Skip to main content

Learning Bandwidth Expansion Using Perceptually-motivated Loss

Author(s): Feng, Berthy; Jin, Zeyu; Su, Jiaqi; Finkelstein, Adam

To refer to this page use:
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.
Publication Date: 2019
Citation: Feng, Berthy, Zeyu Jin, Jiaqi Su, and Adam Finkelstein. "Learning Bandwidth Expansion Using Perceptually-motivated Loss." IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) (2019): pp. 606-610. doi:10.1109/ICASSP.2019.8682367
DOI: 10.1109/ICASSP.2019.8682367
ISSN: 1520-6149
EISSN: 2379-190X
Pages: 606 - 610
Type of Material: Conference Article
Journal/Proceeding Title: IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)
Version: Author's manuscript

Items in OAR@Princeton are protected by copyright, with all rights reserved, unless otherwise indicated.