Research on Pitch Extraction Algorithm of Noisy Speech

Abstract:

Pitch detection in noisy environment plays an important role in speech analyzing and recognition. In the light of the properties of Hilbert-Huang transform and the EMD soft-threshold de-noising method, an effective pitch detection method for noisy speech signal is proposed in this paper. Firstly, the EMD soft-threshold de-noising method is applied to realize the background noise reduction, secondly, using the Hilbert-Huang transform to detect the pitch period of the de-noising speech signal. The analysis proposed in this paper show that, compared with the conventional methods of the pitch detection of the noisy speech, especially for the low signal to noise ratio (SNR), this approach has a higher accuracy.

• Periodical: Advanced Materials Research (Volumes 433-440)
• Pages: 4675-4678
• DOI: https://doi.org/10.4028/www.scientific.net/AMR.433-440.4675
• Online since: January 2012
• Authors: Hong Yan Xing, Cui Hua Yu, Peng Li
• Keywords: EMD Soft-Threshold De-Noising Method, Hilbert-Huang Transform (HHT), Pitch Detection

References

[1] Zhao Zh D, Hu X M, Tian J F. An effective pitch detection method for speech signals with low signal- to –noise ratio,. Proceedings of the Seventh International Conference on Machine Learning and Cybernetics, Kun ming, 12-15 July 2008: 2775-2778.

DOI: 10.1109/icmlc.2008.4620879

[2] L. R. Rabiner. On the use of the autocorrelation analysis for pitch detection, . IEEE Transactions on Acoustics Speech and Signal Processing, 1977, 25(1): 24-33.

DOI: 10.1109/tassp.1977.1162905

[3] M. J Ross, H L Shaffer, A Cohen, et al. Average magnitude difference function pitch extractor". IEEE Transactions on Acoustics Speech Signal Processing. 1974, 22(5): 353-362.

DOI: 10.1109/tassp.1974.1162598

[4] S Kadame, G. F Boudreaux-Bartels. Application of wavelet transform for pitch detection,. IEEE Trans on IT, 1992, 38(2): 917-924.

DOI: 10.1109/18.119752

[5] Huang N E, Shen Zh, Long S R, etc. The empirical mode decomposition and the Hilbert spectrum for nonlinear and non-stationary time series analysis [J]. Proc. R. Soc. Lond A 1998, 454: 903-995.

DOI: 10.1098/rspa.1998.0193

[6] Li R L, Bao Ch Ch, Dou H J. Pitch Detection Method for Noisy Speech Signals Based on Pre-Filter and Weighted Wavelet coefficients,. ICSP2008 Proceedings: 530-533.

DOI: 10.1109/icosp.2008.4697187

[7] Du X L, He L ZH, Hou W. A study of Wavelet Threshold Denoising Based on Empirical Mode Decomposition (EMD) [J]. Journal Of Beijing University of Technology, 2007, 33(3): 265-272. (in Chinese).

[8] DONOHO D L. De-noising by soft- thresholding [J]. IEEE Transaction on Information, 1995, 41 (3): 623-627.