Active noise cancellation (ANC) technology in pioneering headphones and headphones has changed the way we experience sound, offering a switch to silence disturbing background noise. However, this often forces users to choose between muting all sound or cutting out the noise for conversation.
Researchers at the University of Washington have developed a solution for their AI-powered “target speech listening” system. This new technology allows the headphones to isolate and amplify specific sounds while cutting out ambient noise. Users only see the speakers, and the AI picks up unique speech patterns, allowing them to focus on the voice even if the user is moving.
The system uses commercial headphones modified with microphones and AI neurons. By pressing a button and looking at the speaker, the headphones “register” the voice for just three to five seconds. The AI then continues to track and separate the voice, even if the user is looking away.
Testing showed a significant improvement in clarity, with users overestimating voice quality by nearly twice as much compared to traditional systems. While the system currently has to produce the loudest volume of the target speaker, this could become more versatile with future developments.
This technology has the potential to help partially hear and improve communication in noisy environments. The researchers aim to integrate the system with branded headphones and hearing aids to provide valuable accessibility features.
Advances in AI and neural networks have opened up new possibilities for speech processing, making conversation clear and accessible in chaotic situations
