ok, thanks for the quick reply. i would like to ask for some more clarification.
as you can see from the code, what i want to estimate is actually one posterior per channel. at least for mu and std. and i would also be ok with adding the same shape parameter to the nu variable.
so, the question is, whether these posteriors somehow depend on each other. i.e. if i would estimate it for channel 1 and channel 2 separately, would the resulting two posteriors be the same as if i would estimate them together (but still would get two posteriors because of the shape parameter).
i have actually tried this out on real data. and the difference between calculating all at once and calculating them in separate calls is roughly the same as when i just do the “all at once” calculation twice but with different random seeds.
but please bear in mind that i am totally new to this field and i might miss something here. this is why i am asking.