top of page

Introducing Stable Audio, an AI Product for Music and Sound Generation


Stability AI, a company that specializes in open generative AI, has announced the launch of Stable Audio, a product that can generate music and sound effects from text descriptions. Stable Audio is the first product of its kind that uses latent diffusion, a state-of-the-art generative AI technique, to deliver faster and higher-quality audio via an easy-to-use web interface.


Stable Audio allows users to create audio tracks of up to 90 seconds by entering descriptive text prompts, such as “epic trailer music intense tribal percussion and brass” or “car passing by”. The product can generate a wide range of music genres and sound effects, from ambient techno to drum solos. Users can also specify the desired length and start time of the audio, giving them more control over the output.


Stable Audio is powered by a large-scale neural network model that was trained on over 800,000 audio files and corresponding text metadata from AudioSparx, a leading music library. The model uses a compressed audio representation to reduce inference time and a timing conditioning mechanism to adjust the audio length. You can read more about the research behind the model here.


Stable Audio is ideal for musicians, producers, podcasters, filmmakers, game developers and anyone who needs high-quality audio content for their projects. Stability AI offers a basic free version of Stable Audio, which can generate and download tracks of up to 45 seconds, and a ‘Pro’ subscription, which delivers 90-second tracks that are downloadable for commercial use.


Stable Audio is the latest product from Stability AI, which also developed Stable Diffusion, an image synthesis model that can generate realistic images from text or sketches. Stability AI is backed by Carlyle Group, a private equity firm that invested $500 million in the company in 20221.


To try Stable Audio for yourself, you can visit www.stableaudio.com. For more information or feedback, you can contact Stability AI at research@stability.ai.

Comments


bottom of page