Analysis of imaging by moonlight

[Pete]

This is a study of moonlight imaging

On Nov 23rd there was a 70% illuminated moon that’s high in the S and on the 25th the moon was high and almost full. C9 (the Cave Nebula) was targeted each night with the ASI2600 mounted on the 102mm triplet and the Optolong LP-P filter. 8 minutes, unbinned, -15C cooling, 200 gain and 1 Hz guide rate.

Processing the images from the 23rd there was vignetting. To the point where processing was impossible. Note: The flat field from 11/20 was used & there’d been no problem on IC 10 with it.

1 nite1 frame nite1 flat s.jpg (296.09 KiB) Viewed 4735 times

But thinking that the problem might be the flat, a new flat field was run on the 25th, moving the illuminated panel around so that if there is a problem with even illumination it’d show up. Now, applying the flat from the 25th on the image from the 23rd:

2 nite1 frame nite2flat s.jpg (296.09 KiB) Viewed 4735 times

About the same. It looks like the problem’s not with the flat field.

Now here’s the image from the 25th with the full moon and a new color flat using the same exposure setup:

4 nite2 nite2 color flats s.jpg (268.96 KiB) Viewed 4735 times

Much worse.

The 25th using a mono flat:

5 nite2 with nite 2 mono flat s.jpg (268.72 KiB) Viewed 4735 times

A bit worse yet.

The image from the 25th using the 11/20 flat:

6 nite2 with nite1 flats s.jpg (274.61 KiB) Viewed 4735 times

Conclusions & Lessons learned:

The different flat fields didn’t change the results. All frames were given a 100% arc-sinH stretch. Even with the Optolong filter it appears that the full moon resulted in more vignetting. Since there was no significant vignetting of IC 10 back on 11/20 it can only be concluded that it’s a waste of time to image dim objects with the moon up.

A secondary observation: On the full moon night the mono flat is a bit worse than the color flat. They are supposed to be equivalent but, perhaps they’re not.

There's a possibility that running shorter exposures would have reduced the vignetting, but we're really talking signal/noise ratio and I don't see where that would be changed.

hgp 6 Dec 2023

[mark.m]

Pete:
At the AAVSO annual meeting last month, Dr. Arne Henden and I got into a discussion about the complexities of flat field images. By the end of that conversation, we both agreed that analysis of what's going on in a flat was beyond the capability of either of us. That says a lot about the challenge of getting a good flat.

One of the root problems is that a lot of the light entering the optics during a flat field exposure is out of focus, and the behavior of an out-of-focus light source is notoriously hard to predict. We kind of agreed that understanding what's going on in a flat field exposure requires blending what opticians call a "far-field" image with "near-field" blur. Arne used the phrase "black magic" to describe the making of a "good flat." One of the challenges is that about the only way to check a flat image to see how "good" it is, is to do exactly what you just did in your moonlight imaging. And all that gives you is a good/bad indication, without making any suggestions about what to do differently the next time.

What you did (imaging in strong moonlight) is a stress test for flats.

As Arne said, "You just keep trying things until you find a magic recipe that works well, and then you never change anything again."

- Mark

[Pete]

Hmmm. So what you're suggesting is that I try a twilight flat........... Could be.