OZSpeech: One-step Zero-shot Speech Synthesis with Learned-Prior-Conditioned Flow Matching

TL;DR

OZSpeech introduces a novel one-step zero-shot speech synthesis method using learned-prior-conditioned flow matching, improving efficiency and attribute modeling for more accurate and natural speech generation.

Contribution

It is the first TTS approach to utilize optimal transport conditional flow matching with one-step sampling and a learned prior, enhancing attribute control and reducing computational costs.

Findings

Achieves high content accuracy and naturalness in speech synthesis.

Effectively preserves speaker style and prosody.

Reduces sampling steps compared to previous flow-based TTS methods.

Abstract

Text-to-speech (TTS) systems have seen significant advancements in recent years, driven by improvements in deep learning and neural network architectures. Viewing the output speech as a data distribution, previous approaches often employ traditional speech representations, such as waveforms or spectrograms, within the Flow Matching framework. However, these methods have limitations, including overlooking various speech attributes and incurring high computational costs due to additional constraints introduced during training. To address these challenges, we introduce OZSpeech, the first TTS method to explore optimal transport conditional flow matching with one-step sampling and a learned prior as the condition, effectively disregarding preceding states and reducing the number of sampling steps. Our approach operates on disentangled, factorized components of speech in token format, enabling accurate modeling of each speech attribute, which enhances the TTS system's ability to precisely clone the prompt speech. Experimental results show that our method achieves promising performance over existing methods in content accuracy, naturalness, prosody generation, and speaker style preservation. Audio samples are available at our demo page https://ozspeech.github.io/OZSpeech_Web/.