Title :
Pole-zero estimation of speech based on L1 norm linear prediction
Author :
Namba, Munehiro ; Kamata, Hiroyuki ; Ishida, Yoshihisa
Author_Institution :
Dept. of Electron. & Commun., Meiji Univ., Kawasaki, Japan
Abstract :
This paper discusses the robustness of the linear prediction whose minimization process is based on a least absolute error criterion. Due to such robustness in the preciseness of prediction, we can estimate a pole-zero model of speech signals and approximate better the spectral envelope compared to that obtained by traditional linear prediction
Keywords :
error analysis; least squares approximations; linear programming; minimisation; poles and zeros; prediction theory; spectral analysis; speech processing; L1 norm linear prediction; least absolute error criterion; linear programming; minimization process; pole-zero estimation; robustness; spectral envelope; speech; Electrical capacitance tomography; Frequency domain analysis; Minimization methods; Poles and zeros; Prediction algorithms; Predictive models; Robustness; Speech analysis;
Conference_Titel :
Communications, Computers, and Signal Processing, 1995. Proceedings., IEEE Pacific Rim Conference on
Conference_Location :
Victoria, BC
Print_ISBN :
0-7803-2553-2
DOI :
10.1109/PACRIM.1995.519595