This pr tackles some core issues with pyrasa_sprint. There was sth wrong with the previous implementation that resulted in some weird artifacts from time to time. This is fixed now alongside #22 , #30 and #31. However, in order to do so I had to go back to use scipy.signal.stft which was recently marked a legacy function from scipy. @thht there are ways to tweak the newer scipy.signal.ShortTimeFFT towards producing an output like scipy.signal.stft (see Comparison with Legacy Implementation in https://docs.scipy.org/doc/scipy/tutorial/signal.html#tutorial-stft), but I didnt manage to do so yet..
This pr tackles some core issues with pyrasa_sprint. There was sth wrong with the previous implementation that resulted in some weird artifacts from time to time. This is fixed now alongside #22 , #30 and #31. However, in order to do so I had to go back to use scipy.signal.stft which was recently marked a legacy function from scipy. @thht there are ways to tweak the newer scipy.signal.ShortTimeFFT towards producing an output like scipy.signal.stft (see Comparison with Legacy Implementation in https://docs.scipy.org/doc/scipy/tutorial/signal.html#tutorial-stft), but I didnt manage to do so yet..