Discrete Unit based Masking for Improving Disentanglement in Voice Conversion

TL;DR

This paper introduces a novel input masking technique that enhances speaker disentanglement in voice conversion systems by reducing phonetic dependency, leading to improved conversion quality across various models.

Contribution

The paper proposes a discrete unit masking mechanism at the input level that improves disentanglement and is compatible with any encoder-decoder voice conversion framework.

Findings

44% relative improvement in objective intelligibility

Enhanced disentanglement in attention-based VC methods

Applicable to multiple VC frameworks

Abstract

Voice conversion (VC) aims to modify the speaker's identity while preserving the linguistic content. Commonly, VC methods use an encoder-decoder architecture, where disentangling the speaker's identity from linguistic information is crucial. However, the disentanglement approaches used in these methods are limited as the speaker features depend on the phonetic content of the utterance, compromising disentanglement. This dependency is amplified with attention-based methods. To address this, we introduce a novel masking mechanism in the input before speaker encoding, masking certain discrete speech units that correspond highly with phoneme classes. Our work aims to reduce the phonetic dependency of speaker features by restricting access to some phonetic information. Furthermore, since our approach is at the input level, it is applicable to any encoder-decoder based VC framework. Our approach improves disentanglement and conversion performance across multiple VC methods, showing significant effectiveness, particularly in attention-based method, with 44% relative improvement in objective intelligibility.