Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This study enhances mask-based beamforming by introducing a multi-channel formulation that applies distinct masks to each microphone, addressing the limitations of single-mask approaches in spatially distributed microphone setups. The method is particularly beneficial in scenarios with significant signal variation across microphones, as it captures spatial diversity more effectively. Experimental results indicate that this multi-channel masking approach outperforms traditional methods in distributed microphone configurations while maintaining comparable performance in compact arrays.
Multi-channel masking outperforms single-mask strategies in diverse acoustic environments, revealing significant gains in speech enhancement for spatially distributed microphones.
Mask-based beamforming is a popular geometry-agnostic approach for speech enhancement, typically applying a single mask across all microphones to estimate the required covariance matrices. While effective for compact arrays, this strategy may be suboptimal for spatially distributed microphones, where signal characteristics may vary strongly across microphones. To effectively capture the spatial diversity across microphones, we extend the mask-based beamformer to a multi-channel formulation, where each microphone is pre-filtered by a separate mask before covariance estimation. To address time-varying acoustic scenes, caused by spectro-temporal nonstationarity, we adopt a frame-causal online implementation with a sliding window. Experiments with simulated compact arrays and distributed microphones show that multi-channel masking yields a benefit over using a single mask when microphone signals differ substantially, while retaining similar performance in compact arrays. We further demonstrate the robustness of the multi-channel masking approach by comparing oracle ideal ratio masks to blind DNN-based mask estimation.