Learning Bandwidth Expansion Using Perceptually-motivated Loss

Berthy Feng; Zeyu Jin; Jiaqi Su; Adam Finkelstein

ICASSP 2019accepted0 citations

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.

BibTeX

@inproceedings{icassp2019_learningbandwidt,
  title = {Learning Bandwidth Expansion Using Perceptually-motivated Loss},
  author = {Berthy Feng and Zeyu Jin and Jiaqi Su and Adam Finkelstein},
  booktitle = {ICASSP 2019},
  year = {2019}
}