What is Frame buffer pixel format?

HI @High500

What does ‘16bit or 32bit precision’ exactly mean?
Does float has 16bits or 32bits of data per colour channel?
As I know, if set color depth 10bit in Window display setting , each RGB channel has 10bits of data and Alpha has 2bits of data.
ie:
R:10 G:10 B:10 A:2 Total = 32bits per pixel
What you mean is 16bits or 32bits per colour channel not per pixel
Is it correct?
If so, in what cases, should i set frame buffer pixel format Float RGBA?

Thank for your answer!