A Style Transfer Approach to Source Separation

Training neural networks for source separation involves presenting a mixture recording at the input of the network and updating network parameters in order to produce an output that resembles the clean source. Consequently, supervised source separation depends on the availability of paired mixture-c...

Full description

Saved in:
Bibliographic Details
Published in2019 IEEE Workshop on Applications of Signal Processing to Audio and Acoustics (WASPAA) pp. 170 - 174
Main Authors Venkataramani, Shrikant, Tzinis, Efthymios, Smaragdis, Paris
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.10.2019
Subjects
consistency loss
deep learning
domain translation
Gallium nitride
Generative adversarial networks
Music
Neural networks
Source separation
Spectrogram
Style transfer
Training
unsupervised learning
Online AccessGet full text

Cover

More Information
Summary:Training neural networks for source separation involves presenting a mixture recording at the input of the network and updating network parameters in order to produce an output that resembles the clean source. Consequently, supervised source separation depends on the availability of paired mixture-clean training examples. In this paper, we interpret source separation as a style transfer problem. We present a variational auto-encoder network that exploits the commonality across the domain of mixtures and the domain of clean sounds and learns a shared latent representation across the two domains. Using these cycle-consistent variational auto-encoders, we learn a mapping from the mixture domain to the domain of clean sounds and perform source separation without explicitly supervising with paired training examples.
ISSN:1947-1629
DOI:10.1109/WASPAA.2019.8937203