A Vector Quantized Approach for Text to Speech Synthesis on Real-World Spontaneous Speech

TL;DR

This paper presents MQTTS, a novel vector quantized TTS system trained on real-world data, which effectively handles speech diversity and alignment issues, outperforming existing systems in quality.

Contribution

The work introduces a vector quantized approach with multiple code groups and monotonic alignment for real-world data TTS, addressing alignment mismatch problems.

Findings

MQTTS outperforms existing TTS systems in objective measures.

Using discrete codes improves speech intelligibility.

Silence prompts enhance synthesis quality.

Abstract

Recent Text-to-Speech (TTS) systems trained on reading or acted corpora have achieved near human-level naturalness. The diversity of human speech, however, often goes beyond the coverage of these corpora. We believe the ability to handle such diversity is crucial for AI systems to achieve human-level communication. Our work explores the use of more abundant real-world data for building speech synthesizers. We train TTS systems using real-world speech from YouTube and podcasts. We observe the mismatch between training and inference alignments in mel-spectrogram based autoregressive models, leading to unintelligible synthesis, and demonstrate that learned discrete codes within multiple code groups effectively resolves this issue. We introduce our MQTTS system whose architecture is designed for multiple code generation and monotonic alignment, along with the use of a clean silence prompt to improve synthesis quality. We conduct ablation analyses to identify the efficacy of our methods. We show that MQTTS outperforms existing TTS systems in several objective and subjective measures.