Title :
Optimal speech data reduction using a stochastic realization algorithm
Author :
Mahalanabis, A.K. ; Kanai, Teruto ; Kanai, T.
Author_Institution :
Pennsylvania State University, University Park, PA
Abstract :
The paper discusses some results of application of a stochastic realization algorithm for developing an ARMA type predictor for a class of speech signals. It is shown that a minimal order adaptive predictor can be obtained for the given speech data using a finite window version of the realization algorithm. The utility of the predictor is first established by processing synthetic signals generated by an ARMA
Keywords :
Covariance matrix; Gaussian noise; Kalman filters; Parameter estimation; Predictive models; Signal generators; Signal processing; Speech processing; State estimation; Stochastic processes;
Conference_Titel :
Acoustics, Speech, and Signal Processing, IEEE International Conference on ICASSP '86.
DOI :
10.1109/ICASSP.1986.1168816
