Word-Level Style Control for Expressive, Non-attentive Speech Synthesis

TL;DR

This paper introduces a novel non-attentive speech synthesis model that enables word-level style control and prosody transfer by disentangling style and phonetic information using dual encoders and a prior model.

Contribution

It proposes a new architecture with dual encoders for style and phonetic representations, allowing fine-grained style control at the word level in speech synthesis.

Findings

Model achieves word-level and global style control.

Enables prosody transfer without reference utterance.

Disentangles style from phonetic content effectively.

Abstract

This paper presents an expressive speech synthesis architecture for modeling and controlling the speaking style at a word level. It attempts to learn word-level stylistic and prosodic representations of the speech data, with the aid of two encoders. The first one models style by finding a combination of style tokens for each word given the acoustic features, and the second outputs a word-level sequence conditioned only on the phonetic information in order to disentangle it from the style information. The two encoder outputs are aligned and concatenated with the phoneme encoder outputs and then decoded with a Non-Attentive Tacotron model. An extra prior encoder is used to predict the style tokens autoregressively, in order for the model to be able to run without a reference utterance. We find that the resulting model gives both word-level and global control over the style, as well as prosody transfer capabilities.