Skip to Main content Skip to Navigation
Conference papers

CycleGAN Voice Conversion of Spectral Envelopes using Adversarial Weights

Rafael Ferro 1 Nicolas Obin 1 Axel Roebel 2
1 Analyse et synthèse sonores [Paris]
STMS - Sciences et Technologies de la Musique et du Son
2 Analyse et synthèse sonores [Paris]
STMS - Sciences et Technologies de la Musique et du Son
Abstract : This paper tackles GAN optimization and stability issues in the context of voice conversion. First, to simplify the conversion task, we propose to use spectral envelopes as inputs. Second we propose two adversarial weight training paradigms, the generalized weighted GAN and the generator impact GAN, both aim at reducing the impact of the generator on the discriminator, so both can learn more gradually and efficiently during training. Applying an energy constraint to the cycleGAN paradigm considerably improved conversion quality. A subjective experiment conducted on a voice conversion task on the voice conversion challenge 2018 dataset shows first that despite a significantly reduced network complexity, the proposed method achieves state-of-the-art results, and second that the proposed weighted GAN methods outperform a previously proposed one.
Complete list of metadatas

Cited literature [26 references]  Display  Hide  Download

https://hal.archives-ouvertes.fr/hal-02929245
Contributor : Nicolas Obin <>
Submitted on : Thursday, September 3, 2020 - 12:00:48 PM
Last modification on : Thursday, September 10, 2020 - 3:16:20 AM

File

1910.12614.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-02929245, version 1

Citation

Rafael Ferro, Nicolas Obin, Axel Roebel. CycleGAN Voice Conversion of Spectral Envelopes using Adversarial Weights. Eusipco, Jan 2021, Amsterdam, Netherlands. ⟨hal-02929245⟩

Share

Metrics

Record views

15

Files downloads

18