CosyVoice 3: Towards In-the-wild Speech Generation via Scaling-up and Post-training

TL;DR

CosyVoice 3 advances in in-the-wild multilingual speech synthesis by scaling data and model size, introducing a new speech tokenizer, and a differentiable reward model, achieving improved naturalness, consistency, and speaker similarity.

Contribution

The paper presents CosyVoice 3, a novel multilingual speech synthesis model with a new speech tokenizer and post-training reward model, trained on a vastly larger dataset and larger model size.

Findings

Enhanced speech naturalness and prosody in multilingual synthesis.

Improved content consistency and speaker similarity.

Successful scaling of data and model size for better performance.

Abstract

In our prior works, we introduced a scalable streaming speech synthesis model, CosyVoice 2, which integrates a large language model (LLM) and a chunk-aware flow matching (FM) model, and achieves low-latency bi-streaming speech synthesis and human-parity quality. Despite these advancements, CosyVoice 2 exhibits limitations in language coverage, domain diversity, data volume, text formats, and post-training techniques. In this paper, we present CosyVoice 3, an improved model designed for zero-shot multilingual speech synthesis in the wild, surpassing its predecessor in content consistency, speaker similarity, and prosody naturalness. Key features of CosyVoice 3 include: 1) A novel speech tokenizer to improve prosody naturalness, developed via supervised multi-task training, including automatic speech recognition, speech emotion recognition, language identification, audio event detection, and speaker analysis. 2) A new differentiable reward model for post-training applicable not only to CosyVoice 3 but also to other LLM-based speech synthesis models. 3) Dataset Size Scaling: Training data is expanded from ten thousand hours to one million hours, encompassing 9 languages and 18 Chinese dialects across various domains and text formats. 4) Model Size Scaling: Model parameters are increased from 0.5 billion to 1.5 billion, resulting in enhanced performance on our multilingual benchmark due to the larger model capacity. These advancements contribute significantly to the progress of speech synthesis in the wild. We encourage readers to listen to the demo at https://funaudiollm.github.io/cosyvoice3.