NVIDIA has unveiled Fugatto, a cutting-edge AI model designed to transform how we create and interact with audio. This powerful tool blends text and sound, allowing users to generate or modify music, voices, and sounds simply by using text prompts.
What Makes Fugatto Special?
Unlike other AI tools, Fugatto stands out for its versatility and precision. It can:
- Compose music snippets.
- Change the mood or accent of a voice.
- Add or remove instruments from songs.
- Create entirely new sounds never heard before.
“This tool is incredible,” said Ido Zmishlany, a multi-platinum producer. “With Fugatto, I can invent new sounds instantly in the studio.”
Powerful and Creative Features
Fugatto can perform unique tasks, such as:
- Making a saxophone sound like it’s meowing.
- Turning a trumpet into a barking sound.
- Gradually changing one sound into another, like thunder turning into birdsong.
These capabilities are made possible by ComposableART, a feature that combines different instructions into one smooth output. For example, users can create a voice with a French accent and adjust the emotional tone as needed.
Applications Across Industries
Fugatto can be used in many industries, including:
- Music Production: Quickly test new song ideas and improve audio quality.
- Advertising: Adapt voices and sounds for different regions and audiences.
- Education: Personalize learning with familiar voices for language lessons.
- Gaming: Create new sound effects or match audio with in-game actions.
Zmishlany believes this tool could change the music industry. “The electric guitar gave us rock. The sampler created hip-hop. AI is the next big instrument,” he said.
The Technology Behind Fugatto
Fugatto is powered by NVIDIA’s expertise in speech and audio processing. It uses NVIDIA DGX systems and 32 NVIDIA H100 Tensor Core GPUs, with 2.5 billion parameters for maximum performance.
An international team of researchers from countries like India, Brazil, and South Korea developed the model, ensuring it supports multiple languages and accents.
A Game-Changer for Audio
The team behind Fugatto experienced many exciting moments during its development. One highlight was when the AI successfully created music from a simple text prompt. Another memorable demo featured electronic music mixed with barking dogs, making the entire team laugh.
With Fugatto, NVIDIA is paving the way for a new era of creativity in music and sound design.
Source: itnewsafrica
