A Robust Binaural Linearly Constrained Minimum Variance with Spatial Cues Preservation for Hearing Aids Beamforming
share
In this work, a robust binaural beamforming algorithm for hearing aid applications is introduced. The robust binaural beamforming algorithm has two main components: a Robust Binaural Linearly Constrained Minimum Variance (BLCMV) based on imposing two constraints around the estimated direction of the target signal, and a post processor based on a complex coherence for classification, selection, and mixing of binaural signals at each time-frequency bin. The robust BLCMV provides a good level of noise reduction and low level of target distortion under realistic conditions.The post processor enhances the beamformer abilities to preserve the binaural cues for the diffuse-like background noise and the directional interferers, while keeping a good level of noise reduction. The algorithm introduced does not require knowledge of the interfering sources' directions nor the second order statistics of the noise-only components. The introduced algorithm requires an estimate of the target speaker direction, but it is designed to be robust to some deviation from the estimated direction. Comparing with a recently proposed state of the art method, comprehensive evaluations have been performed under complex acoustic scenarios generated in both anechoic and reverberant environments, considering a mismatch between the estimated and the true direction of arrivals for the target speakers as well as for the interferers. Mismatch between the anechoic propagation models used for the design and the reverberant propagation models used to generate the directional signals is also considered. The results illustrate the robustness of the proposed algorithm to mismatches generated from the direction of arrivals and from the reverberation.
READ FULL TEXT
