Posted Reaction by PublMe bot in PublMe

Want to hear a saxophone meow? Just ask Nvidia’s new AI music generator, FugattoHearing a trumpet bark or getting a saxophone to meow might sound like the product of one of your fever dreams, but it won’t remain so for long with Nvidia’s new AI music generator.
Called Fugatto (short for Foundational Generative Audio Transformer Opus 1), the tool is said to be capable of generating “any combination of music, voices and sounds” using texts and audio inputs. It can even let you produce sounds “never heard before”, says the chip giant.

READ MORE: “You get to a point where you’ve gotten good enough so that nothing is good enough”: Billy Joel describes the “curse” of being a songwriter

Researchers at Nvidia described Fugatto as a “Swiss army knife for sound”, and found that it could handle tasks it was not pretrained on, like generating a high-quality singing voice from a text prompt.
Unlike most AI music generators on the market which can only recreate the training data they’ve been exposed to, Fugatto allows users to “create soundscapes it’s never seen before”, such as a “thunderstorm easing into a dawn with the sound of birds singing” — or a saxophone that howls and barks for that matter.
To put it simply, “whatever users can describe, the model can create.”
Musicians can use Fugatto to quickly compose or edit ideas for a song, try out different styles, instruments and voices (even accents and emotions!), as well as isolate vocal and instrumental stems. They can also add effects and enhance the overall audio quality of an existing track.
Want to replace a MIDI melody for some power female opera vocals? Or desecrate Beethoven’s Moonlight Sonata with some sick drum beats? Fugatto can do it for you.
“We wanted to create a model that understands and generates sound like humans do,” said Rafael Valle, a manager of applied audio research at NVIDIA and one of dozens of people behind Fugatto. “Fugatto is our first step toward a future where unsupervised multitask learning in audio synthesis and transformation emerges from data and model scale.”
To build Fugatto, researchers had to put together a dataset containing millions of audio samples used for training. The team then generated “data and instructions that considerably expanded the range of tasks the model could perform, while achieving more accurate performance and enabling new tasks without requiring additional data.”
“The history of music is also a history of technology. The electric guitar gave the world rock and roll. When the sampler showed up, hip-hop was born,” says Ido Zmishlany, a music producer and member of Nvidia’s Inception program. “With AI, we’re writing the next chapter of music. We have a new instrument, a new tool for making music — and that’s super exciting.”
Check out what Fugatto can do in the video below.

Learn more at Nvidia.
The post Want to hear a saxophone meow? Just ask Nvidia’s new AI music generator, Fugatto appeared first on MusicTech.

Hearing a trumpet bark or getting a saxophone to meow might sound like the product of one of your fever dreams, but it won’t remain so for long with Nvidia’s new AI music generator.