beta 34 vs beta 35 on star registration (bad results)

Dear Mathias @mrtipsy,

Thank you very much for reporting your issue.

The actual number of stars reported by star analysis does not mean if more stars reported is better actually. It could very well be that beta34 also detected a lot of non-stars as stars and thus gave a higher star count. On all our test data we also see this on some frames and we clearly see that beta35 is way more robust in detecting real stars instead of noise, hot pixels, cosmic rays and such.

Without seeing your data, it will be very hard to say what actually happens in your case.

What happens in registration, I assume that the frames still register okay with default settings in both 3) analyse stars and 4) register?

I have tested beta35 on a multitude of data sets and for all sets beta35 was way better in the star analysis result (you can visually check with the star map function above the image viewer) and the registration. If beta35 is not better on your data, then clearly we need to see your data. Are you willing to share it?

" reference pic for ic 434 and it found almost 2000 stars on version 34 and a little over 800 on version 35"

800 is very decent actually I think and should give no problems for registering the frames.

"and its the same on the fit files. before it found over 100 stars on each fit, and now 20-40 stars on the same fits

Ps: Seestar s50 fit files"

The Seestar uses very short exposures, it is very likely that the 100+ star count in beta34 is quite bad, it could detect many noise artefacts for stars actually. So do a double check visually with the starmap and let me know how that looks.

Mabula