Re: Curious problem with UnitStep
- To: mathgroup at smc.vnet.net
- Subject: [mg67909] Re: Curious problem with UnitStep
- From: dh <dh at metrohm.ch>
- Date: Wed, 12 Jul 2006 05:06:17 -0400 (EDT)
- References: <e8vtkn$sp0$1@smc.vnet.net>
- Sender: owner-wri-mathgroup at wolfram.com
Hi Hanspeter,
try:Trace[akf[t, 1.0]]
and you will see that the variable t is captured by the dummy variable
of the integral. This can e.g. be fixed by using a local variable:
akf[x_, Ï?_] := Module[{t},...]
Daniel
hanspi wrote:
> Dear colleagues,
>
> I am trying to calculate the ACF of a combination of unit steps and its
> fourier transform. There I see something very curious.
>
> I do:
>
> sw[t_, \[Tau]_] := UnitStep[\[Tau] - Abs[t]]
> Plot[sw[t, 1.0], {t, -2, 2}]
>
> which looks OK. Then:
>
> akf[x_, \[Tau]_] :=
> Assuming[{t \[Epsilon] Reals, \[Tau] > 0},
> Integrate[
> sw[t, \[Tau]] sw[x - t, \[Tau]], {t, -\[Infinity], \[Infinity]}]]
> Plot[akf[t, 1.0], {t, -2, 2}]
>
> Which also looks OK.
>
> But when I just type
>
> akf[t, 1.0]
>
> I get as an output
>
> 2.
>
> which is wrong, and also
>
> FourierTransform[akf[t, 1.0], t, \[Omega]]
>
> gives a wrong result. Why can I plot akf properly, but not evaluate
> it?
>
> Hanspeter
>