AI Headphones Can Hear Single Person in a Crowd
Researchers at the University of Washington have developed an AI system, “Target Speech Hearing” (TSH), that allows headphone users to focus on a single speaker in noisy environments by looking at them for three to five seconds. Unlike current noise-canceling headphones, which offer limited control over environmental sounds, TSH isolates and amplifies the desired speaker’s voice even as the listener and speaker move. This system, presented at the ACM CHI Conference, utilizes machine learning to identify and continuously enhance the clarity of the enrolled speaker’s voice.
The proof-of-concept, tested on 21 subjects, showed significant improvement in voice clarity. While it can only enroll one speaker at a time and struggles with overlapping voices, future developments aim to expand its application to earbuds and hearing aids. The research, supported by several funding sources, is available for further development but is not yet commercially available.