EUROSPEECH 2003 - INTERSPEECH 2003
8th European Conference on Speech Communication and Technology

Geneva, Switzerland
September 1-4, 2003

        

Time Delay Estimation Based on Hearing Characteristic

Zhaoli Yan, Limin Du, Jianqiang Wei, Hui Zeng

Chinese Academy of Sciences, China

This paper proposes a new time delay estimation model, Summary Cross-correlation Function (SCCF). It is based on a hearing model of the human ear, which comes from a pitch perception model. The inherent relation between some time delay estimation (TDE) and pitch perception method is mentioned, and propose an idea - some pitch perception models' pre-processing can be used for references in TDE model and vice versa. The new TDE model is proposed based on this viewpoint. Then SCCF is analyzed further, and compares its performance with Phase Transform (PHTA) and Modified Cross-power Spectrum (M-CPSP). The simulated experiments show that the new model is more robust to noise than PHAT and M-CPSP.

Full Paper

Bibliographic reference.  Yan, Zhaoli / Du, Limin / Wei, Jianqiang / Zeng, Hui (2003): "Time delay estimation based on hearing characteristic", In EUROSPEECH-2003, 557-560.