Stability AI, a leading company in AI technology, is making waves in the audio industry with the release of Stable Audio Open 1.0. While Stability AI is known for its stable diffusion text-to-image generation AI technology, it has a broader portfolio that includes models for code, text, and now, audio. Since its public launch in September 2023, Stable Audio has gained popularity as a text-to-audio generative AI tool. The recent release of Stable Audio 2.0 on April 3 has brought even more clarity and length to the generated audio.
Stable Audio Open: A Limited but Accessible Tool
Stable Audio Open is a more restricted version of the full Stable Audio tool. While the full tool can generate audio of up to 3 minutes, Stable Audio Open focuses on shorter pieces such as sound effects. This limited version aims to provide audio researchers and producers with hands-on access to one of Stability AI’s generative audio models, accelerating research, adoption, and practical creative use of these tools.
Open Access with Responsible Training
As the name implies, Stable Audio Open is an open model. However, it is not technically open source. Instead, it is available to users under the Stability AI non-commercial research community agreement license. This license grants open access to the model but imposes limitations on what users can do with it. Stability AI has taken a responsible approach to training the model by using audio data from FreeSound and the Free Music Archive. This ensures that no copyrighted or proprietary material was used without permission, maintaining ethical standards in their AI development.
Fine-tuning for Unleashing Creativity
One of the key benefits of Stable Audio Open is the ability for users to fine-tune the model on their own custom audio data. This means that musicians and producers can train the model on their own drum recordings, for example, to generate unique beats. The fine-tuning process is made possible through the Stable Audio Tools library, which is licensed under an actual open-source license. This empowers users to have control and creativity over their generated audio.
Looking Towards the Future
Zach Evans, head of audio research at Stability AI, expressed the company’s dedication to continuously improving the quality and controllability of their generative audio models. He mentioned that they are excited about future releases, both commercial and open model, that will reflect the progress made by their research. With Stable Audio Open 1.0, Stability AI is not only providing a valuable tool for audio researchers and producers but also paving the way for further advancements in generative AI technology.
Transform 2024: The Place to Be
If you’re interested in diving deeper into the world of AI and generative technology, don’t miss out on Transform 2024. This exclusive AI event will take place in San Francisco from July 9 to 11. It offers a unique opportunity to connect with industry leaders, explore the opportunities and challenges of generative AI, and learn how to integrate AI applications into your industry. Register now to be a part of this transformative event!
