Fast Fourier Transform Power Spectra

The FFT provides a computationally fast way to transfer between the time and frequency domains.

Power density spectrum with data plotted by frequency. Note the general overall steady decrease in power with increasing frequency, interrupted by sharp spikes where the series is periodic.
The same power density spectrum with data plotted by period. Note the general overall steady increase in power with increasing period, interrupted by sharp spikes where the series is periodic.

 


Table with the data from the plots. Note the general decline in frequency with increasing bin number, but that periods around 12 and 24 hours have anomalously high power, suggesting periodicity in the data at those periods.

Bin    Frequency   Period         Power
    1       0.0    infinite    1.348073
    3   0.00293  341.333333    0.085273
    2   0.00195  512.000000    0.057467
    4   0.00391  256.000000    0.030551
    5   0.00488  204.800000    0.024552
    6   0.00586  170.666667    0.016627
    7   0.00684  146.285714    0.010556
    8   0.00781  128.000000    0.010149
   82   0.08008   12.487805    0.006268
    9   0.00879  113.777778    0.006149
   10   0.00977  102.400000    0.005995
   11   0.01074   93.090909    0.004356
   83   0.08105   12.337349    0.004074
   43   0.04199   23.813953    0.003910
   12   0.01172   85.333333    0.003597
   13   0.01270   78.769231    0.003324
   40   0.03906   25.600000    0.002679
   15   0.01465   68.266667    0.002505
   14   0.01367   73.142857    0.002443
   42   0.04102   24.380952    0.002357
   16   0.01563   64.000000    0.001971
   17   0.01660   60.235294    0.001869
   18   0.01758   56.888889    0.001701
   20   0.01953   51.200000    0.001432
   19   0.01855   53.894737    0.001422
   81   0.07910   12.641975    0.001408

Last revision 2/2/2009