Speech Endpoint Detection Method Based on LMS Noise Reduction and Improved Dual-threshold

Authors

Chunli Zhu, 1. School of Mechatronics Engineering and Automation, Shanghai University, Shanghai 200072, China;;
Li Xin, 1. School of Mechatronics Engineering and Automation, Shanghai University, Shanghai 200072, China;;2. Department of Electronic and Information Engineering, Nanjing University, Nanjing 210093, China;;3. State Key Lab of Pattern Recognition, Institute of Automation, Chinese Academy of Sciences, Beijing 100080, China;

Abstract

Abstract: Endpoint detection of speech in low signal-to-noise ratio plays a very important role in voice processing. For a speech signal that is polluted by additive noise, it is possible to suppress the noise and keep the original speech relatively constant. A dual-threshold speech endpoint detection algorithm was proposed based on the least mean squares error (LMS) adaptive filtering. Dual median filtering smoothing was performed before and after double-parameter double-threshold detection. Through the Matlab simulation, the speech endpoint detection method was compared with other methods. In the noise environment with low signal-to-noise ratio, the endpoint detection effect of speech has better accuracy and robustness.

Recommended Citation

Zhu, Chunli and Xin, Li (2020) "Speech Endpoint Detection Method Based on LMS Noise Reduction and Improved Dual-threshold," Journal of System Simulation: Vol. 29: Iss. 9, Article 11.
DOI: 10.16182/j.issn1004731x.joss.201709011
Available at: https://dc-china-simulation.researchcommons.org/journal/vol29/iss9/11

First Page

1950

DOI Link

https://doi.org/10.16182/j.issn1004731x.joss.201709011

Last Page

1960

CLC

TN912.3

Recommended Citation

Zhu Chunli, Li Xin. Speech Endpoint Detection Method Based on LMS Noise Reduction and Improved Dual-threshold[J]. Journal of System Simulation, 2017, 29(9): 1950-1960.

DOI

10.16182/j.issn1004731x.joss.201709011