Path: utzoo!utgpu!utstat!jarvis.csri.toronto.edu!mailrus!uflorida!stat!fsu!prism!mg27
From: mg27@prism.gatech.EDU (Michael J. Gourlay)
Newsgroups: comp.lang.fortran
Subject: Re: Fourier Transformation Routine Needed
Message-ID: <3274@hydra.gatech.EDU>
Date: 9 Nov 89 20:23:30 GMT
References: <1021@pacific.mps.ohio-state.edu>
Reply-To: mg27@prism.gatech.EDU (Michael J. Gourlay)
Organization: Georgia Institute of Technology
Lines: 30

sun@pacific.mps.ohio-state.edu (Mingqiu Sun) writes:
)recently my research project is involved in a Fourier Transformation
)integration. it is just integrate over f(x)cos(x), where f(x) is a 
)relatively smooth function. i tried a Fourier transformation routine
)in the IMSL liabrary. however it is extremely slow. do netters have
)similar experience? i need a good routine badly. all suggestions are
)welcome.

Have you heard of a book called Numerical Recipies by Press, Flannery,
Teukolsky, and Vetterling?  There are versions of the book for ForTran,
and other languages. Published by University of Cambridge Press. The
book optionally comes with a disk for IBM PC/AT/etc, which you could
port to your Mainfraim, or whatever. I'd upload it, but for one thing,
it's against the copyright, and also, I can't seem to get my transfer
program to work right now. 

I highly recommend the book. It has an entire chapter on Fourier Transform
Spectral Methods, and there are sin and cos tranforms in there, too. The
routines are fast.

May I also suggest IEEE transoactions on Antennas and Propogation, Vol 37,
number 3, March 1989, pp 404-6, which explains how to do such integrals
with higher precision.

                          my opinions
-- 
Michael J. Gourlay					      GT Box 35431
Georgia Institute of Technology, School of Physics, Atlanta Georgia, 30332
uucp:	  ...!{decvax,hplabs,ncar,purdue,rutgers}!gatech!prism!gt5431b
Internet: gt5431b@prism.gatech.edu