AI noise-cancelling headphones let you focus on just one voice

Prototype noise-cancelling headphones allow you to select which background noises to drown out, letting you put an “audio spotlight” on one specific voice so you can concentrate on it.

Conventional noise-cancelling headphones reduce unwanted sounds like the rumble of a bus engine, but because the technology cancels out certain frequencies entirely, it can also suppress sounds we want to hear.

Now, Shyam Gollakota at the University of Washington in Seattle and his colleagues have created headphones that can remove any unwanted noises while leaving others intact, regardless of their frequencies. It can also be trained with the press of a button to home in on a specific person’s voice and exclude all other noise.

The researchers are presenting their prototype at a joint meeting of the Acoustical Society of America and the Canadian Acoustical Association this week. The device uses an artificial intelligence system called a neural network that has been trained on many examples of 20 different types of sound, including alarm clocks, crying babies and birdsong. The user can choose to turn on or off each category of sound from an app, allowing it to pass through the headphones or be blocked.

The prototype consists of commercially available headphones with a microphone attached on the outside of the housing that covers each ear. These microphones record ambient sound and pass it to either a small Orange Pi microcontroller or a smartphone on which the neural network runs. This AI then removes unnecessary sounds and transmits the edited audio feed to the headphones. Gollakota says the equipment could be built into a set of headphones.

The technology works in the same way as the AI that was used to isolate individual instruments and voices amid a noisy jumble recorded during work on The Beatles’s 1970 album Let It Be, allowing director Peter Jackson to create the documentary series The Beatles: Get Back.

That process took some time, but this prototype can process audio within just 8 milliseconds – because the team kept the neural network small and simple enough for a mobile device to run quickly – to avoid confusing delays between things happening and you hearing them.

Gollakota says that the effect is like an “audio spotlight” being turned onto a noise source, allowing you to focus intently on it even in chaotic and loud environments.

“This has new capabilities which give more control to the user. We’re taking the first steps of human acoustic perception augmentation right now,” says Gollakota.