Stability AI Launches Revolutionary “Stable Audio” for Creative Music Generation

In a bold leap forward in audio generation technology, Stability AI has introduced “Stable Audio,” a groundbreaking artificial intelligence model that allows users to craft custom audio clips from simple text prompts. This innovation builds upon the success of Stability AI’s text-to-image generation technology, Stable Diffusion, and marks a significant advancement in the field of audio composition.

Direct Interaction with Raw Audio

Traditional methods of generating audio tracks from text prompts often involve the use of MIDI files and symbolic generation, resulting in repetitive and constrained musical compositions. Stable Audio takes a novel approach by directly engaging with raw audio samples, liberating creators from these limitations and paving the way for entirely new and innovative musical expressions.

A Symphony of Data

Stable Audio’s prowess is underpinned by a comprehensive training dataset comprising over 800,000 licensed music pieces from the AudioSparks library. This vast repository not only ensures superior audio quality but also provides essential metadata, enriching the capabilities of text-based models and enhancing the overall user experience.

Freedom of Expression

Unlike image generation models that mimic specific artistic styles, Stable Audio doesn’t seek to emulate iconic bands or artists. Instead, it empowers users to embark on their creative journeys without the constraints of rigid stylistic boundaries. It celebrates artistic freedom and encourages individuals to explore their unique musical ideas.

1.2 Billion Parameters: Powering Creativity

The Stable Audio model boasts an impressive 1.2 billion parameters, putting it on par with the original Stable Diffusion model known for its image generation prowess. Text prompts, an integral component of audio generation, were meticulously developed and trained using the Contrastive Language Audio Pretraining (CLAP) technique. To assist users in crafting effective prompts, Stability AI is releasing a prompt guide in tandem with the Stable Audio launch.

Accessible Pricing Tiers

Recognizing the importance of accessibility, Stability AI offers two versions of Stable Audio. The free version allows users up to 20 monthly generations, with each generation producing tracks of up to 20 seconds. For those seeking more extensive creative opportunities, the Pro version extends these limits, enabling 500 generations and extending track duration to a generous 90 seconds. 

Conclusion: 

Stability AI’s introduction of Stable Audio represents a monumental shift in audio generation technology. By harnessing advanced AI techniques and offering accessible pricing options, this innovative tool opens new horizons for creative expression in the realm of music and audio production. Musicians, composers, and audio creators, both aspiring and professional, now have a potent ally to bring their unique musical visions to life. Stable Audio promises to democratize the world of music composition, empowering creators to explore, innovate, and redefine the boundaries of musical artistry.