סמינר: Graduate Seminar
Speech Reinforcement with Acoustic Feedback Suppression and Region-of-Interest Beamforming for Speech Communication and Wearable Devices
Date:
January,14,2026
Start Time:
12:30 - 13:30
Location:
1061, Meyer Building
Add to:
Lecturer:
Yehav Alkaher
Research Areas:
| Speech perception in complex acoustic environments remains a significant challenge for both normal-hearing and hearing-impaired individuals. Speech Reinforcement (SR) applications, ranging from in-room and remote conference calls to in-car communication and hearing aids, aim to bridge this gap by amplifying a desired speaker. However, these systems are fundamentally limited by acoustic feedback and the resulting “howling” phenomenon caused by the coupling between microphones and loudspeakers.In this seminar, I will address the challenges of high-quality speech acquisition and the difficulty of maintaining high amplification gain without compromising output quality in feedback-prone environments. I will first examine the mechanisms of acoustic feedback in several scenarios where SR systems are required, alongside the limitations of traditional suppression methods. I will then introduce our novel approach to spatial filtering: Region-of-Interest (ROI) beamforming in the time-domain. Unlike conventional beamformers that require precise source localization, ROI beamforming accounts for the spatial uncertainties of sound sources that are inherent in wearable devices, such as hearing glasses. Finally, I will present our adaptive time-domain ROI beamforming solution, demonstrating its efficacy in suppressing feedback while maintaining robust speech acquisition in dynamic, real-world scenarios. |
| Ph.D. student under the supervision of Prof. Israel Cohen.
|