Researchers led by the University of Washington have developed AI-powered headphones capable of creating a personal sound bubble, minimizing background noise and enhancing nearby conversations. This innovation has the potential to revolutionize how we interact with our auditory environment.
Imagine sitting in a bustling office or a noisy restaurant and being able to focus solely on the conversation happening directly in front of you. Thanks to groundbreaking research led by the University of Washington, that scenario could become a reality. The researchers have developed a prototype for AI-powered headphones that create what they call a “sound bubble.”
“Our abilities to focus on the people in our vicinity can be limited in places like loud restaurants, so creating sound bubbles on a hearable has not been possible so far. Our AI system can actually learn the distance for each sound source in a room and process this in real-time, within 8 milliseconds, on the hearing device itself,” senior author Shyam Gollakota, a professor in the Paul G. Allen School of Computer Science & Engineering at UW, said in a news release.
The technology utilizes artificial intelligence algorithms in conjunction with six small microphones attached to the headband of commercially available noise-canceling headphones. The proprietary neural network tracks when different sounds reach each microphone, then suppresses sounds outside the bubble while amplifying those within it. This configuration allows users to clearly hear conversations or sounds within a 3-6 feet radius while minimizing background noise by an average of 49 decibels.
Published in the journal Nature Electronics, this innovative research stands out for its practical applications. Unlike current noise-canceling devices, which primarily rely on tracking head position and amplifying directional sound, this new system functions based on distance, allowing it to simultaneously amplify multiple nearby speakers while diminishing distant noise.
“We’d worked on a previous smart-speaker system where we spread the microphones across a table because we thought we needed significant distances between microphones to extract distance information about sounds,” added Gollakota. “But then we started questioning our assumption. Do we need a big separation to create this ‘sound bubble’? What we showed here is that we don’t.”
To develop an effective AI model, the researchers collected a distance-based sound dataset using an ingenious setup: they mounted the prototype headphones on a mannequin head attached to a robotic platform that rotated and emulated various environments with moving speakers. This allowed the team to gather comprehensive data in diverse indoor settings.
The implications of this research extend beyond just creating a more pleasant auditory experience. It has potential applications in assistive hearing devices, transforming how people with hearing impairments interact with their environments.
While the current system is optimized for indoor use, the team is working toward making the technology compatible with earbuds and hearing aids, which presents additional challenges in microphone placement.
“Humans aren’t great at perceiving distances through sound, particularly when there are multiple sound sources around them,” Gollakota added.
The breakthrough lies in training AI to discern sound sources’ distances by comparing the phases of multiple frequencies, effectively allowing the headphones to create a personalized, quiet auditory space.