dlagf01

Fourier cosine and sine transforms using lagged convolutions in double-precision (dlagf0 and dlagf1)

Conversions of these routines into Matlab.From:http://pubs.usgs.gov/of/1995/ofr-95-0077/of-95-77.htmlAnderson, W.L, 1983, Fourier cosine and sine transforms using lagged convolutions in double-precision (Subprograms DLAGFO/DLAGF1): U.S. Geological Survey Open-File Report 83-320, 38 p. 026 DLAGF0/ DLAGF1 Fourier cosine/sine transforms (in double-precision). Numerical integration by lagged convolution. USGS Open-File Report 83-320 (1983) Fortran-77:VAX-11/780 (VMS) Subroutine Package

%%%%%%%%%%%%%
From the help:
%C THE PURPOSE OF SUBPROGRAM DLAGF0 IS TO PROVIDE IN DOUBLE PRECISION
%C A GENERAL ALGORITHM FOR FAST FOURIER COSINE TRANSFORMS OF THE FORM:
%C INTEGRAL FROM 0 TO INFINITY OF FUN(G)*DCOS(B*G)*DG, WHERE WE LAG B
%C IN[BMIN,BMAX]--USING IMPLIED VALUE OF BMIN=BMAX*DEXP(-.1*(NB-1)).
%C
%C DLAGF0 USES A NEW DOUBLE-PRECISION COSINE DIGITAL FILTER FOR USE
%C IN LAGGED CONVOLUTION, WHERE THE ALGORITHM IS PATTERNED AFTER THE
%C DOUBLE-PRECISION HANKEL TRANSFORM ROUTINE DHANKL PUBLISHED IN THE
%C FOLLOWING REFERENCE:
%C
%C ANDERSON, W.L., 1982, FAST HANKEL TRANSFORMS USING RELATED AND LAGGED
%C CONVOLUTIONS: ACM TRANS. ON MATH. SOFTWARE, V.8, N.4, P.344-368.
%C
%C DLAGF0 USES THE LAGGED CONVOLUTION METHOD IN THE ABOVE REFERENCE,
%C BUT USES INTERNAL WORK ARRAYS DWORK(787) AND KEY(787)--INSTEAD OF
%C PASSING DWORK IN THE CALLING SEQUENCE AS IN DHANKL. THE MINIMUM
%C RELATIVE AND/OR ABSOLUTE ERRORS ARE ABOUT 10**-10 USING TOL<=10**-12.