Microsoft Releases VibeVoice for Long-Form Audio
The new 1.5-billion-parameter text-to-speech model is designed to generate natural, multi-speaker audio for podcasts and other long-form content.
Microsoft has released VibeVoice-1.5B, a new open-source model aimed at generating high-quality, long-form speech. At 1.5 billion parameters, it's a notable new entry in the text-to-speech (TTS) landscape, focusing on a particularly challenging area: creating natural-sounding, multi-speaker conversations.
The model is specifically designed to produce audio that mimics the style of podcasts. It supports both English and Chinese, making it versatile for a wide range of applications. Importantly, VibeVoice is released under a permissive MIT license, which allows for broad use in both research and commercial projects without significant restrictions.
Key Capabilities
- Long-form Generation: Capable of producing extended audio clips beyond typical short sentences.
- Multi-speaker Support: Can synthesize conversations involving different voices.
- Bilingual: Supports both English and Chinese text input.
- Permissive Licensing: Released under the MIT license, encouraging wide adoption.
The release of VibeVoice matters because it provides a strong open-source alternative for creating sophisticated audio content that has often been the domain of proprietary services. Developers and creators can now experiment with generating entire podcast episodes, dynamic audiobooks, or more complex conversational agents. You can find the model and usage instructions on its Hugging Face repository.
Sources
- Visit
microsoft/VibeVoice-1.5B
Hugging Face
More from Microsoft
All Microsoft releases →
Microsoft's Mage-VL-Cowork takes on computer-use tasks
A new vision-language agent model aims to operate GUIs, write code, and work the terminal.
Microsoft Releases FrogNano-4B Under MIT License
The compact 4-billion-parameter text model lands on Hugging Face with a permissive license aimed at developers.
Microsoft's VibeVoice ASR brings streaming speech-to-text
A 7-billion-parameter model targets real-time, multilingual transcription with an open release on Hugging Face.
More in Text → Speech
All Text → Speech →
Liquid AI's LFM3-Omni-2B brings any-to-any AI to the edge
The 2-billion-parameter model combines vision, audio and speech-to-speech in a package small enough to run on device.
IBM's Granite Switch 4.1 Unifies Chat, Logic, Speech
A single endpoint now routes text conversation, decision-making, and speech synthesis across IBM's Apache-licensed Granite line.
Fish Audio's Drama 3 brings MoE to open TTS
The open-source speech model pairs a 35B mixture-of-experts backbone with voice cloning and directed performance control.
0 comments
No comments yet. Be the first to weigh in.