Fig. 8From: Parallel processing of distributed beamforming and multichannel linear prediction for speech denoising and deverberation in wireless acoustic sensor networksPESQ improvement of the evaluated methods with varying degrees of VAD errors. a Accurate RETFs, b inaccurate RETFsBack to article page