With the CMOS cameras any binning must be done after the image acquisition in a suitable software package.
The only software, so far, I've found is PIPP.
https://sites.google.com/site/astropipp/
This offers 2x2, 3x3, 4x4 binning
You can run a SER video (taken with Firecapture) through PIPP with say 3x3 binning and get an output file (SER).
This can then be processed through AS3! etc but with the larger plate scale.
I have a SER video taken with the ASI 1600 which gives a plate scale of 0.79 arcsec/pixel. The image size is 4640 x 3496 (as a TIF, 93 Mb)
Run through PIPP then AS3! the size is 1192 x 1192 (as a TIF, 8.1 Mb), plate scale 2.4 arcsec/pixel.
I also use Irfanview viewer, and can reduce the size of an image....
https://www.irfanview.com/
If I reduce the original image (4640 x 3496) to (1582 x 1192) the TIF is only 725 Kb. It seems to use the Lanczos "compression"
https://en.wikipedia.org/wiki/Lanczos_resampling
So, the question is - What's the difference between the PIPP binning and the Irfanview re-sizing?????