At which point in processing should i BIN/resize my OSC CMOS images if i need to?

[ONIKKINEN]

I have roughly 0.9 arcsec/pixel resolution with my 8 inch newtonian and i find that my skies are really never (so far literally not a single time) good enough to warrant this. Binning 2x2 to 1.8 as/pixel is much more appropriate, if not ideal. I use DSS, SIRIL and photoshop for processing.

So at what point does this make most sense, processing wise? I capture all my images unbinned so that i don't lose the option to use the higher resolution data if the conditions allowed that for that night, but as far as i have understood binning a CMOS camera is the same in capture or post as the read noise is read on each pixel anyway, unlike CCD. Should i bin the stack before or after stretching to get the best results or does it matter at all? Right now i mostly finish processing and the final step might be a resize and JPEG save.

[Budgie1]

I do it before noise reduction and before stretching in PixInsight. 

Even after binning 2x2, my final process is to resize and PNG save the image.

 

[vlaiv]

Depends how you bin, but it should be done in linear stage.

If you split bin - then obviously before stacking and after calibration (this is advanced thing, so don't worry about it if you don't know how to split bin and/or don't have software to do it), otherwise - you can either do it after calibration and before stacking or right after stacking.

Easiest way to do it is to do it right after stacking before you start any processing on the data as that way you bin only one image - resulting stack (instead of each individual sub).

If you are going to resize image in the end any way - bin to that size before processing. That will let you pull out more in processing or deal with noise much easier.