Defocused star images - to better show color - challenging data set

I took a series of intentionally defocused (wide field) images in order to better show star colors. (Tiny dots, especially faint ones, do not convey color information well. Larger dots convey color information better.)

The blur circles were about 40 - 50 pixels in diameter (after 2x2 binning).

That proved to be a challenging/weird data set for Astro Pixel Processor. Analyze stars? Register images? Calibrate background? That was all impossible.

But removing light pollution/gradients seemed to work well. (And for a wide angle image, that's important.)

Here is the result from stacking with no registration.

After correcting light pollution/gradients:

(Note that I did important post processing in GIMP. My goal was to compress/flatten brightness differences so that colors of faint stars could be better shown. I did this by boosting lightness in HSL color space, then going back into RGB color space to darken the sky background.)

This is not meant to be a scientific analysis/display of star color. It's a crude attempt at showing relative color differences, especially for fainter stars.

If you want the set of raw images for APP testing, just ask. (I understand if you don't want to improve star analysis/registration of images that are intentionally defocused. This is not a typical way to do astrophotography.)