Distribution augmentation for low-resource expressive text-to-speech

TL;DR

This paper introduces a data augmentation method for low-resource expressive TTS that syntactically preserves text and audio fragments, enhancing diversity, reducing overfitting, and improving speech quality and robustness.

Contribution

The novel augmentation technique generates new training examples without extra data, significantly benefiting low-resource TTS by increasing diversity and robustness.

Findings

Improves speech quality across multiple datasets and architectures.

Enhances robustness of attention-based TTS models.

Reduces overfitting in low-resource settings.

Abstract

This paper presents a novel data augmentation technique for text-to-speech (TTS), that allows to generate new (text, audio) training examples without requiring any additional data. Our goal is to increase diversity of text conditionings available during training. This helps to reduce overfitting, especially in low-resource settings. Our method relies on substituting text and audio fragments in a way that preserves syntactical correctness. We take additional measures to ensure that synthesized speech does not contain artifacts caused by combining inconsistent audio samples. The perceptual evaluations show that our method improves speech quality over a number of datasets, speakers, and TTS architectures. We also demonstrate that it greatly improves robustness of attention-based TTS models.