Active Learning for Text-to-Speech Synthesis with Informative Sample Collection

TL;DR

This paper introduces an active learning approach for building efficient text-to-speech datasets by iteratively selecting informative samples, resulting in higher-quality speech synthesis with less data.

Contribution

The paper presents a novel active learning-based method for TTS corpus construction that improves data efficiency and speech quality over traditional approaches.

Findings

Constructed corpus yields higher-quality speech synthesis.

Method reduces data requirements for high-quality TTS.

Iterative sample collection enhances model training effectiveness.

Abstract

The construction of high-quality datasets is a cornerstone of modern text-to-speech (TTS) systems. However, the increasing scale of available data poses significant challenges, including storage constraints. To address these issues, we propose a TTS corpus construction method based on active learning. Unlike traditional feed-forward and model-agnostic corpus construction approaches, our method iteratively alternates between data collection and model training, thereby focusing on acquiring data that is more informative for model improvement. This approach enables the construction of a data-efficient corpus. Experimental results demonstrate that the corpus constructed using our method enables higher-quality speech synthesis than corpora of the same size.