Summary
Smart Info-Media Systems in Asia
2019
Session Number:RS1
Session:
Number:RS1-2
Refined Autocorrelation Function for Pitch Detection of Speech
Md. Saifur Rahman, Yosuke Sugiura, Tetsuya Shimamura,
pp.72-77
Publication Date:2019/9/4
Online ISSN:2188-5079
DOI:10.34385/proc.57.RS1-2
PDF download (1.3MB)
Summary:
In this paper, a noise robust method for pitch extraction using the autocorrelation function (ACF) and homomorphic deconvolution is proposed. In the proposed method, we extract the vocal source information eliminating the vocal tract information in the autocorrelation domain. The accuracy of pitch extraction is significantly enhanced in an efficient way based on the ACF calculation. Through experiments superior performance of the proposed method relative to some conventional methods is shown.