Title :
A robust linear prediction method for noisy speech
Author :
Shimamura, Tetsuya ; Kunieda, Nobuyuki ; Suzuki, Jouji
Author_Institution :
Dept. of Inf. & Comput. Sci., Saitama Univ., Urawa, Japan
fDate :
31 May-3 Jun 1998
Abstract :
This paper proposes a method to improve the performance of the LP method in a noisy environment. A relationship for the autocorrelation function (ACF) of a periodical signal is derived, and the ACF estimates of noisy speech are transformed into their noiseless ones. The normal equation in which the noiseless ACF is used is efficiently solved by the Levinson-Durbin algorithm. Computer simulations demonstrate that the proposed method provides an improvement in the LPC cepstrum distance
Keywords :
cepstral analysis; correlation methods; prediction theory; speech processing; LPC cepstrum distance; Levinson-Durbin algorithm; autocorrelation function; noisy speech; periodical signal; robust linear prediction method; Additive noise; Autocorrelation; Computer simulation; Equations; Phase measurement; Prediction methods; Robustness; Speech analysis; Time measurement; Working environment noise;
Conference_Titel :
Circuits and Systems, 1998. ISCAS '98. Proceedings of the 1998 IEEE International Symposium on
Conference_Location :
Monterey, CA
Print_ISBN :
0-7803-4455-3
DOI :
10.1109/ISCAS.1998.698809
