A trumpet note, a human scream, and pure electrical static all look, plotted as waveforms over time, like completely unrelated shapes — one smoothly periodic, one chaotic and shrieking, one an unpredictable jumble. The Fourier transform, one of the foundational tools in signals and systems engineering, reveals that this apparent unrelatedness is superficial: every one of these signals, no matter how complex it looks, can be broken down mathematically into a sum of simple sine waves at different frequencies, added together in different amounts.
Any signal, decomposed into a recipe of pure tones
The core claim behind the Fourier transform is that a wide class of signals, however complicated they appear in the time domain (the ordinary way of plotting a signal as it changes moment to moment), can be represented equivalently in the frequency domain — as a specific combination of pure sine waves, each at its own frequency and amplitude, that when added together perfectly reconstruct the original signal. A trumpet's rich, characteristic tone is actually a fundamental frequency plus a specific set of higher harmonics layered on top, each contributing a particular amount; that specific harmonic recipe is exactly what distinguishes a trumpet's sound from a flute playing the identical musical note. The Fourier transform is the mathematical tool that calculates that recipe for any given signal, converting a description of "how the signal changes over time" into an equivalent description of "which frequencies are present, and how much of each."
Why decomposing into frequencies is so useful
This frequency-domain view isn't just a mathematical curiosity — many practical problems in engineering become dramatically easier once a signal is viewed this way. Removing a specific unwanted frequency, like an electrical hum at a fixed frequency, from an audio recording is far more straightforward in the frequency domain, where that hum shows up as an isolated spike that can be directly identified and removed, than in the tangled time-domain waveform, where the hum is smeared invisibly throughout the entire signal. Compression, filtering, and countless other signal-processing techniques rely on this same underlying idea: a change that would be complicated to express in the time domain often becomes simple and mechanical once expressed in the frequency domain instead, which is exactly why the Fourier transform sits at the mathematical core of an enormous range of technologies, from audio processing to image compression to wireless communication.
What we're still unsure about
The mathematics of the Fourier transform, and its ability to decompose signals into frequency components, are completely settled, rigorously proven mathematics with an enormous body of successful engineering application behind them — there's no scientific dispute about the core technique. What remains a genuinely active area of engineering practice is choosing the right variant and implementation of frequency-domain analysis for a specific real-world problem, since practical signals are finite, noisy, and often changing character over time in ways the idealised mathematical theory doesn't directly address, and selecting the appropriate windowing, sampling, and transform technique for a given application remains a matter of engineering judgement rather than one fixed formula that works everywhere.
This sits inside Signals & Systems (Fourier, Laplace Transforms), one of eight topics in Electrical Engineering, one of four domains in Engineering, one of seventeen subjects the app can quiz you on.