Skip to content 
A revolutionary artificial intelligence system developed by researchers at the University of Washington enhances the way users can focus on a speaker’s voice amidst noise. With noise-canceling headphones equipped with this technology, a user can “enroll” a speaker by simply looking at them for a few seconds. This action prompts the system to isolate and amplify the enrolled speaker’s voice, maintaining clarity even as the listener moves or turns away.
This innovation builds on the team’s previous work in creating a “semantic hearing” system for noise-canceling headphones, allowing users to filter out unwanted noises while preserving sounds of interest. The new “Target Speech Hearing” system extends this capability with a user-friendly process involving off-the-shelf headphones augmented with binaural microphones. By tapping a button and facing the speaker, the system captures the speaker’s voice patterns and continuously adjusts to focus on this specific voice, improving its accuracy with prolonged speech.
The technology showcases a level of personal audio control beyond what is currently offered by leading noise-canceling headphones, proposing a solution for clear auditory communication in noise-polluted environments. This system, although still in prototype, represents a significant leap forward in personal audio technology, offering a glimpse into a future where listening can be tailored in real-time to the listener’s needs.
The University of Washington team shared their breakthrough at the ACM CHI Conference on Human Factors in Computing Systems and has made the proof-of-concept device’s code accessible for further innovation. Although not yet available for purchase, the system promises a new realm of possibilities for enhancing speech clarity in any noisy scenario.
