Parallel Tacotron: Non-Autoregressive and Controllable TTS

TL;DR

Parallel Tacotron is a non-autoregressive TTS model that enhances efficiency and naturalness through variational autoencoders, lightweight convolutions, and iterative spectrogram loss, achieving comparable quality to autoregressive models with faster inference.

Contribution

It introduces a highly parallelizable non-autoregressive TTS model with variational autoencoder-based residual encoder and iterative refinement techniques, improving naturalness and efficiency.

Findings

Matches autoregressive baseline in subjective evaluations

Significantly reduces inference time

Utilizes variational autoencoder for better naturalness

Abstract

Although neural end-to-end text-to-speech models can synthesize highly natural speech, there is still room for improvements to its efficiency and naturalness. This paper proposes a non-autoregressive neural text-to-speech model augmented with a variational autoencoder-based residual encoder. This model, called \emph{Parallel Tacotron}, is highly parallelizable during both training and inference, allowing efficient synthesis on modern parallel hardware. The use of the variational autoencoder relaxes the one-to-many mapping nature of the text-to-speech problem and improves naturalness. To further improve the naturalness, we use lightweight convolutions, which can efficiently capture local contexts, and introduce an iterative spectrogram loss inspired by iterative refinement. Experimental results show that Parallel Tacotron matches a strong autoregressive baseline in subjective evaluations with significantly decreased inference time.