2
$\begingroup$

I am interested in evaluating the discrete Fourier transform for a list of data. The problem is the results given by the function Fourier depends on the length of the data.

For example, if the list of data is simply Table[Sin[20 π t], {t, 0, 10, x}] for different values of x I will obtain the same shape from Fourier when I plot the obtained numbers but the values on the $x$ axis will change.

For $x=0.001$ I will get:

m1

But for $x=0.01$ I get:

different result

$\endgroup$
  • 1
    $\begingroup$ Possible duplicates: 105439, 33149 and 44237 $\endgroup$ – Lukas Lang Aug 26 '17 at 21:16
  • $\begingroup$ The answer to your question is here. $\endgroup$ – andre314 Aug 26 '17 at 22:04
6
$\begingroup$

The x axis of your plot has no meaning - it is simply the index of the value: The finer the data in your original table, the finer the frequency resolution and therefore the bigger the number of values in your output.

You can transform the axes however: The frequency spacing $\delta f$ and total bandwidth $\Delta f$ is defined the following way:

$\delta f=\frac{1}{\Delta t}$

$\Delta f=\frac{1}{\delta t},$

where $\delta t$ is the time domain spacing and $\Delta t$ is the total time span of the data.

Note: Despite the many duplicates (see comments), I thought it would be nice to have the relations written down somewhere without the need to extract them from code (as far as I could tell, they're not in any of the duplicates). It might be better to move this answer to one of the duplicates - if so, to which one should I move this?

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.