From help-request at octave dot org Mon Jan 24 05:18:35 2005
Subject: Re: question about FFT computation
From: =?iso-8859-1?Q?"J=F6rg=20Kampmann"?= <joerg-at-ibk at web dot de>
To: help at octave dot org
Date: Mon, 24 Jan 2005 12:22:40 +0100

"J=F6rg Kampmann" <joerg-at-ibk at web dot de> schrieb am 24.01.05 11:58:15:
>=20
>=20
> "Cesare CRUCCAS" <cesare dot cruccas at samcef dot com> schrieb am 24.01.05 09:23:2=
1:
> >=20
> > Good morning Joerg,
> >=20
> > I've found your name in the mailing list of octave about FFT...
> >=20
> > I've need to make a FFT computation of a signal,=20
> > but with the particularity that the times aren't equidistants.=20
> > =20
> > I've two vectors of dimension N.=20
> > The first is the times (i.e. T=3D[0, 0.01, 0.04, ...])=20
> > and the second is another thing (i.e. Accel=3D[0, 4500, 4300, ...])
> >=20
> > Could You help me=3F
>=20
> thanks for the message, ad hoc I would say:
>=20
> Interpolate the signals (e.g. with bezier or any other interpolation). A=
nd make an equidistant sample series
> from that.=20
>=20
> When non-equidistancy is not that grave, i.e. when the sampling-points c=
an be considered as statistically
> equidistant, you should try to leave it as is ...
>=20
> However if it is a real-time problem you possibly have to consider also =
these issues, maybe a RTOS like QNX
> www.qnx.com
>=20
>=20
--=20
IBK-Consult - D-31228 Peine
http://www.ibk-consult.de



-------------------------------------------------------------
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
-------------------------------------------------------------