Stability AI Unveils Stable Audio 2.5 for Professional Sound Production
Stability AI has unveiled Stable Audio 2.5, a groundbreaking audio generation model tailored for professional sound production. This new tool promises high-quality, customizable audio content at scale, responding better to mood descriptions and musical jargon.
The model's speed and efficiency are achieved through a unique training method called Adversarial Relativistic-Contrastive (ARC). It can generate music pieces up to three minutes long in just a few seconds, with a processing time of less than two seconds on industrial Nvidia H100 GPUs.
Stable Audio 2.5 can generate audio from text and even complete tracks based on existing audio files through its Audio Inpainting feature. It creates improved musical structures, including multi-part compositions with intro, development, and outro.
In May, Stability AI introduced Stable Audio Open Small for smartphones, using the same ARC training method. The company has partnered with Amp, a subsidiary of advertising agency WPP, to create specialized solutions for large clients.
Stable Audio 2.5 opens up new possibilities in various fields, such as advertising spots, game intros, mall music, and characteristic tones for credit card transactions. With its ability to generate high-quality, customizable audio content at scale, it is set to revolutionize the professional sound production industry.
