These AI Headphones Isolate a Single Voice in a Noisy Space
Researchers at the University of Washington have developed an advanced AI system — “Target Speech Hearing” — for noise-canceling headphones that allows users to focus on a single speaker in noisy environments.
“We tend to think of AI now as web-based chatbots that answer questions,” said senior author Shyam Gollakota , a UW professor in the Paul G. Allen School of Computer Science & Engineering. “But in this project, we develop AI to modify the auditory perception of anyone wearing headphones, given their preferences. With our devices you can now hear a single speaker clearly even if you are in a noisy environment with lots of other people talking.”
Transcript
00:00:02 Noise cancellation headphones are a really interesting piece of technology that has changed the landscape of headphones hearing aids and everything. So basically they have the ability to clear out everything in the environment and artificially create an environment of silence around you. The next natural step, we thought, could be not only clear out everything, but also introduce whatever you want to hear. So this is a headphone system. It's called Look Once To Hear. We have these off the shelf noise canceling headphones
00:00:43 and we've attached two binary microphones on the right and left. We use the system to let people pick and choose the specific speakers that they want to listen to in a noisy environment. While your friend is speaking, you can just look at them for a couple seconds and by pressing a button, that would let the headphones know which voice to focus on based on the direction. When you look at a particular person their mouth is located equal distances towards your ears
00:01:44 and so a sound would arrive at the same time across both ears, which is a cue for the headphone system. And the headphone system uses AI technology to extract the voice that a wearer wants to hear while ignoring all sounds in that environment from that point on. Once I have this example of how they sound like I can just extract their sound even when I’m not looking at them. AI and especially neural networks have made great strides in speech processing. That application is really exciting -
00:02:40 especially useful for people with hearing challenges where they want to amplify the voice of the person they want to hear. Ideally, this is something that can be deployed on earbuds or hearing aids so that lots of people can use them. This can be big. This can impact lots of people. And being able to be at the forefront of this type of research is very exciting.