NaturalSpeech 2: Latent Diffusion Models are Natural and Zero-Shot Speech and Singing Synthesizers

TL;DR

NaturalSpeech 2 introduces a diffusion-based TTS system that leverages neural audio codecs and zero-shot learning to produce high-quality, diverse speech and singing synthesis, outperforming previous models on large-scale datasets.

Contribution

It presents a novel diffusion-based TTS framework with a speech prompting mechanism for zero-shot speech and singing synthesis, scaling to 44K hours of data.

Findings

Outperforms previous TTS systems in prosody and voice quality

Achieves high-quality zero-shot singing synthesis with only a speech prompt

Demonstrates robustness and diversity in large-scale datasets

Abstract

Scaling text-to-speech (TTS) to large-scale, multi-speaker, and in-the-wild datasets is important to capture the diversity in human speech such as speaker identities, prosodies, and styles (e.g., singing). Current large TTS systems usually quantize speech into discrete tokens and use language models to generate these tokens one by one, which suffer from unstable prosody, word skipping/repeating issue, and poor voice quality. In this paper, we develop NaturalSpeech 2, a TTS system that leverages a neural audio codec with residual vector quantizers to get the quantized latent vectors and uses a diffusion model to generate these latent vectors conditioned on text input. To enhance the zero-shot capability that is important to achieve diverse speech synthesis, we design a speech prompting mechanism to facilitate in-context learning in the diffusion model and the duration/pitch predictor. We scale NaturalSpeech 2 to large-scale datasets with 44K hours of speech and singing data and evaluate its voice quality on unseen speakers. NaturalSpeech 2 outperforms previous TTS systems by a large margin in terms of prosody/timbre similarity, robustness, and voice quality in a zero-shot setting, and performs novel zero-shot singing synthesis with only a speech prompt. Audio samples are available at https://speechresearch.github.io/naturalspeech2.