A real-time architecture for dual-microphone speech enhancement systems

In this paper, we propose a real-time architecture for a dual-microphone speech enhancement to be employed in a hands-free mobile communication. The proposed method is based on a new SNR estimation and a coherence based Voice Activity Detector (VAD). The enhancement system was implemented using MATLAB tools and DSP processors. The results show that these two implementations are comparable in terms of PESQ score. We also calculate the computational complexity of the enhancement method. The obtained results show that the proposed method is suitable from the hardware implementation point of view.