High quality, lightweight and adaptable TTS using LPCNet

TL;DR

This paper introduces a lightweight, high-quality neural TTS system based on LPCNet that is fast, adaptable to new voices with minimal data, and capable of near-natural speech synthesis.

Contribution

The paper presents a modular neural TTS system combining prosody prediction, acoustic features, and LPCNet vocoder, enabling fast, adaptable, and high-quality speech synthesis.

Findings

Synthesizes speech close to natural quality.

Runs 3 times faster than real-time on CPU.

Effectively adapts to unseen voices with minimal data.

Abstract

We present a lightweight adaptable neural TTS system with high quality output. The system is composed of three separate neural network blocks: prosody prediction, acoustic feature prediction and Linear Prediction Coding Net as a neural vocoder. This system can synthesize speech with close to natural quality while running 3 times faster than real-time on a standard CPU. The modular setup of the system allows for simple adaptation to new voices with a small amount of data. We first demonstrate the ability of the system to produce high quality speech when trained on large, high quality datasets. Following that, we demonstrate its adaptability by mimicking unseen voices using 5 to 20 minutes long datasets with lower recording quality. Large scale Mean Opinion Score quality and similarity tests are presented, showing that the system can adapt to unseen voices with quality gap of 0.12 and similarity gap of 3% compared to natural speech for male voices and quality gap of 0.35 and similarity of gap of 9 % for female voices.