Fugatto: NVIDIA unveils powerful generative video tool

By: Anry Sergeev | 25.11.2024, 19:05

NVIDIA has introduced Fugatto (Foundational Generative Audio Transformer Opus 1), a universal generative AI model capable of creating and transforming music, voices, and sounds based on text or audio requests. The company called the development a "Swiss Army knife for sound".


Video: NVIDIA Fugatto demonstration

Here's What We Know

Fugatto is unique in that it combines the capabilities previously performed by separate models. It is capable of performing complex tasks such as:

  • Generating music from text: For example, the model can create a "melancholic piano composition with a light jazz rhythm".
  • Sound transformation: You can change the instrument in a song, say, turn a guitar into a cello or add drums to an existing melody.
  • Voice modification: The model can adjust the emotional tone of your voice or even change its accents.
  • A special feature of Fugatto is the ability to synthesise sounds that have not yet existed. For example, it can create a unique noise that would be appropriate for a particular film or game.

Why It Matters

NVIDIA's development sets new standards for generative AI in audio. With Fugatto, creative professionals get a tool that not only automates routine processes, but also provides endless possibilities for creativity. Fugatto's ability to adapt to many scenarios means that in the future, the model could find application in areas where audio is crucial, from advertising to educational programmes.

This technology also confirms the trend of AI models moving towards multitasking. Previously, separate models were required for different tasks, but now Fugatto demonstrates that multifunctional solutions are not only possible but also effective.

What's behind the word Fugatto?

Obviously, it comes from the word "fugue". A fugue is a musical form and compositional technique characterised by a polyphonic structure, where one or more themes (melodies) are repeated and developed in different voices. The term is derived from the Latin word fuga, meaning "running" or "escape", which reflects the idea of a theme "running" from one voice to another. The fugue became a central genre in Baroque music, especially thanks to composers such as Johann Sebastian Bach, who is considered one of the greatest masters of this genre.

The word "fugatto" (Italian: fugato) means a polyphonic musical form that is related to the fugue. Fugatto originated in works of strict style and encompasses a technique of imitation where one or more themes are repeated and developed in different voices.

Source: blogs.nvidia.com