Speech enhancement using a critical point based Wiener Filter

This paper presents an new approach to noise reduction for voice communication over Bluetooth technology. In the literature, several authors have compared the performance of different filtering techniques, such as the well-known Spectral Subtraction (SS), and Wiener Filter (WF) using simulated data, whereas this research uses real-time data samples collected from cars subjected to a noisy environment with varying sound levels in search of an optimal solution. The cars were driven at different speeds with the windows and fan set to different configurations. The tests were analyzed both subjectively and objectively. The proposed algorithm is based on two superior voice detection and noise estimation techniques, namely Log-Energy Voice Activity Detection (VAD) and First-in First-out (FIFO), and it has shown a considerable improvement in audio quality over the VAD and FIFO only algorithms.