Hi Alejandro,
the image it is a close up of the NGC 2264 @1016mm focal lenght.
it is hard to place the right samples.
The gradient comes from a bad flat and light pollution: light pollution is not severe but the integration time is long enough to make it noticeable.
I was thinking to find a way to get rid of the flat issue and then do my best for the light pollution residual gradient.
The image itself is an experiment (good ha data, RGB 2x2) so I am trying to find a quick way to fix data to a reasonable point before experimenting with the HaRGB blend (it seems difficult to blend ha data with 2x2 RGB...).
By the way, for now I have followed your suggestion and re run a model in DBE, which is "almost good".
thank you very much,
Andrea