In modern hearing assistant devices, it is common practice to implement adaptive beamforming. Microphone array adaptive beamforming allows the wearer of the hearing aids to listen target sound source more clearly by spatially filtering the acoustic environment. The minimum variance distortionless response (MVDR) beamformer is a widely utilized optimal beamforming design. The equation for MVDR is:

min_w w^H S_{xx} w subject to w^H d=1

The goal is to minimize the interference and noise sources, while preserving the desired signal in a specific direction.
Binaural cues are important features for humans to determine the lateral location of a sound source. For hearing aids, it is important to preserve these auditory cues of the desired speech signal. Implementing a microphone array with MVDR for each ear independently runs the risk of destroying this binaural information with each array could be solving their own local optimization problem. A solution in which the left and right beamformers are working cooperatively towards the same target source signal is referred to as binaural beamforming. A binaural MVDR beamformer uses a wireless communication channel to share the single channel audio result from each hearing aid and iteratively update the beamforming weights.

binaural mvdr beamforming

The potential drawback of the binaural MVDR is that auditory cues of the interfering sources are attenuated as part of algorithm. In “Binaural LCMV Beamforming” it will be discussed how this information can be preserved.
VOCAL Technologies offers custom designed solutions for beamforming with a robust voice activity detector, acoustic echo cancellation and noise suppression. Our custom implementations of such systems are meant to deliver optimum performance for your specific beamforming task.


More Information