Fourier transform

The Fourier transform takes a signal written as a shape in time, x(t), and re-expresses it as a recipe of frequencies, X(f) — telling you exactly which pure tones, and how strong each one is, are hidden inside. It's like passing white light through a prism: the prism doesn't change the light, it just spreads it into its rainbow of constituent colours. The Fourier transform is the prism for any signal, splitting a complicated wiggle into the spectrum of frequencies it's made of.

This dual view is the most powerful trick in engineering. A bass rumble and a treble hiss overlap hopelessly in time but sit in completely different places in frequency, so to remove the hiss you switch to the frequency domain, cut the high band, and switch back — that's a filter. The transform also turns the painful convolution operation into plain multiplication, X(f)·H(f), which is why almost every audio, image, radio, and Wi-Fi system computes spectra (usually via the fast FFT) millions of times a second.