matrixinv.3

speech signal process tools

.\" Copyright (c) 1987-1990 Entropic Speech, Inc.
.\" Copyright (c) 1997 Entropic Research Laboratory, Inc. All rights reserved.
.\" @(#)matrixinv.3	1.9 18 Apr 1997 ESI/ERL
.ds ]W (c) 1997 Entropic Research Laboratory, Inc.
.TH MATRIX_INV 3\-ESPSsp 18 Apr 1997
.SH "NAME"
matrix_inv \- invert a matrix with single precision
.br
matrix_inv_d \- invert a matrix with double precision
.SH "SYNOPSIS"
.B #include <stdio.h>
.br
.B double matrix_inv (mat_in, inv_out, order)
.br
.B float **mat_in;
.br
.B float **inv_out;
.br
.B int order;

.br
.B double matrix_inv_d (mat_in, inv_out, order)
.br
.B double **mat_in;
.br
.B double **inv_out;
.br
.B int order;
.SH "DESCRIPTION"
.PP
.I matrix_inv
inverts an input matrix
.I mat_in
of order
.I order
and stores the result in
.I inv_out.
If the input matrix is singular to working precision, then the output
matrix
.I inv_out
is left unchanged.
.PP
The two parameters, \fImat_in\fP and \fIinv_out\fP, are interpreted as
pointers to matrices of floats (or doubles)
containing \fIorder\fP rows and \fIorder\fP columns.  Space
for these matrices must be allocated by the calling program \- such a
pointer can be assigned by means of \fIarr_alloc\fP(3\-ESPSu).
.PP
.I matrix_inv
returns the condition number of the matrix.  If the condition number
is greater than 1e6 for
.I matrix_inv,
or greater than 1e14 for
.I matrix_inv_d,
then the matrix is considered singular and \-1.0
is returned.
.PP
.I matrix_inv_d
is the double precision version of
.I matrix_inv.
.SH "EXAMPLE"
.PP
int	order = 4;	/* order of matrix to invert */
.br
float	**matrix;	/* input matrix to invert */
.br
float	**inv_matrix;	/* inverse matrix */

/* allocate memory and initialize matrices */
.sp
if (matrix_inv (matrix, inv_matrix, order) == \-1.0)
.br
   /* input matrix is singular to working precision */
.br
else
.br
   /* inv_matrix contains the inverse of input matrix */
.br
.SH DIAGNOSTICS
.PP
.SH "BUGS"
.PP
None Known.
.SH "SEE ALSO"
.PP
.nf
\fIarr_alloc\fP(3-\ESPSu)
.fi
.SH REFERENCES
[1] Forsythe, G. E., M. A. Malcolm, C. B. Moler (1977), \fIComputer Methods for
Mathematical Computations\fP. New Jersey:  Prentice-Hall, Inc.,
51\-55.
.SH "AUTHOR"
.PP
Ajaipal S. Virdy.