre-process the images with Bias and Darks. (No flats, which ‘could’ be the issue), register, stack, remove green, color calib and then I attempt BG subtraction.

Usually take 15 samples per line avoid any samples near the comet’s tails, try 4 harmonics, 3, 2, 1, or a combination to get it as flat as I can. In the manual there is a UI where you can see the BG model, but I assume that is not available for MAC as I can’t seem to locate that. I tried taking samples in the high BG regions only, high BG regions only, only along the sides of the comet, leaving all the middle lines sample free… none of which works…

Here is an example. (Yes I also don’t understand how the stretching correlates to output, but that’s a different topic). ANY help is appreciated… you experts probably recognize the issue from a lightyear away.

rbarbera · July 23, 2020, 3:54pm

Welcome Peter, I’m afraid that background subtraction is not substitute for flat field correction. I’ve suffered the same problem each time I try to compensate for vignetting when I didn’t have flats available.

Perhaps you can try to don’t apply the background subtraction in SiriL. Instead, export the stacked image and try to apply a vignneting correction on a photo processing app. But if your registration has move each frame an apreciable amount, perhaps this correction will not be effective.

PeterdH · July 24, 2020, 7:33am

Hi Rafa,

thanks for your imput. So you think this is indeed a Flat Field problem. I don’t understand why there’s a ring then… however, with 4 polynomials you can get complex surfaces. Great. Well, that gives me motivation to create some flat fields.

Re your other suggestion, wouldn’t it be better to remove the vignetting before loading into SiriL? I was searching for people pre-processing images before SiriL, but I haven’t found much. I don’t know yet if the SiriL sw takes all advantages from RAW files, meaning if I pre-process a RAW file I loose dynamic range etc… So much to learn always

Cheers

Peter

lock042 · July 24, 2020, 7:46am

Fore sure it is related to the flat field. Background extraction will never fix a lack of flat field.

rbarbera · July 24, 2020, 2:52pm

Because you lens has rotational symmetry around the optical axis :).

with 4 polynomials you can get complex surfaces

Yes but it seems that not precise enough to follow all the complex vignetting some lens are causing. The remaining, once you remove the interpolated low frequency signal is this medium / hight frequency difference that shows as concentric rings once you star to push the contrast.

Look at this extreme case:

I’ve tried to use background removal… on an image that has:

Vignneting
Sky light pollution gradient
Internal reflection on the teleobjective.

I’ve applied flat fields, but the flat fields didn’t show the reflection from Venus, so the annulus and the external third ruined the picture

rbarbera · July 27, 2020, 6:34am

Hi Peter, I’ve experienced a clear example of rings generated with the lack of flat frames. This weekend I’ve tested a new optical setup: A 5" Mak with 0.6x reducer. I expected some vignetting, but not so much, anyway. This is the output from the setup:

If I apply Histogram Stretch it will be more evident that I’ve only an strong vignetting and no other aberrations. Also, you can see that I’ve choose a 20x20 grid for the background model.

I didn’t take flat frames, because I was running out of time and I was only interested on star profiles (how much aberrations it has near the borders). So, I tried to “fix it” with a 4 order background fit:

And as expected it didn’t work.

So, from now, the only suggestion I’ve is: Remember to always take flat frames

PeterdH · July 27, 2020, 8:42am

Wow, thank you so much for doing this experiment Rafa!

lock042 · July 27, 2020, 9:12am

For me, flats are the most important frames.
If you are too lazy: so don’t shoot dark frames, but do not miss flats.
Really :).

PeterdH · July 27, 2020, 9:46am

Thank you too for all your help and insights. Very insightful.

sguyader · July 27, 2020, 11:53am

Isn’t it possible to take flats afterwise?

vinvin · July 27, 2020, 12:58pm

no, because dust will not be at the exact same place, so it will show them more than without, and it’s hard to get the exact same focus and orientation of camera (and the rest of the optical train) once it has been unmounted.

Mockus · July 27, 2020, 1:37pm

I totally agree with Vincent: If you want to have perfect results, you have to shoot the flats with the same settings (And the same dust) as the lights.

What you can try and what I’ve done in the past: I’ve tried to replicate the original conditions as close as possible. So I went out focused again on a bright star and then shot the flats with this setup. This will of course not remove any dust (it may introduce light spots from new dust!), but it will remove the vignetting quite good.

So if you lights were shot with a very clean sensor, and the sensor is still clean, it may be worth the effort to shoot the flats afterwards, though it is not optimal.

sguyader · July 27, 2020, 1:40pm

Well, to me flat field frames are more important for the correction of the vignetting than for dust removal. Dust removal easier to correct in post.
Regarding focus, if you focus at infinity both for your astro images and for generating the flat field frames, I would say that slight focus variation would not affect significantly the flat field correction.
At last, I don’t understand the importance of camera orientation on flat field frames. Or maybe I totally missed the point?

sguyader · July 27, 2020, 1:41pm

My point was that it would still be better than no flats at all, even if it’s not optimal.

Mockus · July 27, 2020, 1:52pm

agreed. Or at least that fits with my own experiences so far.

lock042 · July 27, 2020, 3:20pm

Here an example of dust removal.

rbarbera · July 28, 2020, 11:57am

Why do you call it flat frames when clearly they are dirty frames?

lock042 · July 28, 2020, 7:54pm

Fair enough. But this is the best evidence that flat frames are the most important frames in the preprocessing chain tool.

rbarbera · July 28, 2020, 8:51pm

Indeed. Nice example of why you should always take flats.