CosyVoice 3: Towards In-the-wild Speech Generation via Scaling-up and Post-training
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866913860880957440 |
|---|---|
| author | Du, Zhihao Gao, Changfeng Wang, Yuxuan Yu, Fan Zhao, Tianyu Wang, Hao Lv, Xiang Wang, Hui Ni, Chongjia Shi, Xian An, Keyu Yang, Guanrou Li, Yabin Chen, Yanni Gao, Zhifu Chen, Qian Gu, Yue Chen, Mengzhe Chen, Yafeng Zhang, Shiliang Wang, Wen Ye, Jieping |
| author_facet | Du, Zhihao Gao, Changfeng Wang, Yuxuan Yu, Fan Zhao, Tianyu Wang, Hao Lv, Xiang Wang, Hui Ni, Chongjia Shi, Xian An, Keyu Yang, Guanrou Li, Yabin Chen, Yanni Gao, Zhifu Chen, Qian Gu, Yue Chen, Mengzhe Chen, Yafeng Zhang, Shiliang Wang, Wen Ye, Jieping |
| contents | 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. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2505_17589 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | CosyVoice 3: Towards In-the-wild Speech Generation via Scaling-up and Post-training Du, Zhihao Gao, Changfeng Wang, Yuxuan Yu, Fan Zhao, Tianyu Wang, Hao Lv, Xiang Wang, Hui Ni, Chongjia Shi, Xian An, Keyu Yang, Guanrou Li, Yabin Chen, Yanni Gao, Zhifu Chen, Qian Gu, Yue Chen, Mengzhe Chen, Yafeng Zhang, Shiliang Wang, Wen Ye, Jieping Sound Artificial Intelligence Audio and Speech Processing 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. |
| title | CosyVoice 3: Towards In-the-wild Speech Generation via Scaling-up and Post-training |
| topic | Sound Artificial Intelligence Audio and Speech Processing |
| url | https://arxiv.org/abs/2505.17589 |