Expressive-VC: Highly Expressive Voice Conversion with Attention Fusion of Bottleneck and Perturbation Features

TL;DR

Expressive-VC is a novel end-to-end voice conversion framework that combines neural bottleneck and perturbation features with attention fusion, achieving high expressiveness, speaker similarity, and intelligibility in converted speech.

Contribution

It introduces a new fusion approach using attention mechanism to combine linguistic and para-linguistic features for expressive voice conversion.

Findings

Outperforms state-of-the-art systems in expressiveness and speaker similarity.

Maintains high intelligibility in converted speech.

Effectively captures source expressiveness while preserving target speaker identity.

Abstract

Voice conversion for highly expressive speech is challenging. Current approaches struggle with the balancing between speaker similarity, intelligibility and expressiveness. To address this problem, we propose Expressive-VC, a novel end-to-end voice conversion framework that leverages advantages from both neural bottleneck feature (BNF) approach and information perturbation approach. Specifically, we use a BNF encoder and a Perturbed-Wav encoder to form a content extractor to learn linguistic and para-linguistic features respectively, where BNFs come from a robust pre-trained ASR model and the perturbed wave becomes speaker-irrelevant after signal perturbation. We further fuse the linguistic and para-linguistic features through an attention mechanism, where speaker-dependent prosody features are adopted as the attention query, which result from a prosody encoder with target speaker embedding and normalized pitch and energy of source speech as input. Finally the decoder consumes the integrated features and the speaker-dependent prosody feature to generate the converted speech. Experiments demonstrate that Expressive-VC is superior to several state-of-the-art systems, achieving both high expressiveness captured from the source speech and high speaker similarity with the target speaker; meanwhile intelligibility is well maintained.