Statistical Parametric Speech Synthesis Using Bottleneck Representation From Sequence Auto-encoder

TL;DR

This paper introduces a unit-level acoustic representation using a recurrent neural network auto-encoder for statistical parametric speech synthesis, reducing computational cost while maintaining synthesis quality.

Contribution

It proposes a novel fixed-dimensional unit-level acoustic representation that enables efficient speech synthesis with comparable quality to traditional methods.

Findings

Achieves speech synthesis quality similar to conventional methods.

Reduces computational requirements significantly.

Demonstrates effectiveness of auto-encoder based unit representations.

Abstract

In this paper, we describe a statistical parametric speech synthesis approach with unit-level acoustic representation. In conventional deep neural network based speech synthesis, the input text features are repeated for the entire duration of phoneme for mapping text and speech parameters. This mapping is learnt at the frame-level which is the de-facto acoustic representation. However much of this computational requirement can be drastically reduced if every unit can be represented with a fixed-dimensional representation. Using recurrent neural network based auto-encoder, we show that it is indeed possible to map units of varying duration to a single vector. We then use this acoustic representation at unit-level to synthesize speech using deep neural network based statistical parametric speech synthesis technique. Results show that the proposed approach is able to synthesize at the same quality as the conventional frame based approach at a highly reduced computational cost.