I am studying how to remove noise from speech audio and I figured the most common error in spectral processing comes in the form of musical noise.
Now what are the techniques on reducing this musical noise from the desired speech signal? My intuition tells me it's all about having an accurate estimation of the real-time noise. But we know real-time noise is far from stationary.
Is smoothing possible for the spectral domain? Where I apply spectral subtraction gains but they are smoothed across adjacent bins?