Skip to main content

Fourier transform of periodic sequences

Short answer

A periodic sequence is not absolutely summable, so its DTFT consists of impulses. Compute the DTFS coefficients a_k, then X(e^{jw}) = sum 2pi a_k delta(w - 2pi k/N).

Derive the expression of Discrete Time Fourier Transform for periodic sequence (how do you calculate the Fourier transform of a periodic sequence?).

A periodic sequence is not absolutely summable, so its DTFT does not exist in the ordinary sense. It is found by allowing impulses in the frequency domain: each Fourier series term becomes a train of impulses.

Step 1: DTFT of a single complex exponential

Consider the spectrum consisting of impulses of area 2π2\pi at ω0\omega_0, repeated every 2π2\pi:

X(ejω)=∑l=−∞∞2π δ(ω−ω0−2πl)X(e^{j\omega}) = \sum_{l=-\infty}^{\infty}2\pi\,\delta(\omega - \omega_0 - 2\pi l)

Its inverse DTFT, taken over one period of 2π2\pi that contains only the impulse at ω0\omega_0, is

x[n]=12π∫2π2π δ(ω−ω0)ejωndω=ejω0nx[n] = \frac{1}{2\pi}\int_{2\pi}2\pi\,\delta(\omega - \omega_0)e^{j\omega n}d\omega = e^{j\omega_0 n}

Hence

ejω0n↔DTFT∑l=−∞∞2π δ(ω−ω0−2πl)e^{j\omega_0 n} \xleftrightarrow{DTFT} \sum_{l=-\infty}^{\infty}2\pi\,\delta(\omega - \omega_0 - 2\pi l)

Step 2: Periodic sequence as a sum of exponentials

A periodic x[n]x[n] with period NN has the DTFS

x[n]=∑k=⟨N⟩akejk(2π/N)n,ak=1N∑n=⟨N⟩x[n]e−jk(2π/N)nx[n] = \sum_{k=\langle N\rangle}a_k e^{jk(2\pi/N)n}, \qquad a_k = \frac{1}{N}\sum_{n=\langle N\rangle}x[n]e^{-jk(2\pi/N)n}

Step 3: Apply linearity

Each term akejk(2π/N)na_k e^{jk(2\pi/N)n} gives impulses of area 2πak2\pi a_k at ω=2πk/N+2πl\omega = 2\pi k/N + 2\pi l. Since aka_k is periodic in kk with period NN, the impulses for all kk and ll can be combined into one sum over all integers kk:

X(ejω)=∑k=−∞∞2π ak δ ⁣(ω−2πkN)\boxed{X(e^{j\omega}) = \sum_{k=-\infty}^{\infty}2\pi\,a_k\,\delta\!\left(\omega - \frac{2\pi k}{N}\right)}

Procedure: find the DTFS coefficients aka_k, then place an impulse of area 2πak2\pi a_k at each frequency 2πk/N2\pi k/N.

 X(e^jw)
     2pi a0   2pi a1        2pi a0 (repeat)
       ^        ^      ^      ^
       |   ^    |      |      |
 ------+---+----+------+------+---- w
       0  2pi/N 4pi/N  ...   2pi

Example: periodic impulse train

x[n]=∑kδ[n−kN]x[n] = \sum_k\delta[n - kN] has ak=1/Na_k = 1/N for all kk, so

X(ejω)=2πN∑k=−∞∞δ ⁣(ω−2πkN)X(e^{j\omega}) = \frac{2\pi}{N}\sum_{k=-\infty}^{\infty}\delta\!\left(\omega - \frac{2\pi k}{N}\right)

An impulse train in time gives an impulse train in frequency.

Example 2: cos⁡ω0n=12(ejω0n+e−jω0n)\cos\omega_0 n = \frac{1}{2}(e^{j\omega_0 n} + e^{-j\omega_0 n}) gives X(ejω)=∑lπ[δ(ω−ω0−2πl)+δ(ω+ω0−2πl)]X(e^{j\omega}) = \sum_l \pi[\delta(\omega - \omega_0 - 2\pi l) + \delta(\omega + \omega_0 - 2\pi l)].

✓ Answer checked by Claude Opus 5.5 (high effort)

All Signals and Systems chapter 3 questions →