Enhancing Low-Resource ASR through Versatile TTS: Bridging the Data Gap

TL;DR

This paper demonstrates that using versatile TTS models to generate synthetic speech data significantly improves low-resource ASR performance, especially by analyzing factors like text and speaker diversity.

Contribution

It introduces a novel approach of leveraging powerful TTS models for data augmentation in low-resource ASR, including the first study of text diversity's impact.

Findings

Consistent performance improvements across various low-resource datasets.

Text diversity in synthesized data significantly enhances ASR accuracy.

Analyzing factors like speaker diversity and data volume informs effective augmentation strategies.

Abstract

While automatic speech recognition (ASR) systems have achieved remarkable performance with large-scale datasets, their efficacy remains inadequate in low-resource settings, encompassing dialects, accents, minority languages, and long-tail hotwords, domains with significant practical relevance. With the advent of versatile and powerful text-to-speech (TTS) models, capable of generating speech with human-level naturalness, expressiveness, and diverse speaker profiles, leveraging TTS for ASR data augmentation provides a cost-effective and practical approach to enhancing ASR performance. Comprehensive experiments on an unprecedentedly rich variety of low-resource datasets demonstrate consistent and substantial performance improvements, proving that the proposed method of enhancing low-resource ASR through a versatile TTS model is highly effective and has broad application prospects. Furthermore, we delve deeper into key characteristics of synthesized speech data that contribute to ASR improvement, examining factors such as text diversity, speaker diversity, and the volume of synthesized data, with text diversity being studied for the first time in this work. We hope our findings provide helpful guidance and reference for the practical application of TTS-based data augmentation and push the advancement of low-resource ASR one step further.