Speech Enhancement with Dual-path Multi-Channel Linear Prediction Filter and Multi-norm Beamforming

In this paper, we propose a speech enhancement method us ing dual-path Multi-Channel Linear Prediction (MCLP) filtersand multi-norm beamforming. Specifically, the MCLP part inthe proposed method is designed with dual-path filters in bothtime and frequency dimensions. For the beamforming part, weminimize the power of the microphone array output as well asthe l1 norm of the denoised signals while preserving source sig nals from the target directions. An efficient method to select theprediction orders in the dual-path filters is also proposed, whichis robust for signals with different reverberation time (T60) val ues and can be applied to other MCLP-based methods. Eval uations demonstrate that our proposed method outperforms thebaseline methods for speech enhancement, particularly in highreverberation scenarios.