Audible Noise Suppression with a Real-Time Broad-Band Superdirective Microphone Array

A novel audible noise suppression (ANS) multichannel processor is proposed and implemented for real-time processing of data from a 15-microphone nested linear array. The ANS processor, based on the masking properties of the human auditory system, has been used successfully in single-channel systems. An enhanced multichannel version has now been developed, taking advantage of the extra information available in the acoustic spatial samples from the microphone array. This is used to improve the clean speech signal estimates used to calculate dynamically the noise hearing thresholds for ANS filtering, which will benefit the perceived and objective quality of the processed signal. While off-line experiments with a multichannel recorded database under different noise and reverberation conditions have previously assessed the performance of the system, several on-line experiments assessing the real-time prototype are described.