Nvidia is taking a giant step forward in music creation with its incredible new tool, Fugatto. This innovation leverages artificial intelligence to generate music, sounds and speech from audio and text inputs. The originality lies in its ability to create sounds never heard beforesuch as a meowing trumpet.

Limitless creativity

Fugatto offers users the ability to compose music tracks based on surprising prompts. For example, by asking him to “create a howling, barking saxophone with dogs,” the tool ensures unique sound production. Here are some notable features:

• Creation of original sounds: Fugatto can generate a wide range of sounds in response to textual descriptions.
• Voice transformation: The tool allows you to change the accent and tone of a voice, whether angry or calm.
• Advanced music editing: Fugatto can isolate vocals in a song, add instruments, or change melodies at will.

A sample of possibilities

To illustrate its potential, Nvidia shared several examples of sound effects that Fugatto can produce. For example, he could compose deep bass pulses accompanied by digital tweetsevoking the sound of a giant machine waking up. These features open up an unprecedented range of creative possibilities for musicians, producers and content creators.

Data for innovation

To design Fugatto, Nvidia researchers gathered a massive dataset containing millions of audio samples. They then developed instructions that greatly expanded the model’s capabilities. Using this approach, AI can perform a variety of tasks with greater precision, while requiring less additional data.

A competitive context

Importantly, Fugatto is entering a market already rich in AI audio tools. Companies like Stability AI, OpenAI, Google DeepMind, ElevenLabs And Adobe also offer music solutions based on artificial intelligence. However, none of these platforms claim to match Fugatto’s ability to create completely new and unheard sounds.

Additionally, some startups are experiencing legal copyright issues for their music creation tools. A recent report also found that companies like Nvidia trained their AI models using subtitles from thousands of YouTube videos, raising ethical questions about the use of data.

What future for Fugatto?

Currently, Nvidia has not specified when, or even if, Fugatto will be widely available on the market. The possibilities offered by this technology push the boundaries of musical creativity and pose exciting questions about the future of music composition and publishing. While we await more information on possible commercialization, there is no denying that Fugatto can transform the way we perceive and create music.