(Web Desk) - Researchers have developed an AI audio generator that they claim can create sounds that have never been heard before.
The new generative artificial intelligence model, called Fugatto, was built by a team at computer chip giant Nvidia who wanted to create a “Swiss Army knife for sound”.
The tool allows users to edit or generate audio using simple text prompts, such as removing a particular instrument from a song or changing the accent of someone’s voice.
In a blog post detailing Fugatto, Nvidia’s Richard Kerris claimed it offered more dexterity than any other AI model and holds the potential to revolutionise music through brand new sounds.
“Fugatto can make a trumpet bark or a saxophone meow,” Mr Kerris wrote. “Whatever users can describe, the model can create.”
A video demonstrating the tool shows Fugatto creating a unique audio sample from a text prompt: “Deep, rumbling bass pulses paired with intermittent, high-pitched digital chirps, like the sound of a massive, sentient machine waking up.”
Another example shows Fugatto transforming the sound of a train into a string orchestra.
“This thing is wild,” said Ido Zmishlany, a producer and songwriter who is part of Nvidia’s Inception program to help new startups.
“Sound is my inspiration. It’s what moves me to create music. The idea that I can create entirely new sounds on the fly in the studio is incredible.”
Fugatto took more than a year to develop, requiring millions of audio samples to train it.
The rapid advancements that generative AI has brought to audio technology has led to concerns about how it might impact people working within creative industries.