/*  ----------------------------------------------------------------	Speech Signal Processing Toolkit (SPTK): version 3.0			 SPTK Working Group		   Department of Computer Science		   Nagoya Institute of Technology				and    Interdisciplinary Graduate School of Science and Engineering		   Tokyo Institute of Technology		      Copyright (c) 1984-2000			All Rights Reserved.  Permission is hereby granted, free of charge, to use and  distribute this software and its documentation without  restriction, including without limitation the rights to use,  copy, modify, merge, publish, distribute, sublicense, and/or  sell copies of this work, and to permit persons to whom this  work is furnished to do so, subject to the following conditions:    1. The code must retain the above copyright notice, this list       of conditions and the following disclaimer.    2. Any modifications must be clearly marked as such.  NAGOYA INSTITUTE OF TECHNOLOGY, TOKYO INSITITUTE OF TECHNOLOGY,  SPTK WORKING GROUP, AND THE CONTRIBUTORS TO THIS WORK DISCLAIM  ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, INCLUDING ALL  IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO EVENT  SHALL NAGOYA INSTITUTE OF TECHNOLOGY, TOKYO INSITITUTE OF  TECHNOLOGY, SPTK WORKING GROUP, NOR THE CONTRIBUTORS BE LIABLE  FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY  DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS,  WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS  ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR  PERFORMANCE OF THIS SOFTWARE. ----------------------------------------------------------------*//****************************************************************    $Id: _amcep.c,v 1.2 2002/12/25 05:28:28 sako Exp $    Adaptive Mel Cepstral Analysis	double amcep(x, b, m, a, lambda, step, tau, pd, eps);	double   x      : input sequence	double   *b     : MLSA filter coefficients	int      m      : order of cepstrum	double   a 	: alpha	double   lambda : leakage factor	double   step   : step size	double   tau    : momentum constant	int      pd     : order of pade approximation	double   eps	: minimum value for epsilon	return value  : prediction error*****************************************************************/#include <stdio.h>#include <SPTK.h>double amcep(x, b, m, a, lambda, step, tau, pd, eps)double x, *b, a, lambda, step, tau, eps;int m, pd;{    register int   i;    static double  *bb = NULL, *d, *ep, *e, xx, gg = 1.0;    static int     size;    double 	   tt, mu, tx, mlsadf(), log();        if(bb == NULL){	bb = dgetmem(3*(m+1)+3*(pd+1)+pd*(m+2));	e  = bb + m + 1;	ep = e  + m + 1;	d  = ep + m + 1;	size = m;    }    if(m > size){	free(bb);	bb = dgetmem(3*(m+1)+3*(pd+1)+pd*(m+2));	e  = bb + m + 1;	ep = e  + m + 1;	d  = ep + m + 1;	size = m;    }	    for(i=1; i<=m; i++)	bb[i] = -b[i];        x = mlsadf(x, bb, m, a, pd, d);    phidf(xx, m, a, e);    xx = x;    gg = gg * lambda + (1.0 - lambda) * x * x;    gg = ( gg < eps )? eps : gg;    b[0] = 0.5 * log(gg);    mu = step / (double) m / gg;    tx = 2 * (1.0 - tau) * x;        for(i=1; i<=m; i++){	ep[i] = tau * ep[i] - tx * e[i];	b[i] -= mu * ep[i];    }        return(x);}phidf(x, m, a, d)double x, a, *d;int m;{    register int   i;    d[0] = a * d[0] + (1.0 - a*a) * x;        for(i=1; i<m; i++)	d[i] += a * (d[i+1] - d[i-1]);    for(i=m; i>=1; i--) 	d[i] = d[i-1];}