 Windowing functions such as Han, Bartlett, Tuki or Hamming Windows to name but four fade a signal in at the beginning of the window and fade it out again at the end of the window. Windowing functions are especially important in the FFT. The number of samples you can feed into the FFT, known as the FFT size, has to be a power of two. But your signal is probably a lot longer than the size of your FFT, so you have to cut it up into small FFT-sized pieces called frames and run the FFT on each individual frame. If we fed an unwindowed frame into the FFT, then that frame would end abruptly and the FFT would think that there are high frequencies present in your signal that aren't really there. They're just a consequence for the fact that you cut your signal up. To reduce this problem, windowing functions ensure that each frame begins and ends at zero or nearly zero, depending on which windowing function you used.