Sample Efficient Adaptive Text-to-Speech

TL;DR

This paper introduces a meta-learning approach for adaptive text-to-speech that enables rapid speaker adaptation with minimal data, achieving state-of-the-art naturalness and voice similarity.

Contribution

It proposes three strategies for quick speaker adaptation in TTS using a multi-speaker WaveNet model, including embedding learning, fine-tuning, and embedding prediction.

Findings

Successful adaptation to new speakers with few minutes of data

Achieved state-of-the-art naturalness in speech synthesis

Improved voice similarity with minimal data

Abstract

We present a meta-learning approach for adaptive text-to-speech (TTS) with few data. During training, we learn a multi-speaker model using a shared conditional WaveNet core and independent learned embeddings for each speaker. The aim of training is not to produce a neural network with fixed weights, which is then deployed as a TTS system. Instead, the aim is to produce a network that requires few data at deployment time to rapidly adapt to new speakers. We introduce and benchmark three strategies: (i) learning the speaker embedding while keeping the WaveNet core fixed, (ii) fine-tuning the entire architecture with stochastic gradient descent, and (iii) predicting the speaker embedding with a trained neural network encoder. The experiments show that these approaches are successful at adapting the multi-speaker neural network to new speakers, obtaining state-of-the-art results in both sample naturalness and voice similarity with merely a few minutes of audio data from new speakers.