2

What you are trying to archive is very hard if at all possible. The pinnacle of research today (including my own) is striving for good results on such tasks. DCASE challanges engage with similar tasks with some interesting results, though the make assumption which cannot be generalized to your case. In DCASE2019 task 3, for example, they assume up to 2 ...


1

A few points first Since you are using digital filters, you should be using residuez() not residue() Matlab's residuez() is not particularly well implemented: in case of higher order or steep filters, you can get a lot of numerically noise or results that are just totally wrong. Always check the result against your original target. Back to the question: ...


1

I think you consider that pulse in the spectrum as the noise. Then you should remove it by a lowpass filter of cutoff frequency $$ f_c = 5 ~\text{kHz.}$$ Slightly less than this should be used for guarding against the transition width of the lowpass filter. So you better select something like $4.8$ kHz for the lowpass filter cutoff frequency. Note that ...


Only top voted, non community-wiki answers of a minimum length are eligible