Financial Markets


Stability AI, the groundbreaking tech company known for its game-changing AI technologies, has once again set the stage for a seismic shift in the digital creative landscape. They have just announced the public release of their latest innovative software, Stable Audio, a technology that morphs simple textual prompts into finely produced audio clips. This cutting-edge technology represents the future of music and audio generation, exploiting the same advanced AI methods perfected by Stability AI's text-to-image generation AI technology, Stable Diffusion.

The birth of Stable Audio is rooted in Stability AI's internal research studio, Harmonai, which focuses on music generation. The uniqueness of Stable Audio lies in its deviation from the conventional 'symbolic generation' methods; instead of relying on MIDI files, it has embraced the raw form of art and works directly with raw audio samples. This revolutionary approach allows for the creation of unprecedented music styles that cannot be achieved through MIDI and symbolic generation.

The AI model powering Stable Audio is nurtured on over 800,000 licensed music pieces from the renowned audio library, AudioSparks. This extensive training fuels its creativity and contributes to its sophisticated audio generation capabilities. Accompanied by the diverse metadata provided by AudioSparks, the model draws upon a wide spectrum of audio data, therefore elevating the quality and diversity of its output.

In an era where AI models are often trained to mimic established artists, Stable Audio veers off the beaten track by not allowing the exact mimicking of any particular artist's style. Instead of spoon-feeding the creativity of users, it champions user innovation and originality, encouraging users to create unique musical pieces. To facilitate the generation of audio clips, Stability AI uses the inventive Contrastive Language Audio Pretraining (CLAP) technology to handle the text prompts required for audio generation.

The generous release of Stable Audio even comes as a free package which allows up to 20 generations per month of up to 20-second tracks. For users craving more, a Pro version is available at a manageable cost of $12 per month. This upgrade permits larger usage, supporting up to 500 generations and increasing the maximum duration of individual tracks to 90 seconds.

Stable Audio is not just another AI technology; it exhibits the future of music and audio generation. This tantalizing intersection of AI and music holds promises of a new kind of digital revolution, one that enhances user creativity. While its full ramifications are yet to be felt, one thing is clear: the world of music generation will never be the same again - a testament to the transformative power of AI technologies and the bright, limitless future they portend.