MOSS-VoiceGenerator: Create Realistic Voices with Natural Language Descriptions
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , , , , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2026
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866914431527550976 |
|---|---|
| author | Huang, Kexin Fan, Liwei Jiang, Botian Jiang, Yaozhou Tu, Qian Zhu, Jie Zhang, Yuqian Zhao, Yiwei Yang, Chenchen Fei, Zhaoye Li, Shimin Yang, Xiaogui Cheng, Qinyuan Qiu, Xipeng |
| author_facet | Huang, Kexin Fan, Liwei Jiang, Botian Jiang, Yaozhou Tu, Qian Zhu, Jie Zhang, Yuqian Zhao, Yiwei Yang, Chenchen Fei, Zhaoye Li, Shimin Yang, Xiaogui Cheng, Qinyuan Qiu, Xipeng |
| contents | Voice design from natural language aims to generate speaker timbres directly from free-form textual descriptions, allowing users to create voices tailored to specific roles, personalities, and emotions. Such controllable voice creation benefits a wide range of downstream applications-including storytelling, game dubbing, role-play agents, and conversational assistants, making it a significant task for modern Text-to-Speech models. However, existing models are largely trained on carefully recorded studio data, which produces speech that is clean and well-articulated, yet lacks the lived-in qualities of real human voices. To address these limitations, we present MOSS-VoiceGenerator, an open-source instruction-driven voice generation model that creates new timbres directly from natural language prompts. Motivated by the hypothesis that exposure to real-world acoustic variation produces more perceptually natural voices, we train on large-scale expressive speech data sourced from cinematic content. Subjective preference studies demonstrate its superiority in overall performance, instruction-following, and naturalness compared to other voice design models. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_28086 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | MOSS-VoiceGenerator: Create Realistic Voices with Natural Language Descriptions Huang, Kexin Fan, Liwei Jiang, Botian Jiang, Yaozhou Tu, Qian Zhu, Jie Zhang, Yuqian Zhao, Yiwei Yang, Chenchen Fei, Zhaoye Li, Shimin Yang, Xiaogui Cheng, Qinyuan Qiu, Xipeng Sound Artificial Intelligence Computation and Language Voice design from natural language aims to generate speaker timbres directly from free-form textual descriptions, allowing users to create voices tailored to specific roles, personalities, and emotions. Such controllable voice creation benefits a wide range of downstream applications-including storytelling, game dubbing, role-play agents, and conversational assistants, making it a significant task for modern Text-to-Speech models. However, existing models are largely trained on carefully recorded studio data, which produces speech that is clean and well-articulated, yet lacks the lived-in qualities of real human voices. To address these limitations, we present MOSS-VoiceGenerator, an open-source instruction-driven voice generation model that creates new timbres directly from natural language prompts. Motivated by the hypothesis that exposure to real-world acoustic variation produces more perceptually natural voices, we train on large-scale expressive speech data sourced from cinematic content. Subjective preference studies demonstrate its superiority in overall performance, instruction-following, and naturalness compared to other voice design models. |
| title | MOSS-VoiceGenerator: Create Realistic Voices with Natural Language Descriptions |
| topic | Sound Artificial Intelligence Computation and Language |
| url | https://arxiv.org/abs/2603.28086 |