Thread Subject: Using fft function

I am trying to plot a fft spectrum for a stairs signal. For example:

time = 1:36;
mag = [1 2 3 4 5 6 6 6 6 6 5 4 3 2 1 0 0 0 1 2 3 4 5 6 6 6 6 6 5 4 3 2 1 0 0 0];
stairs(time, mag)

Question is how do I use this signal to create an fft spectrum? Any leads will be greatly appreciate.

"Timothy " <timothya@haywardtyler.com> wrote in message <hcsasr$e80$1@fred.mathworks.com>...
> I am trying to plot a fft spectrum for a stairs signal. For example:
>
> time = 1:36;
> mag = [1 2 3 4 5 6 6 6 6 6 5 4 3 2 1 0 0 0 1 2 3 4 5 6 6 6 6 6 5 4 3 2 1 0 0 0];
> stairs(time, mag)
>
> Question is how do I use this signal to create an fft spectrum? Any leads will be greatly appreciate.


using the fft command?

On Nov 5, 5:42 am, "Timothy " <timot...@haywardtyler.com> wrote:
> I am trying to plot a fft spectrum for a stairs signal. For example:
>
> time = 1:36;
> mag = [1 2 3 4 5 6 6 6 6 6 5 4 3 2 1 0 0 0 1 2 3 4 5 6 6 6 6 6 5 4 3 2 1 0 0 0];
> stairs(time, mag)
>
> Question is how do I use this signal to create an fft spectrum? Any leads will be greatly appreciate.

But fft won't see it as a stairs signal.
What it will see is:
plot(t,mag,'o')
For the stairs signal, you have two data points at each time stamp.
But fft requires data that are equispaced in time.

TideMan <mulgor@gmail.com> wrote in message <f074b9d6-8740-4fc2-bbb2-ba8dac9ba2bd@y10g2000prg.googlegroups.com>...
> On Nov 5, 5:42?am, "Timothy " <timot...@haywardtyler.com> wrote:
> > I am trying to plot a fft spectrum for a stairs signal. For example:
> >
> > time = 1:36;
> > mag = [1 2 3 4 5 6 6 6 6 6 5 4 3 2 1 0 0 0 1 2 3 4 5 6 6 6 6 6 5 4 3 2 1 0 0 0];
> > stairs(time, mag)
> >
> > Question is how do I use this signal to create an fft spectrum? Any leads will be greatly appreciate.
>
> But fft won't see it as a stairs signal.
> What it will see is:
> plot(t,mag,'o')
> For the stairs signal, you have two data points at each time stamp.
> But fft requires data that are equispaced in time.
>

Thank you for the reply. I was expecting Matlab to know how to calculate the fft for this signal, but I realised it is not straight forward. Do you have any way around this?

On Nov 5, 9:22 pm, "Timothy " <timot...@haywardtyler.com> wrote:
> TideMan <mul...@gmail.com> wrote in message <f074b9d6-8740-4fc2-bbb2-ba8dac9ba...@y10g2000prg.googlegroups.com>...
> > On Nov 5, 5:42?am, "Timothy " <timot...@haywardtyler.com> wrote:
> > > I am trying to plot a fft spectrum for a stairs signal. For example:
>
> > > time = 1:36;
> > > mag = [1 2 3 4 5 6 6 6 6 6 5 4 3 2 1 0 0 0 1 2 3 4 5 6 6 6 6 6 5 4 3 2 1 0 0 0];
> > > stairs(time, mag)
>
> > > Question is how do I use this signal to create an fft spectrum? Any leads will be greatly appreciate.
>
> > But fft won't see it as a stairs signal.
> > What it will see is:
> > plot(t,mag,'o')
> > For the stairs signal, you have two data points at each time stamp.
> > But fft requires data that are equispaced in time.
>
> Thank you for the reply. I was expecting Matlab to know how to calculate the fft for this signal, but I realised it is not straight forward. Do you have any way around this?

Why would you expect Matlab to know how to do such a thing?
The FFT is an algorithm that operates on equispaced time series data,
where time is monotonically increasing.
But you want to transform data in which time is not monotonically
increasing, so the FFT doesn't work, no matter which package you use:
Fortran, C++, Python, whatever.