Fast Fourier Transform, FFT analysis
According to Fourier theorem, any continuous signal can be decomposed as a sum of sines. The FFT analysis is a narrow band spectral analysis.
The Fourier transform is the representation of each sine as a single line in frequency domain:
- The Fast Fourier Transformation (also called FFT) is an algorithm enabling to efficiently transform a sampled signal in a frequency spectrum.
- This algorithm can be applied only on finite durations such as periodic signal. This is rare to satisfy this condition so windowing is often used to force the signal to be a periodic one. The choice of the window functions depends on the application: for rotating analysis the hanning window will be preferred, whereas for bump test a force/response window is more appropriate.
- Due to the windowing, area at the time block ends are not taken into account as the central area of the time block. Overlapping allows to minimize the windowing effect.
- Dynamic signal analyzers use digital signal processing to sample the input signal and convert it from the time domain to the frequency domain. This conversion is done using the FFT algorithm.