From help-request at octave dot org Tue Feb 28 11:27:27 2006
Subject: Re: Problems with FFT and IFFT
From: "Dmitri A. Sergatskov" <dasergatskov at gmail dot com>
To: "John W. Eaton" <jwe at bevo dot che dot wisc dot edu>
Cc: "Keith Goodman" <kwgoodman at gmail dot com>,         "Sascha Berkenkamp" <sberkenkamp@gmx.net>, help@octave.org
Date: Tue, 28 Feb 2006 10:25:11 -0700

On 2/28/06, John W. Eaton <jwe at bevo dot che dot wisc dot edu> wrote:
> On 28-Feb-2006, Keith Goodman wrote:
>
> | Here's what I get (no imaginary part; I get the same result in 2.1.72):
> |
> | $ octave
> | Octave 2.9.4 Forge 20050613
> | >> fs = 0:0.1:10;
> | >> x = sin(2*pi*fs);
> | >> real(x)
>
> What kind of hardware?  I see imaginary parts, but all less than eps.
> Perhaps due to extra internal precision of amd64/x86 floating point
> hardware?  Also, I am using fftw, are you?  Try
>
>    octave_config_info ("FFTW_LIBS")
>
> to see.
>
> In any case, while it might be nice to understand where the
> differences are coming from, I don't think this is a bug.

max(imag(x2) on Athlon64 is 2x as big:

octave:6> max(imag(x2))
ans =  2.57566273587379e-16
that is slightly larger than eps
octave:7> eps
eps =  2.22044604925031e-16

Is not Matlab using (hacked) fftw2 (while octave is using fftw3)?

>
> jwe
>

Dmitri.
--



-------------------------------------------------------------
Octave is freely available under the terms of the GNU GPL.

Octave's home on the web:  http://www.octave.org
How to fund new projects:  http://www.octave.org/funding.html
Subscription information:  http://www.octave.org/archive.html
-------------------------------------------------------------