RWEN-TTS: Relation-aware Word Encoding Network for Natural Text-to-Speech Synthesis

TL;DR

RWEN-TTS introduces a relation-aware encoding network that effectively incorporates syntactic, semantic, and adjacency information to significantly improve naturalness in text-to-speech synthesis.

Contribution

It proposes a novel relation-aware encoding network with modules for semantic and adjacent word relations, addressing limitations of previous TTS models.

Findings

Significant improvement over previous TTS models.

Effective utilization of syntactic and semantic information.

Enhanced naturalness and expressiveness in synthesized speech.

Abstract

With the advent of deep learning, a huge number of text-to-speech (TTS) models which produce human-like speech have emerged. Recently, by introducing syntactic and semantic information w.r.t the input text, various approaches have been proposed to enrich the naturalness and expressiveness of TTS models. Although these strategies showed impressive results, they still have some limitations in utilizing language information. First, most approaches only use graph networks to utilize syntactic and semantic information without considering linguistic features. Second, most previous works do not explicitly consider adjacent words when encoding syntactic and semantic information, even though it is obvious that adjacent words are usually meaningful when encoding the current word. To address these issues, we propose Relation-aware Word Encoding Network (RWEN), which effectively allows syntactic and semantic information based on two modules (i.e., Semantic-level Relation Encoding and Adjacent Word Relation Encoding). Experimental results show substantial improvements compared to previous works.