Learning Bandwidth Expansion Using Perceptually-Motivated Loss

Published in ICASSP, 2019

Recommended citation: Berthy Feng, Zeyu Jin, Jiaqi Su, and Adam Finkelstein. "Learning Bandwidth Expansion Using Perceptually-Motivated Loss." ICASSP, May 2019. https://pixl.cs.princeton.edu/pubs/Feng_2019_LBE/Feng_2019_bwe.pdf

We introduce a perceptually motivated approach to bandwidth expansion for speech. Our method pairs a new three-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.

Project page