Skip to main content

CTFT convolution; DTFT of 2ⁿu[n]

Short answer

Convolution in time becomes multiplication in frequency: x(t) * h(t) ↔ X(jω)H(jω). For x[n] = 2^n u[n], the DTFT does not exist because the series diverges; if intended as (1/2)^n u[n], X(e^(jω)) = 1/(1 - 0.5e^(-jω)).

Show that convolution in time domain results multiplication in frequency domain using continuous time Fourier transform. Determine discrete time Fourier transform of the discrete time signal x[n] = 2ⁿ u[n] and also plot magnitude and phase spectrum.

Convolution property of CTFT

Statement: If x(t)↔X(jω)x(t) \leftrightarrow X(j\omega) and h(t)↔H(jω)h(t) \leftrightarrow H(j\omega), then

y(t)=x(t)∗h(t)  ⟷  Y(jω)=X(jω) H(jω)y(t) = x(t) * h(t) \;\longleftrightarrow\; Y(j\omega) = X(j\omega)\,H(j\omega)

Proof: The convolution is y(t)=∫−∞∞x(τ)h(t−τ)dτy(t) = \int_{-\infty}^{\infty} x(\tau)h(t-\tau)d\tau. Its FT is

Y(jω)=∫−∞∞[∫−∞∞x(τ)h(t−τ)dτ]e−jωtdt=∫−∞∞x(τ)[∫−∞∞h(t−τ)e−jωtdt]dτ\begin{aligned} Y(j\omega) &= \int_{-\infty}^{\infty}\left[\int_{-\infty}^{\infty} x(\tau)h(t-\tau)d\tau\right]e^{-j\omega t}dt \\ &= \int_{-\infty}^{\infty} x(\tau)\left[\int_{-\infty}^{\infty} h(t-\tau)e^{-j\omega t}dt\right]d\tau \end{aligned}

The inner integral is the FT of h(t)h(t) shifted by τ\tau, which by the time-shifting property is e−jωτH(jω)e^{-j\omega\tau}H(j\omega):

Y(jω)=∫−∞∞x(τ) e−jωτH(jω) dτ=H(jω)∫−∞∞x(τ)e−jωτdτ=X(jω)H(jω)\begin{aligned} Y(j\omega) &= \int_{-\infty}^{\infty} x(\tau)\,e^{-j\omega\tau}H(j\omega)\,d\tau \\ &= H(j\omega)\int_{-\infty}^{\infty} x(\tau)e^{-j\omega\tau}d\tau = X(j\omega)H(j\omega) \end{aligned}

Hence convolution in time becomes multiplication in frequency. This is why the output of an LTI system is found simply as Y=XHY = XH, where H(jω)H(j\omega) is the frequency response.

DTFT of x[n]=2nu[n]x[n] = 2^n u[n]

X(ejω)=∑n=0∞2ne−jωn=∑n=0∞(2e−jω)nX(e^{j\omega}) = \sum_{n=0}^{\infty} 2^n e^{-j\omega n} = \sum_{n=0}^{\infty}\left(2e^{-j\omega}\right)^n

This geometric series converges only if ∣2e−jω∣<1|2e^{-j\omega}| < 1, i.e. 2<12 < 1, which is false. Also ∑∣x[n]∣=∑2n=∞\sum|x[n]| = \sum 2^n = \infty, so the sequence is not absolutely summable. The DTFT of 2nu[n]2^n u[n] does not exist (its z-transform 11−2z−1\frac{1}{1-2z^{-1}}, ROC ∣z∣>2|z|>2, does not include the unit circle).

The question is usually intended as x[n]=(12)nu[n]x[n] = \left(\frac{1}{2}\right)^n u[n], which is solved below.

X(ejω)=∑n=0∞(12e−jω)n=11−0.5 e−jωX(e^{j\omega}) = \sum_{n=0}^{\infty}\left(\tfrac12 e^{-j\omega}\right)^n = \frac{1}{1 - 0.5\,e^{-j\omega}} ∣X(ejω)∣=11.25−cos⁡ω,∠X(ejω)=−tan⁡−10.5sin⁡ω1−0.5cos⁡ω|X(e^{j\omega})| = \frac{1}{\sqrt{1.25 - \cos\omega}}, \qquad \angle X(e^{j\omega}) = -\tan^{-1}\frac{0.5\sin\omega}{1 - 0.5\cos\omega}
ω\omega∣X∣\lvert X\rvertPhase
0020∘0^\circ
π/3\pi/31.155−30∘-30^\circ (minimum)
π/2\pi/20.894−26.57∘-26.57^\circ
π\pi0.6670∘0^\circ
 |X|
     2 |                        *
       |                     *     *
       |
       |                  *           *
       |               *                 *
       |         *  *                       *  *
  0.67 |*  *  *                                   *  *  *
       +-------------------------------------------------> w
       -pi                      0                      pi

 phase (degrees)
   +30 |            *  *  *
       |         *           *
       |      *
       |   *
     0 |*                       *                       *
       |                                             *
       |                                          *
       |                           *           *
   -30 |                              *  *  *
       +-------------------------------------------------> w
       -pi                      0                      pi

The magnitude is even (low-pass, peak 2 at ω=0\omega = 0), the phase is odd (extremes ±30∘\pm30^\circ at ω=∓π/3\omega = \mp\pi/3), and both repeat every 2π2\pi.

✓ Answer checked by Claude Opus 5.5 (high effort)

All Signals and Systems chapter 3 questions →