Long-range speech acquirement and enhancement with dual-point laser Doppler vibrometers

Abstract

This paper presents a long-range speech acquirement and speech enhancement system by utilizing two laser Doppler vibrometers (LDV) to measure two separate points on only one vibration object or two different vibration objects. This proposed LDV system can provide dual-channel synchronous signals, and thus the coherent-to-diffused ratio-based and the multi-channel linear prediction-based algorithms can be introduced to reduce the reverberation and the noise of the acquired speech signals. The two algorithms are combined together to further improve the speech enhancement performance. Experimental results show that the proposed algorithm can significantly improve the PESQ scores.