Thread Subject:

fft of spatial coordinate domain

I have a large set of spatial coordinate points. I want to take the FFT of a set of grid points in a line in order to get f(x) where x is the wavenumber domain.

I've tried using meshgrid and linspace to get a function of the spatial coordinates and applying the FFT to it. However, I'm getting all 0's as my answer. Can anyone offer any suggestions on how to apply an FFT along a linear set of data points.

Thank you.

On Apr 7, 10:13=A0am, "Eric " <ericbch...@hotmail.com> wrote:
> I have a large set of spatial coordinate points. =A0I want to take the FF=
T of a set of grid points in a line in order to get f(x) where x is the wav=
enumber domain. =A0
>
> I've tried using meshgrid and linspace to get a function of the spatial c=
oordinates and applying the FFT to it. =A0However, I'm getting all 0's as m=
y answer. =A0Can anyone offer any suggestions on how to apply an FFT along =
a linear set of data points. =A0
>
> Thank you. =A0

---------------------------------------------------------------------------=
--------------------------------------------
You need to have a signal. If all you have is coordinates, and the
value at the coordinates is all zero, then you will get all zeros in
your Fourier domain. What is your signal?

I also don't exactly understand your phrase "apply an FFT along a
linear set of data points." You can take the FFT "of" a set of data,
but I'm not sure what you mean when you say you "apply" the FFT
"along" a linear set. Are you simply trying to take the 1D FFT of a
set of data that has been generated by sampling a signal linearly in
space?

ImageAnalyst <imageanalyst@mailinator.com> wrote in message <bff9aac2-c48c-45d2-8ad2-1689f455efe4@s12g2000prc.googlegroups.com>...
> On Apr 7, 10:13=A0am, "Eric " <ericbch...@hotmail.com> wrote:
> > I have a large set of spatial coordinate points. =A0I want to take the FF=
> T of a set of grid points in a line in order to get f(x) where x is the wav=
> enumber domain. =A0
> >
> > I've tried using meshgrid and linspace to get a function of the spatial c=
> oordinates and applying the FFT to it. =A0However, I'm getting all 0's as m=
> y answer. =A0Can anyone offer any suggestions on how to apply an FFT along =
> a linear set of data points. =A0
> >
> > Thank you. =A0
>
> ---------------------------------------------------------------------------=
> --------------------------------------------
> You need to have a signal. If all you have is coordinates, and the
> value at the coordinates is all zero, then you will get all zeros in
> your Fourier domain. What is your signal?
>
> I also don't exactly understand your phrase "apply an FFT along a
> linear set of data points." You can take the FFT "of" a set of data,
> but I'm not sure what you mean when you say you "apply" the FFT
> "along" a linear set. Are you simply trying to take the 1D FFT of a
> set of data that has been generated by sampling a signal linearly in
> space?

Yes I have a set of coordinate points and set of time response values. I want to take the 1D FFT of straight line of data points. Although I am finding that this yields a dirac delta function.
I also want to be able to take an FFT of a set of time history data in order to get the response in the frequency domain.

"Eric " <ericbchang@hotmail.com> wrote in message <grg5o1$8v7$1@fred.mathworks.com>...
> ImageAnalyst <imageanalyst@mailinator.com> wrote in message <bff9aac2-c48c-45d2-8ad2-1689f455efe4@s12g2000prc.googlegroups.com>...
> > On Apr 7, 10:13=A0am, "Eric " <ericbch...@hotmail.com> wrote:
> > > I have a large set of spatial coordinate points. =A0I want to take the FF=
> > T of a set of grid points in a line in order to get f(x) where x is the wav=
> > enumber domain. =A0
> > >
> > > I've tried using meshgrid and linspace to get a function of the spatial c=
> > oordinates and applying the FFT to it. =A0However, I'm getting all 0's as m=
> > y answer. =A0Can anyone offer any suggestions on how to apply an FFT along =
> > a linear set of data points. =A0
> > >
> > > Thank you. =A0
> >
> > ---------------------------------------------------------------------------=
> > --------------------------------------------
> > You need to have a signal. If all you have is coordinates, and the
> > value at the coordinates is all zero, then you will get all zeros in
> > your Fourier domain. What is your signal?
> >
> > I also don't exactly understand your phrase "apply an FFT along a
> > linear set of data points." You can take the FFT "of" a set of data,
> > but I'm not sure what you mean when you say you "apply" the FFT
> > "along" a linear set. Are you simply trying to take the 1D FFT of a
> > set of data that has been generated by sampling a signal linearly in
> > space?
>
> Yes I have a set of coordinate points and set of time response values. I want to take the 1D FFT of straight line of data points. Although I am finding that this yields a dirac delta function.
------------------------------

Then your data points are probably constant.



> I also want to be able to take an FFT of a set of time history data in order to get the response in the frequency domain.
----------------------------

Why would the normal fft() function not allow you to do that?

On Apr 7, 2:24=A0pm, "Eric " <ericbch...@hotmail.com> wrote:
> ImageAnalyst <imageanal...@mailinator.com> wrote in message <bff9aac2-c48=
c-45d2-8ad2-1689f455e...@s12g2000prc.googlegroups.com>...
> > On Apr 7, 10:13=3DA0am, "Eric " <ericbch...@hotmail.com> wrote:
> > > I have a large set of spatial coordinate points. =3DA0I want to take =
the FF=3D
> > T of a set of grid points in a line in order to get f(x) where x is the=
 wav=3D
> > enumber domain. =3DA0
>
> > > I've tried using meshgrid and linspace to get a function of the spati=
al c=3D
> > oordinates and applying the FFT to it. =3DA0However, I'm getting all 0'=
s as m=3D
> > y answer. =3DA0Can anyone offer any suggestions on how to apply an FFT =
along =3D
> > a linear set of data points. =3DA0
>
> > > Thank you. =3DA0
>
> > -----------------------------------------------------------------------=
----=AD=3D
> > --------------------------------------------
> > You need to have a signal. =A0If all you have is coordinates, and the
> > value at the coordinates is all zero, then you will get all zeros in
> > your Fourier domain. =A0What is your signal?
>
> > I also don't exactly understand your phrase "apply an FFT along a
> > linear set of data points." =A0You can take the FFT "of" a set of data,
> > but I'm not sure what you mean when you say you "apply" the FFT
> > "along" a linear set. =A0Are you simply trying to take the 1D FFT of a
> > set of data that has been generated by sampling a signal linearly in
> > space?
>
> Yes I have a set of coordinate points and set of time response values. =
=A0I want to take the 1D FFT of straight line of data points. =A0Although I=
 am finding that this yields a dirac delta function. =A0

---------------------------------------------------------------------------=
------------------------------
Yep. Sounds right. You might even also see a negative delta function
since the derivative of a delta function is a positive and a negative
delta function. See http://mathworld.wolfram.com/FourierTransformRampFunct=
ion.html
Regards,
ImageAnalyst