Latent Filling: Latent Space Data Augmentation for Zero-shot Speech Synthesis

TL;DR

This paper introduces a latent space data augmentation method called Latent Filling (LF) for zero-shot speech synthesis, which improves speaker similarity without degrading speech quality by augmenting in the speaker embedding space.

Contribution

The paper proposes a novel latent space data augmentation technique for ZS-TTS that enhances speaker similarity without additional training stages.

Findings

LF significantly improves speaker similarity.

LF preserves speech quality.

Seamless integration into existing ZS-TTS systems.

Abstract

Previous works in zero-shot text-to-speech (ZS-TTS) have attempted to enhance its systems by enlarging the training data through crowd-sourcing or augmenting existing speech data. However, the use of low-quality data has led to a decline in the overall system performance. To avoid such degradation, instead of directly augmenting the input data, we propose a latent filling (LF) method that adopts simple but effective latent space data augmentation in the speaker embedding space of the ZS-TTS system. By incorporating a consistency loss, LF can be seamlessly integrated into existing ZS-TTS systems without the need for additional training stages. Experimental results show that LF significantly improves speaker similarity while preserving speech quality.