Stability AI unveils 'Stable Audio': A versatile platform for AI Music Generation

Stability AI has launched a new AI platform, Stable Audio, which uses a diffusion model similar to its AI Generated Art platform, Stable Diffusion XL. The platform, designed to generate songs or background audio, offers the ability to create audio of varying lengths, providing added versatility in music production. This feature is made possible by training the model on music and integrating text metadata for song commencement and termination times.

Stable Audio is the result of advanced research from Stability AI's generative audio research lab, Harmonai. The platform was developed using an extensive dataset of over 800,000 audio files, including music, sound effects, and single-instrument stems, totaling more than 19,500 hours of sounds. The company partnered with AudioSparx, a stock music licensing firm, to incorporate text metadata and secure permissions for copyrighted content.

The platform offers three pricing levels. The free tier allows up to 45 seconds of audio for 20 tracks per month, but prohibits commercial use. The Professional tier, priced at $11.99, permits 500 tracks up to 90 seconds long, while the Enterprise subscription offers customizable usage and pricing. Stable Audio is primarily used to create background music for podcasts or videos, simplifying content production. This new Stability AI tool is undoubtedly one of the most significant players to consider in the field of AI Music Generators, alongside other alternatives such as Google MusicLM, OpenAI's Jukebox, or Boomy.