The NU Voice Conversion System for the Voice Conversion Challenge 2020: On the Effectiveness of Sequence-to-sequence Models and Autoregressive Neural Vocoders

TL;DR

This paper evaluates the effectiveness of sequence-to-sequence models and autoregressive neural vocoders in voice conversion, demonstrating improvements in speech naturalness and similarity through experimental comparisons.

Contribution

It introduces NU's voice conversion systems utilizing seq2seq models and AR vocoders, highlighting their impact on speech quality and similarity in VC tasks.

Findings

Seq2seq models improve conversion similarity.

AR vocoders enhance speech naturalness.

Systems outperform baseline models.

Abstract

In this paper, we present the voice conversion (VC) systems developed at Nagoya University (NU) for the Voice Conversion Challenge 2020 (VCC2020). We aim to determine the effectiveness of two recent significant technologies in VC: sequence-to-sequence (seq2seq) models and autoregressive (AR) neural vocoders. Two respective systems were developed for the two tasks in the challenge: for task 1, we adopted the Voice Transformer Network, a Transformer-based seq2seq VC model, and extended it with synthetic parallel data to tackle nonparallel data; for task 2, we used the frame-based cyclic variational autoencoder (CycleVAE) to model the spectral features of a speech waveform and the AR WaveNet vocoder with additional fine-tuning. By comparing with the baseline systems, we confirmed that the seq2seq modeling can improve the conversion similarity and that the use of AR vocoders can improve the naturalness of the converted speech.