Skip to content
Tutorials Center
My learning

Topics

Lesson 21 of 36 · Fourier Transform

Transform Pairs, Impulses and Parseval's Theorem

Build a table of Fourier transform pairs: impulse, constant, cosine, Gaussian, sinc. Learn energy spectral density and Parseval's theorem with Python checks.

With the properties from the last lesson, a short table of transform pairs unlocks nearly every problem you will meet. This lesson builds that table, explains why impulses appear in spectra of signals that never die out, and introduces Parseval's theorem, which says that energy is the same whether you add it up in time or in frequency.

The essential table

x(t) X(f)
δ(t) 1
1 δ(f)
δ(t - t0) e^(-j2πf t0)
e^(j2πf0 t) δ(f - f0)
cos(2πf0 t) ½ δ(f - f0) + ½ δ(f + f0)
sin(2πf0 t) (1/2j) δ(f - f0) - (1/2j) δ(f + f0)
rect(t/T) T sinc(fT)
sinc(2Wt) (1/2W) rect(f/2W)
e^(-at) u(t), a > 0 1 / (a + j2πf)
e^(-a|t|), a > 0 2a / (a² + (2πf)²)
e^(-πt²) e^(-πf²)
u(t) ½ δ(f) + 1 / (j2πf)

Read the first two rows together: an infinitely short event contains every frequency equally, while a signal that never changes contains only the frequency zero. This is the extreme version of the narrow-in-time, wide-in-frequency rule.

The pair for the Gaussian is special: e^(-πt²) is its own Fourier transform, and it is the only common shape that achieves the minimum time-bandwidth product. Row 8 shows duality: because a pulse becomes a sinc, a sinc becomes a pulse. An ideal low-pass filter has a rectangular frequency response, and its impulse response is therefore a sinc.

Python The Gaussian is its own transform

Press Run (or Ctrl + Enter) to see the result here.

Program input · what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

Impulses in the spectrum

A cosine never decays, so its integral against e^(-j2πft) is large only at f = ±f0 and zero elsewhere. A true impulse cannot be plotted on a computer, but the DFT of a sampled cosine shows the idea: if the cosine fits a whole number of periods in the window, all energy falls into two bins.

Python A cosine shows up as two spikes

Press Run (or Ctrl + Enter) to see the result here.

Program input · what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

The two spikes sit at +50 Hz and -50 Hz (bin 950 is the same as -50 Hz) and each has height ½, matching the table row for the cosine.

Energy and Parseval's theorem

The energy of a signal is E = ∫ |x(t)|² dt. Parseval's theorem states

∫ |x(t)|² dt = ∫ |X(f)|² df.

So |X(f)|² is an energy spectral density: it tells you how the energy is distributed over frequency. This allows you to ask practical questions such as "what fraction of the signal's energy lies below 100 Hz?" by integrating |X(f)|² over that band.

For x(t) = e^(-at) u(t) the time-domain energy is ∫ e^(-2at) dt = 1/(2a). In the frequency domain, |X(f)|² = 1 / (a² + (2πf)²), and its integral over all f is again 1/(2a). The next example checks both numerically for a = 2, where the answer is 0.25.

Python Parseval's theorem numerically

Press Run (or Ctrl + Enter) to see the result here.

Program input · what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

Periodic signals and power

A periodic signal has infinite energy but finite average power. Parseval's version for periodic signals says the average power equals the sum of |c_k|² over the Fourier-series coefficients. Impulses in the spectrum of a periodic signal correspond to those coefficients, so the idea carries over smoothly.

Do not apply the energy formula to a sine wave or a constant. Their energy is infinite; use power instead.

Check your understanding

Quick check · 4 questions

of 4 answered You scored / 4 · perfect!

  1. 1. What is the Fourier transform of the unit impulse δ(t)?

    An impulse contains every frequency with equal strength, so its transform is the constant 1.
  2. 2. The transform of cos(2πf0 t) consists of...

    cos is the sum of two complex exponentials with weight 1/2, one at +f0 and one at -f0.
  3. 3. Which statement of Parseval's theorem is correct?

    The integral of |x(t)|² over time equals the integral of |X(f)|² over frequency.
  4. 4. The ideal low-pass filter has a rectangular frequency response. Its impulse response is...

    By duality, a rectangle in frequency corresponds to a sinc in time. This is also why the ideal filter cannot be built: the sinc is non-causal.

Summary

  • Impulses and constants form the extreme ends of the time-frequency trade-off.
  • A cosine has two impulses of height ½ at ±f0.
  • The Gaussian e^(-πt²) is its own transform; pulse and sinc are duals.
  • Parseval's theorem lets you compute energy in either domain, and |X(f)|² is the energy spectral density.

Next: The DTFT, DFT and FFT