I've heard the term "dithering" used for this, in electronics and a sort of related (reverse?) idea in displaying images on low color depth displays. I think. F

I've heard the term "dithering" used for this, in electronics and a sort of related (reverse?) idea in displaying images on low color depth displays. I think. For example if you have a DAC generated 8-bit data, but you want finer resolution, then you could imagine averaging maybe every pair of samples, which if you add (and divide by two) the samples, you get an extra bit of information. But that only works if your signal (including the measurement circuit) is not rock steady, which would make both samples identical and so you don't gain any information. Adding noise helps the samples be different which can increase the effective resolution. If you were just on the brink of a discretization bucket, a kick of noise will likely push you into the next bucket