Yes, stacked in APP, and yes using ST guidelines such as "remove light pollution" off in Calibrate, "neutralize background" off in Normalize, etc.
Optical train is Celestron 120, 0.6x reducer, spacers setup for 55mm back focus from the reducer as specified and illustrated by ZWO. The 1.25 L-Pro filter was mounted inside the ASI294MC as recommended by ZWO...in other words as close to the sensor as it can be.
The data I uploaded was for two sessions, each involving a meridian flip. Yes, because of the artifacts around the edge, I cropped the image before putting ST to work.
- On August 27 I installed the reducer, successfully imaged IC 134, and created flats and dark flats the next morning before successful processing the results with ST.
- On September 4th I imaged NGC 6960 with no change in the image train and produced a good dual band image with APP and ST.
- On September 6th I imaged M31, again with no change in the image train.
- On September 10th I again imaged M31, also with no change in the image train and created a new set of flats.
My last image, in post #12 above, was created with the new flats and just the Sept 10 lights. I'm terrible a keeping a useful log so I can't honestly say why I created the new flats—but it makes me suspicious that I changed something.
I suspect, thanks to your help, that the mess is the result of careless application of the new flats to both the new and old data.—that and over-enthusiastic post-processing.
I'm going to reprocess the Sept 6th data with the Aug 27 flats and then combine that result with the Sept 10 result that used the new flats and see what ST Compose produces. I'll post the result here
Edited by Tailspin45, 28 September 2021 - 01:51 PM.