Medium Voice & Audio · 1 min read
Stable Audio 2.0: stereo tracks up to 3 minutes and audio-to-audio
In one sentence Stability AI launches Stable Audio 2.0: 44.1 kHz stereo tracks up to three minutes with coherent musical structure, prompt-driven transformation of uploaded audio, and a diffusion-transformer architecture.
Reading level
Stable Audio 2.0 generates full music tracks up to three minutes long, with coherent musical structure and 44.1 kHz stereo audio.
The headline addition is audio-to-audio: you upload your own audio sample and transform it by describing the result you want in words. The model also produces sound effects and supports style transfer.
At launch it is free on the Stable Audio website; Stability AI says it will soon also be available via the API.
Companies
Stability AI
Tools
Stable Audio 2.0
Tags
Stability AIStable AudioMusic GenerationAudio DiffusionStereoGenerative Audio
Sources