TAIPEI (TVBS News) — Soundskrit Chief Technology Officer Hsu Wan-thai (徐萬泰) showcased the company's directional MEMS microphone technology at the SEMICON 2026 pre-event forum on Monday (Aug. 31). The technology is designed to isolate specific sounds in noisy environments, with potential applications in AI glasses, self-service machines and smart accommodations.
Hsu highlighted a key challenge facing AI smart glasses: capturing relevant sounds while filtering out background noise. Soundskrit's directional MEMS microphones address this by detecting the location of a speaker and focusing on the intended sound source, reducing noise interference and the need for high-performance edge computing chips.
Soundskrit plans to incorporate the technology into AI glasses to improve speaker recognition while reducing reliance on intensive computing resources. By isolating the intended audio source, the microphones could improve voice-based interaction in a range of environments.
The technology could also benefit smart accommodations and self-service machines. In settings such as hotel lobbies, directional microphones can focus on users operating the machines, improving voice recognition and human-machine interaction even when there is significant background noise.
Hsu said the microphones can improve audio quality through precise spatial hearing, enabling 3D sound-source localization, spatial audio capture and noise suppression. By processing audio more efficiently at the microphone level, the technology can reduce the computing resources required for applications while maintaining accurate sound detection.