Here's what I was able to come up with using your M13 data.
I used DBE twice, which eliminated most of the gradients that were present in your version, although I was unable to remove every gradient, especially near the corners of the image, so in the end I just cropped the edges off. I suspect it could be a flats issue, so if your imaging train is still in the same layout, I suggest re-taking your flat frames and see if that eliminates the gradients on the edges.
For my DBE settings, I used a tolerance of 10, smoothing factor of 0.3, radius sample of 20 and the amount of samples per row to 20. I then generated the samples, where I just moved any samples that were on any stars. I then used DBE again once I stretched the image with GHS, with the same settings, although I selectively placed more samples where I thought it was needed. I didn't use SXT as I didn't feel it was necessary for an image of this type of target. BXT did work to a certain extent, but it is evident where the trailed stars were left if you look closely.
Here's my PixInsight workflow I used for your reference:
DBE
SPCC
BXT
NXT
GHS
SCNR
DBE
SCNR
Correct Magenta Stars Script to remove the magenta background cast
Curves Colour Balance
Curves Contrast Enhancement
HistogramTransformation
ACDNR
HistogramTransformation
NXT
Export as a 16-bit TIFF into Photoshop for final tweaks
You have good data! You even managed to pick up some small background galaxies. There was a decent sized one near the bottom left-hand corner of the image, but due to the gradient I had to crop it out unfortunately. I haven't actually shot M13 yet, although I might try and see what I can do when Hercules rises high enough in the sky where I live in Australia.
Zak
