Stable Audio Open is a free AI model that can generate audio samples, sound effects, and production elements from text descriptions. The open source model is designed for sound designers, musicians, and creative professionals.
Stability AI, the company that made the popular AI image generator Stable Diffusion, has released Stable Audio Open, an open-source model for generating audio data. The AI model can create high-quality audio samples up to 47 seconds long from simple text prompts.
The model is trained to generate drum beats, instrument riffs, ambient sounds, foley recordings, and other audio elements for music production and sound design.
Stable Audio Open aims to show the potential of generative AI for sound design while ensuring responsible development with creative communities, Stability AI claims. Audio data from FreeSound and the Free Music Archive was used to train Stable Audio Open in order to protect creators' rights.
To get started, you can download the Stable Audio Open model on Hugging Face. The open-source release also lets users change and customize the model with their own audio data. Stability AI wants sound designers, musicians, developers, and audio fans to download the model and provide feedback.
Unlike the paid product Stable Audio 2, which can make full songs up to three minutes long, Stable Audio Open focuses on shorter audio samples and sound effects. It's not designed for generating full songs, melodies, or vocals.