Towards Low-Resource StarGAN Voice Conversion using Weight Adaptive Instance Normalization

TL;DR

This paper introduces a novel low-resource StarGAN-based voice conversion model that employs weight adaptive instance normalization to improve data efficiency and performance across many speakers with limited training data.

Contribution

The work proposes a new model using speaker embeddings and adaptive weight normalization to enhance many-to-many voice conversion in low-resource scenarios.

Findings

Outperforms baseline methods in objective evaluations.

Achieves higher naturalness and similarity in subjective tests.

Effective with as few as 5 samples per speaker.

Abstract

Many-to-many voice conversion with non-parallel training data has seen significant progress in recent years. StarGAN-based models have been interests of voice conversion. However, most of the StarGAN-based methods only focused on voice conversion experiments for the situations where the number of speakers was small, and the amount of training data was large. In this work, we aim at improving the data efficiency of the model and achieving a many-to-many non-parallel StarGAN-based voice conversion for a relatively large number of speakers with limited training samples. In order to improve data efficiency, the proposed model uses a speaker encoder for extracting speaker embeddings and conducts adaptive instance normalization (AdaIN) on convolutional weights. Experiments are conducted with 109 speakers under two low-resource situations, where the number of training samples is 20 and 5 per speaker. An objective evaluation shows the proposed model is better than the baseline methods. Furthermore, a subjective evaluation shows that, for both naturalness and similarity, the proposed model outperforms the baseline method.