Lesson 6 of 30 intermediate 4 min read

The Fourier transform

Key takeaways The Fourier transform re-describes a signal from the time domain (samples over time) into the frequency domain (how much of each frequency is present). Its foundation: any signal is a sum of sine waves. In the frequency domain, jobs like isolating a channel or spotting a carrier become easy — which is why the transform is DSP’s most-used tool.

This lesson introduces the single most powerful idea in signal processing. Don’t worry about the equations — the concept is what unlocks everything after it.

Any signal is a sum of sines

Here’s the surprising fact underneath all of it: every signal, however complicated, can be built by adding together plain sine waves of different frequencies, strengths, and phases. A square wave is a sum of sines; a voice is a sum of sines; a digital transmission is a sum of sines.

The Fourier transform runs that idea in reverse: given the signal, it works out which sines, and how much of each, are in the mix. That recipe — amount of energy at each frequency — is the signal’s spectrum.

Two views of the same thing

The same signal, two views. That messy waveform on the left is just two tones — obvious as two peaks on the right.

The waveform and its spectrum hold the same information — but some questions are trivial in one view and painful in the other. “Is there a carrier near 851 MHz?” is almost impossible to eyeball in the time domain and blindingly obvious as a peak in the frequency domain.

Why radio lives in the frequency domain

Nearly every radio task is naturally a frequency question:

Finding a signal — a control channel is a peak at a known frequency.
Isolating a channel — keep a band of frequencies, discard the rest (that’s filtering).
Tuning — shift the whole spectrum so your channel sits at zero (that’s mixing).

There’s even a deep shortcut the transform reveals: filtering in the time domain (convolution) is just multiplication in the frequency domain — the link you’ll meet in convolution & impulse response.

From idea to computation

The Fourier transform as pure math runs over a continuous signal. On sampled data we use the Discrete Fourier Transform (DFT), which takes a block of samples and returns the energy in a set of frequency bins. Computed naively it’s slow — but a clever algorithm, the FFT, makes it fast enough to run in real time on a live radio stream. That’s the next lesson, and it’s what powers the waterfall display you may have seen.

Quick check: the Fourier transform turns a signal described over time into one described over…

Recap

Any signal is a sum of sine waves; the Fourier transform finds the recipe.
It converts the time domain (samples) into the frequency domain (a spectrum).
Radio tasks — finding, isolating, tuning — are naturally frequency questions.
On sampled data we use the DFT, made fast by the FFT — coming next.

Next up: the FFT, and what a waterfall display is really showing you.

Frequently asked questions

What does the Fourier transform actually do?

It takes a signal described over time (a list of samples) and re-describes it over frequency — telling you how much of each frequency is present. The remarkable underlying fact is that any signal can be built by adding up sine waves of different frequencies, amplitudes, and phases, and the transform finds that recipe.

What is the difference between the time domain and the frequency domain?

The time domain is the signal as it varies moment to moment — the raw samples. The frequency domain is the same signal expressed as its component frequencies. They contain the same information in two views; some operations (like removing an interfering tone) are trivial in one view and hard in the other.