Fast Fourier Transform (FFT) Animation using Matlab

We show the progress of Fast Fourier Transform (FFT) of a time-domain signal as it changes in time. Matlab's fft() function is used for illustration, hence it should be noted that the function is assumed to be periodic. The number of FFT point is assumed to be same as the time domain signal to prevent zero padding when the full domain is filled. In the beginning a sinusoidal signal at 50 Hertz (1*sin(2pi*50*t)) starts to develop and slowly fills the full domain. Then, another sinusoidal signal at a higher amplitude and 100 Hertz (1.5*sin(2pi*100*t)) is added to this signal. Later, a third sinusoidal signal at 200 Hz replaces the 100 Hz one. Following this, a DC component (0 Hz) is inserted and finally, the DC component is removed. At each stage, the the development of frequency components can clearly be observed. In the frequency spectrum, first a 50Hz component starts to build up along with lower and higher frequencies. The reason for wider spectrum is the fact that sinusoidal signal fills the domain within an increasing time window (hence introduction of sinc components). Once the full domain is filled with the sinusoidal, then only 50 Hz component remains due to the fact that the signal assumes to be periodic in Matlab fft function.