AdaSpeech 4: Adaptive Text to Speech in Zero-Shot Scenarios

TL;DR

AdaSpeech 4 is a zero-shot adaptive TTS system that enhances speaker generalization and quality by systematic speaker modeling, enabling high-quality voice synthesis without fine-tuning.

Contribution

The paper introduces a novel speaker modeling approach using basis vectors, conditional layer normalization, and a distribution-based supervision loss for improved zero-shot TTS.

Findings

Achieves better voice quality and similarity than baselines

Operates effectively without fine-tuning on new speakers

Demonstrates strong generalization across multiple datasets

Abstract

Adaptive text to speech (TTS) can synthesize new voices in zero-shot scenarios efficiently, by using a well-trained source TTS model without adapting it on the speech data of new speakers. Considering seen and unseen speakers have diverse characteristics, zero-shot adaptive TTS requires strong generalization ability on speaker characteristics, which brings modeling challenges. In this paper, we develop AdaSpeech 4, a zero-shot adaptive TTS system for high-quality speech synthesis. We model the speaker characteristics systematically to improve the generalization on new speakers. Generally, the modeling of speaker characteristics can be categorized into three steps: extracting speaker representation, taking this speaker representation as condition, and synthesizing speech/mel-spectrogram given this speaker representation. Accordingly, we improve the modeling in three steps: 1) To extract speaker representation with better generalization, we factorize the speaker characteristics into basis vectors and extract speaker representation by weighted combining of these basis vectors through attention. 2) We leverage conditional layer normalization to integrate the extracted speaker representation to TTS model. 3) We propose a novel supervision loss based on the distribution of basis vectors to maintain the corresponding speaker characteristics in generated mel-spectrograms. Without any fine-tuning, AdaSpeech 4 achieves better voice quality and similarity than baselines in multiple datasets.