Title :
Improvement of pitch estimation using harmonic wavelet transform
Author :
Chisaki, Y. ; Usagawa, T. ; Ebata, M.
Author_Institution :
Dept. of Comput. & Sci., Kumamoto Univ., Japan
Abstract :
A pitch detection method for a harmonic signal based on the wavelet transform is proposed. A modified wavelet transform is derived from a general continuous wavelet transform. Errors of pitch estimation obtained by the proposed method are presented. Moreover, the authors examine how the frequency and phase errors in analyzing the wavelet against the observed signal affect the accuracy of the pitch estimation. Furthermore, the pitch estimation is performed for vowels. As a result, the pitch peak power is enhanced. So, the pitch frequency is estimated with ease among candidates for the power power
Keywords :
bandlimited signals; harmonic analysis; parameter estimation; signal detection; speech enhancement; wavelet transforms; band-limited speech signal; frequency error; general continuous wavelet transform; generalized harmonic analysis; harmonic signal; harmonic wavelet transform; modified wavelet transform; observed signal; phase error; pitch detection; pitch estimation accuracy; pitch estimation errors; pitch frequency; pitch peak power; speech enhancement; vowels; wavelet analysis; Continuous wavelet transforms; Equations; Frequency estimation; Harmonic analysis; Power harmonic filters; Signal analysis; Signal resolution; Speech analysis; Wavelet analysis; Wavelet transforms;
Conference_Titel :
TENCON 99. Proceedings of the IEEE Region 10 Conference
Conference_Location :
Cheju Island
Print_ISBN :
0-7803-5739-6
DOI :
10.1109/TENCON.1999.818486
