Learning Bandwidth Expansion Using Perceptually-motivated Loss

Berthy Feng, Zeyu Jin, Jiaqi Su, Adam Finkelstein

Abstract: We introduce a perceptually motivated approach to bandwidth expansion for speech. Our method pairs a new 3-way split variant of the FFTNet neural vocoder structure with a perceptual loss function, combining objectives from both the time and frequency domains. Mean opinion score tests show that it outperforms baseline methods from both domains, even for extreme bandwidth expansion.