This one was a lot tougher than I thought it would be. At meridian crossing it only achieved an elevation of around 22 deg. High airmass, and very noticeable differential refraction between the 3 bands.

[ I normally avoid imaging below 30 deg elevation... But there's no chance to get Scorpio or Sagitarius unless I break my rule... ]

At first I thought something must be obviously wrong here. That semicircular feature at the lower left looks like it is a camera flat field artifact. But I checked my master flat, and not a hint of such a feature even after removing the dome and stretching to the same degree as this image. Must be real.

[ It IS an artifact! I just did a second batch of images taken a few deg south of this one, and lo and behold, the same semicircular pattern appears in exactly the same place. So pardon my dust... ]

The signal from the nebula was very weak, and very noisy. In fact, this image offered proof that 1/F noise exists in spatial dimensions from the sky too. As each tiny scale length was denoised, the next higher scale length dominated. So the denoising became a compromise between destroying information content and smoothness.

At high magnification you can clearly see the separation between red and blue channels on stars. That's the strong differential refraction. It also elongates the stars in the vertical direction.

-------

I think the artifact is the Moon, some internal reflection. The shape is a severely defocused source of some kind, well off axis. Perhaps I should wait another week or so, when the Moon is gone.

-------

Wow, I spent a whole day trying to track down image artifacts. New sky flats were obtained, and indeed the "moon ring" is now gone. Looking extremely closely at the images in the stack, and exploring the dark frame master after removing the overall dome structure, shows several things:

1. This data was tainted by high thin clouds, in addition to high amounts of light pollution from looking at low elevations over the city of Tucson. That markedly reduces inherent contrast in the image, and grossly increases the noise level, and

2. Decreased contrast, and high noise levels, encourages excessive stretching to "bring out the details". But noise, when overly stretched, will bring out what look like plausible structures in nebula background regions, even when there really isn't any data to justify those structures - just like seeing "horses and rabbits" in the clouds.

So the revision puts up what is probably justified by the data and tries for little more. A better image will happen when there are no clouds to mess things up.