Scribd
Upload
50 views5 pages

Exercises 2 With Solutions

This document contains solutions to exercises on signals and systems from a course. It includes: 1) The analysis of sampling a signal with two frequency components, showing the sampled signal's spectrum is zero due to cancellation. 2) Examining the spectrum of a signal sampled and filtered, determining the 1 kHz frequency component is what can be heard. 3) Analyzing sampling rates and reconstructing signals from sampled data for various signals and sampling scenarios.

Uploaded by

quanmh16082003
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
50 views5 pages

Exercises 2 With Solutions

This document contains solutions to exercises on signals and systems from a course. It includes: 1) The analysis of sampling a signal with two frequency components, showing the sampled signal's spectrum is zero due to cancellation. 2) Examining the spectrum of a signal sampled and filtered, determining the 1 kHz frequency component is what can be heard. 3) Analyzing sampling rates and reconstructing signals from sampled data for various signals and sampling scenarios.

Uploaded by

quanmh16082003
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 5

WS 2023-24 Prof. Dr. B.

Lange SS II

Solutions for 2nd Exercises in signals and systems II

Exercise 5: The time-continuous signal

x(t) = cos(2πt) − cos(10πt)

is sampled with a frequency fs = 4 Hz. Sketch the spectra of both the time-
continuous and the sampled signal. Discuss your findings!
Solution: The spectrum of the time-continuous signal is
X(f )
1

0
f
−1
−6 −5 −4 −3 −2 −1 0 1 2 3 4 5 6

The spectrum of the sampled signal (xn )n∈Z is zero for all frequencies because
the original and frequency-shifted components completely cancel:
Xs (f )
2

0
f
−1

−2
−6 −5 −4 −3 −2 −1 0 1 2 3 4 5 6

This example illustrates that reconstruction of the original signal from the
samples is impossible if the sampling frequency is too low.

Exercise 6: A sine generator generates a sinusoidal signal at f = 9 kHz.


It is sampled with a sampling frequency fs = 8 kHz. The sampled signal is
then filtered by an (ideal) low-pass filter with cut-off frequency fc = 4 kHz
and finally output on a loudspeaker. Explain what frequency can be heard.
Solution: The spectrum of the time-continuous signal has the form
π
X(jω) = · (δ(ω − ω0 ) − δ(ω + ω0 )),
j

version: October 9, 2023 page 1 of 5


WS 2023-24 Prof. Dr. B. Lange SS II

or, expressed with frequencies rather than angular frequencies,


π
X(jf ) = · δ(2πf − 2πf0 ) − δ(2πf + 2πf0 )
j
π 1
= · δ(f − f0 ) − δ(f + f0 ) = · δ(f − f0 ) − δ(f + f0 ) .
2πj 2j

The spectrum of the time-continuous signal can be sketched as follows (frequency


axis scaled in kHz):

ℑ X(f )
1

0
f
−1
−15 −10 −5 0 5 10 15

The spectrum of the sampled signal is


∞
X
Xs (jf ) = fs X(j · (f − kfs ))
k=−∞
∞
X 1
= fs · δ(f − kfs − f0 ) − δ(f − kfs + f0 ) .
k=−∞
2j

It can be sketched as follows (frequency axis again scaled in kHz):



ℑ X(f )
4
3
2
1
0
−1 f
−2
−3
−4
−15 −10 −5 0 5 10 15

version: October 9, 2023 page 2 of 5


WS 2023-24 Prof. Dr. B. Lange SS II

An (ideal) low-pass filter with cut-off frequency fc = 4 kHz reduces it to


(frequency axis again scaled in kHz):

ℑ X(f )
4
3
2
1
0
−1 f
−2
−3
−4
−15 −10 −5 0 5 10 15

So only the frequency f = 1 kHz can be heard.

Exercise 7: Consider the following (real) spectrum X(jω).


