Side-by-side comparison
| Parameter | Fourier Series | Fourier Transform |
|---|---|---|
| Input signal type | Periodic signals with period T0 | Aperiodic (non-periodic) signals; also generalised for periodic |
| Output spectrum | Discrete line spectrum at harmonics nf0 | Continuous spectrum X(f) or X(jω) defined for all frequencies |
| Mathematical form | x(t) = Σ cn·e^{j2πnf0t}, n = −∞ to ∞ | X(f) = ∫_{-∞}^{∞} x(t)·e^{-j2πft} dt |
| Coefficients / values | Discrete complex coefficients cn | Continuous complex function X(f) |
| Fundamental requirement | Signal must be periodic (Dirichlet conditions) | Signal must be absolutely integrable or square-integrable |
| Energy/power signal | Applied to power signals (periodic → infinite energy) | Applied to energy signals (finite energy) |
| Relationship | cn = (1/T0)·X(f)|_{f=n/T0} — FS coeffs are samples of FT | FT of periodic signal gives impulses: Σ cn·δ(f − nf0) |
| Real example | Harmonic analysis of 230 V, 50 Hz AC waveform distortion | Spectrum of a single RADAR pulse or a Gaussian window |
| Gibbs phenomenon | Overshoot at discontinuities when series is truncated | Does not apply directly; windowing introduces spectral leakage |
| MATLAB function | Computed via FFT on one period; manual cn formula | fft() on aperiodic signal or freqz() for system response |
Key differences
Fourier Series decomposes a periodic signal into harmonics at integer multiples of f0 = 1/T0; the spectrum is discrete. Fourier Transform handles aperiodic energy signals and produces a continuous spectrum. The deep link: as T0 → ∞ in the FS, harmonics get closer together and the discrete spectrum becomes the continuous FT. For a 50 Hz square wave, FS gives cn = (2A/nπ)sin(nπ/2) at n = 1, 3, 5 … — all odd harmonics. A single square pulse's FT is A·τ·sinc(fτ), a continuous envelope.
When to use Fourier Series
Use Fourier Series when the signal is periodic — for example, analysing the harmonic content of the output of a PWM inverter to check compliance with IEEE 519 harmonic standards.
When to use Fourier Transform
Use the Fourier Transform when the signal is aperiodic — for example, finding the bandwidth of a single Gaussian pulse used in ultra-wideband (UWB) radar where the pulse duration is 1 ns.
Recommendation
For university exams, choose Fourier Series for any periodic waveform and Fourier Transform for single pulses or decaying signals. If the problem says "find the spectrum" without specifying, check periodicity first. For GATE, the relationship cn = X(nf0)/T0 is a guaranteed 2-mark question.
Exam tip: Examiners ask you to find FS coefficients of a full-wave rectified sine and then sketch the spectrum — show discrete lines at 0, 2f0, 4f0 (no odd harmonics) and label cn magnitudes.
Interview tip: Interviewers ask why we can't use the regular Fourier Series for a non-periodic signal — explain that a non-periodic signal has no f0, so harmonics have no meaning; you need the continuous FT integral instead.