Stability AI has announced the launch of Stable Audio, a new text-to-music sound generator powered by generative AI. Stable Audio can generate high-quality clips of music and sound effects up to 90 seconds and respond to detailed instructions on genre, instrumentation, mood, BPM and more.
Stable Audio was trained on a dataset of over 800,000 audio files from the stock music library AudioSparx. Its creators have published a blog post that shows several impressive examples of what Stable Audio can do. Judging from these, it looks like it could be set to surpass the capabilities of MusicGen and MusicLM, AI music generators developed by Meta and Google that were released over the past year.
Stability AI say that Stable Audio is "ideal for musicians seeking to create samples to use in their music". The examples published by its creators show that the tool can generate clips of any given BPM, something we found its competitors unable to do: the prompt "116 BPM rock drums loop clean production" produced a clean, usable rock drum loop at exactly 116 BPM. It doesn't seem to understand tonality, though, as we asked it to give us a piano loop in G major and received something in A minor.
Stable Audio can create clips of varying lengths and is particularly effective at creating structured musical ideas, due to the fact that it was trained to account for "audio file duration and start time" in addition to musical attributes like genre and instrumentation. This means that it's less likely to generate random sections of a song that start or end in the middle of a phrase, something that other AI music generators often do.
Stable Audio is available in two tiers. The free tier, which is available now for anyone to use, will generate audio clips of up to 20 seconds, while the Pro tier is available to subscribers at $11.99 a month. Pro subscribers can generate up to 500 tracks a month of up to 90 seconds in duration and these can be used in commercial projects.
“As the only independent, open and multimodal generative AI company, we are thrilled to use our expertise to develop a product in support of music creators,” said Emad Mostaque, CEO of Stability AI. “Our hope is that Stable Audio will empower music enthusiasts and creative professionals to generate new content with the help of AI, and we look forward to the endless innovations it will inspire.”
Stability AI is the company behind Stable Diffusion, the popular image-generation tool released in 2022. Stable Diffusion was previously used to generate music by the creators of text-to-music AI tool Riffusion.