Skip to content
Tutorials Center
My learning

Topics

Lesson 17 of 36 ยท Fourier Series

Properties of Fourier Series

Learn the key properties of Fourier series coefficients, linearity, time shift, reversal, differentiation, multiplication and symmetry, and use them to skip integrals.

Computing a Fourier series from the defining integral is tedious. Properties let you get new series from known ones with one line of algebra. Throughout, x(t) has period T0, w0 = 2 pi / T0 and Fourier coefficients c_k (we write x <-> c_k).

Linearity

If x <-> c_k and y <-> d_k (same period), then a x + b y <-> a c_k + b d_k. The spectrum of a sum is the sum of spectra.

Time shift

Delaying the signal, x(t - t0), multiplies each coefficient by a phase factor:

x(t - t0) <-> c_k e^(-j k w0 t0).

The magnitudes |c_k| do not change; a shift only changes phases, by an amount proportional to frequency (a linear phase). The power at each harmonic is unaffected by where in time the signal starts.

Python A shift changes only the phase

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

Program input ยท 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.

Time reversal

x(-t) <-> c_{-k}. For real signals c_{-k} = c_k*, so reversal conjugates the coefficients (flips the sign of the phases).

Time scaling

x(a t) has period T0 / a, but the same coefficients c_k; only the fundamental frequency w0 becomes a w0. Speeding a signal up moves every harmonic up in frequency without altering how strong it is.

Differentiation and integration

If x(t) = sum of c_k e^(j k w0 t), then differentiating term by term:

dx/dt <-> j k w0 c_k.

Differentiation multiplies the coefficient of the k-th harmonic by j k w0: high harmonics are boosted by their frequency. This is why derivatives amplify noise and why a signal with a jump (whose derivative contains an impulse) has slowly decaying coefficients. Integration divides by j k w0 (valid for k != 0 when the DC term is zero), which smooths the signal and speeds up the decay.

The triangle wave is the integral of the square wave, so its coefficients should be those of the square wave divided by j k w0, giving 1/k^2. Check this with the relationship:

Python Differentiating a triangle gives a square

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

Program input ยท 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.

Multiplication and convolution

  • Multiplying two periodic signals in time convolves their coefficients: x y <-> sum of c_m d_{k - m}.
  • Periodic convolution in time (over one period) multiplies coefficients: (1/T0) integral of x(tau) y(t - tau) d tau <-> c_k d_k.

The second fact is the heart of frequency-domain filtering. Combined with the idea that LTI systems scale each exponential e^(j k w0 t) by a complex number, it gives the output of a system fed a periodic input directly.

Response of an LTI system to a periodic input

An LTI system with impulse response h(t) turns e^(j w t) into H(w) e^(j w t), where H(w) = integral of h(t) e^(-j w t) dt is the frequency response. So the output coefficients are

y <-> c_k H(k w0).

For an RC low-pass with H(w) = 1 / (1 + j w RC), each harmonic of a square wave is attenuated and delayed by a different amount, rounding the corners.

Python Square wave through an RC low-pass

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

Program input ยท 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 input (truncated at 49 harmonics) rings up to 1.179 because of the Gibbs effect; the output peaks near 1.01, with rounded corners. That is a direct picture of low-pass filtering: higher harmonics are cut more.

Symmetry summary

Property of x(t) Effect on c_k
real c_{-k} = c_k*
even c_k even
odd c_k odd
real and even c_k real and even
real and odd c_k purely imaginary and odd
half-wave symmetry x(t + T0/2) = -x(t) only odd harmonics

Check your understanding

Quick check ยท 5 questions

of 5 answered You scored / 5 ยท perfect!

  1. 1. Shifting x(t) to x(t - t0) changes its Fourier coefficients' ...

    c_k is multiplied by e^(-j k w0 t0), a unit-magnitude factor that only changes phase.
  2. 2. The coefficients of dx/dt are:

    Differentiating e^(j k w0 t) gives j k w0 e^(j k w0 t).
  3. 3. A periodic input x with coefficients c_k enters an LTI system with frequency response H(w). The output coefficients are:

    Each exponential is scaled by the frequency response at its own frequency.
  4. 4. A real, odd signal has Fourier coefficients that are:

    Odd real signals contain only sines, giving c_k = -j b_k / 2, imaginary and odd in k.
  5. 5. Why does differentiating a signal speed up the decay problem for noisy data?

    The factor j k w0 grows with k, so high-frequency noise is amplified.

Summary

  • Linearity; time shift multiplies c_k by e^(-j k w0 t0) (phase only); reversal gives c_{-k}.
  • Differentiation multiplies by j k w0; integration divides by it.
  • Product in time corresponds to convolution of coefficients; periodic convolution corresponds to a product.
  • LTI output coefficients: c_k H(k w0).
  • Symmetries of x(t) restrict the shape of c_k.

Next: Parseval's Theorem and the Power Spectrum