Nvidia unveils innovative AI for sound design

A groundbreaking AI model was introduced by Nvidia, showcasing advanced capabilities in audio and music generation. Known as Fugatto, the model can create novel sounds, modify voices, and even transform existing audio. Unlike other AI tools, it can take a piano melody and convert it into a human voice or adjust accents and emotional tones in spoken recordings.

Fugatto builds on generative AI’s potential to reshape creative industries like music, film, and gaming. Nvidia’s vice president of applied deep learning, Bryan Catanzaro, highlighted how computers have already revolutionised music through synthesizers, suggesting AI will usher in even greater innovation. While promising, the technology is not yet slated for public release due to concerns over ethical misuse and potential copyright issues.

The model was developed using open-source data and joins a growing trend of tools from companies like Meta and Runway, which also generate audio and video from text prompts. Nvidia’s innovation stands out for its focus on transforming existing recordings into entirely new formats, a feature that could significantly enhance creative possibilities.

Generative AI remains under scrutiny as industry leaders grapple with ethical concerns. The entertainment industry, already wary after disputes involving voice imitation, is debating how to integrate such technologies responsibly. Nvidia and others have acknowledged the risks of misuse, prompting a cautious approach to public rollouts.