StarGAN-VC: Non-parallel many-to-many voice conversion with star generative adversarial networks

Hirokazu Kameoka, Takuhiro Kaneko, Kou Tanaka, Nobukatsu Hojo
6/6/2018
cs.SDcs.LGeess.AS

Official code: kamepong/StarGAN-VC (23 stars, last commit 2021-12-14).

Abstract

This paper proposes a method that allows non-parallel many-to-many voice conversion (VC) by using a variant of a generative adversarial network (GAN) called StarGAN. Our method, which we call StarGAN-VC, is noteworthy in that it (1) requires no parallel utterances, transcriptions, or time alignment procedures for speech generator training, (2) simultaneously learns many-to-many mappings across different attribute domains using a single generator network, (3) is able to generate converted speech signals quickly enough to allow real-time implementations and (4) requires only several minutes of training examples to generate reasonably realistic-sounding speech. Subjective evaluation experiments on a non-parallel many-to-many speaker identity conversion task revealed that the proposed method obtained higher sound quality and speaker similarity than a state-of-the-art method based on variational autoencoding GANs.

View on arXivView PDF

Code Implementations(1)

kamepong/StarGAN-VCOfficial100%
233Shell, PythonDec 2, 20214 years ago

Similar Papers

AUTOVC: Zero-Shot Voice Style Transfer with Only Autoencoder Loss

88% match1 repo5232019

SLMGAN: Exploiting Speech Language Model Representations for Unsupervised Zero-Shot Voice Conversion in GANs

83% match1 repo162023

Multi-modal Adversarial Training for Zero-Shot Voice Cloning

82% match2024

High Fidelity Speech Synthesis with Adversarial Networks

81% match1 repo1302019

HiFi-GAN: Generative Adversarial Networks for Efficient and High Fidelity Speech Synthesis

81% match1 repo2.4k2020

BigVGAN: A Universal Neural Vocoder with Large-Scale Training

81% match1 repo1.2k2022

Cite this paper

@article{kameoka2018starganvc,
  title  = {StarGAN-VC: Non-parallel many-to-many voice conversion with star generative adversarial networks},
  author = {Hirokazu Kameoka and Takuhiro Kaneko and Kou Tanaka and Nobukatsu Hojo},
  year   = {2018},
  eprint = {1806.02169},
  archivePrefix = {arXiv},
  url    = {https://arxiv.org/abs/1806.02169}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/1806.02169.svg)](https://codeofpaper.com/paper/1806.02169)

Discussion