X(jω)
2
1
−0
ω
−18 −15 −12 −9 −6 −3 0 3 6 9 12 15 18

a) The corresponding time signal x(t) is sampled with an angular sampling


frequency of ωs = 7 rad/s. Sketch the spectrum of the sampled signal (xk )k∈Z
and explain why there is no spectral overlap.
b) Examine whether the signal x(t) can be reconstructed from the sampled
signal (xk )k∈Z , and if yes, explain how.
Solution:
a) The spectrum of the signal sampled with ωs = 7 rad/s has the spectrum
∞
ωs X
Xs (jω) = X j(ω − kωs ) .
2π k=−∞

It can be sketched as follows:


X(jω)
2
1
−0
ω
−18 −15 −12 −9 −6 −3 0 3 6 9 12 15 18

There is no overlap in the spectrum since the bandwidth of the signal

17 rad/s − 15 rad/s = 2 rad/s

is less than half of the angular sampling frequency.

version: October 9, 2023 page 3 of 5


WS 2023-24 Prof. Dr. B. Lange SS II

b) Since there is no overlap, rescaling of the spectrum by a factor of 2π 7


and then applying a customized band-pass filter with two pass-bands, from
−17 rad/s to −15 rad/s and from 15 rad/s to 17 rad/s, yields the original
spectrum. Inverse Fourier transform then results in the original time signal.

Exercise 8: The two (real) spectra X(jω) and Y (jω) of the two signals
x(t) and y(t) are sketched as follows:
X(jω) Y (jω)
2 2
1 1
0 0
ω ω
−3 0 3 −3 0 3
a) Determine and sketch the spectrum of the product signal z(t) = x(t) · y(t).
b) Can the product signal z(t) be reconstructed if it is sampled with an
angular sampling frequency of ωs = 5 rad/s?
Solution: a) The two spectra can be expressed using step functions
X(jω) = σ(ω + 1) − σ(ω − 1) and Y (jω) = σ(ω + 2) − σ(ω − 2).
The spectrum of the product z(t) = x(t)·y(t) of the signals is the convolution
of their spectra
1 1
Z(jω) = · X(jω) ∗ Y (jω) = · Y (jω) ∗ X(jω)
2π Z 2π
∞
1
= X j · (ω − Ω) · Y (jΩ) dΩ
2π −∞
Z ∞
1
= σ(ω + 1 − Ω) − σ(ω − 1 − Ω) · σ(Ω + 2) − σ(Ω − 2) dΩ
2π −∞
Z ∞
1
= σ(ω + 1 − Ω) · σ(Ω + 2) − σ(ω − 1 − Ω) · σ(Ω + 2)
2π −∞

− σ(ω + 1 − Ω) · σ(Ω − 2) + σ(ω − 1 − Ω) · σ(Ω − 2) dΩ
(∗) 1
= · σ(ω + 3) · (ω + 3) − σ(ω + 1) · (ω + 1) − σ(ω − 1) · (ω − 1) + σ(ω − 3) · (ω − 3)
2π
1
= · ρ(ω + 3) − ρ(ω + 1) − ρ(ω − 1) + ρ(ω − 3) .
2π
In the step marked with (∗),
Z ∞ Z b
′ ′ ′
σ(t − a) · σ(b − t ) dt = σ(b − a) · dt′
−∞ a
= σ(b − a) · (b − a)
= ρ(b − a)
was used as shown in lecture 4 of SS1. The sketch of this spectrum is:

version: October 9, 2023 page 4 of 5


WS 2023-24 Prof. Dr. B. Lange SS II

Z(jω)
1

0
ω
−4 −2 0 2 4

b) Since the spectrum Z(jω) of the product function extends to ωc = 3 rad/s,


sampling with an angular sampling frequency of ωs = 5 rad/s < 2ωc violates
the sampling theorem. Thus, the original signal can not be reconstructed.

version: October 9, 2023 page 5 of 5

You might also like