Scribd
Upload
7 views6 pages

06.09.2021 en

The document outlines the exam details for Signal Processing at Graz University of Technology, including permitted aids and point distribution for multiple-choice and short problems. It contains various theoretical questions related to signal processing concepts such as linear systems, Fourier transforms, and aliasing effects. Additionally, it includes practical problems involving LTI systems, pole-zero diagrams, and multirate systems.

Uploaded by

bookmytrain93
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
7 views6 pages

06.09.2021 en

The document outlines the exam details for Signal Processing at Graz University of Technology, including permitted aids and point distribution for multiple-choice and short problems. It contains various theoretical questions related to signal processing concepts such as linear systems, Fourier transforms, and aliasing effects. Additionally, it includes practical problems involving LTI systems, pole-zero diagrams, and multirate systems.

Uploaded by

bookmytrain93
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/ 6

Signal Processing and Speech Communication Laboratory — Prof. G.

Kubin
Graz University of Technology

Exam for LE 442.001 Signal Processing, 06.09.2021

Name:

Matriculation No.:

Duration: 150 minutes, Achievable points: 100


Permitted aids: provided SPSC formulary, scientific calculator (not alphanumeric)
The problem statement and the formulary must be returned at the end of the
exam!

Theory – Multiple Choice Questions (20 Points)

Please tick the correct answers. There is exactly one correct answer to each question. For each
correct answer you will receive 4 points, for each wrong answer 2 points will be deducted. For
unanswered questions or questions with multiple answers ticked, we will neither give nor deduct
points. In total, you cannot receive less than 0 points for the multiple-choice part.

(a) Which of the following systems is linear and time-invariant?

( )|n|
□ y[n] = 1
4
x[n].

□ y[n − 1] = x[n − 1] + tan(4) x[n] − cos(0.4πn) y[n].


∑
n+1
□ y[n] = cos2 (π/3) x[m].
m=−∞

□ y[n] = x[12n].

1
ℑ{z}

b a 1 ℜ{z}

Abbildung 1: Pole/zero plot of H(z).

(b) How are the Fast-Fourier-Transformation (FFT) and the Discrete-Time Fourier Transform
(DTFT) related?

□ The DTFT equals the FFT for discrete-time signals.

□ The FFT is an efficient algorithm for the computation of the DTFT.

□ The DTFT is the FFT evaluated at the unit circle.

□ The FFT is an efficient algorithm for the computation of the sampled DTFT.

(c) Let H(z) be represented by the pole/zero plot in Fig. 1. For what region of convergence
(ROC) does the Fourier transform exist?

□ |z| > b

□ a < |z| < b

□ |z| < a

□ In this case a Fourier transform does not exist because there is a zero on the unit circle.

2
(d) Which of the following DTFT functions corresponds to a real-valued discrete time se-
quence?

□ X(ejθ ) = δ2π (θ − π)

□ X(ejθ ) = πj δ2π (θ + π4 )

□ X(ejθ ) = sin (θ)



π if π4 < θ mod 2π < 3π
□ X(ejθ ) = 4
0 otherwise

(e) Under what circumstances will the aliasing effect certainly occur?

□ When downsampling, if not a lowpass filter is used.

□ When sampling a continuous-time signal whose highest frequency component exceeds the
sampling frequency by a rational factor of 11
10
.

□ When upsampling by a rational factor, e.g., 102


100
.

□ When sampling a continuous-time signal, if not a lowpass filter is used.

3
Short Problems (20 Points)

(a) The basis B consists of N base signals bk [n] ∈ C of the form


2πkn
bk [n] = ej N ,

where k, n ∈ {0, 1, . . . , N − 1}. Show that B is an orthogonal basis and determine the scaling
factor α for a general N such that the signals b̃k [n] = α · bk [n] span an orthonormal basis. Hint:
∑N −1 n q N −1
n=0 q = q−1 .

(b) A causal LTI system satisfies the difference equation


1 1 1 1
y[n] − y[n − 1] = x[n] + x[n − 1] − x[n − 2].
4 2 4 4
Determine the transfer function H(z) and sketch the pole/zero plot. Is the system stable?
Determine the impulse response h[n] of the system. Hint: Use a partial fraction decomposition
to split up H(z).

4
Problem 1 (30 Points)

Consider a causal, discrete-time LTI-system with system function H1 (z). Its poles are given as
z∞,1 = −0.8 and z∞,2 = 0.5, its zeros are given as z0,1 = 1.25 and z0,2 = −0.5j

(a) Sketch the pole-zero diagram and determine the region of convergence (ROC). Is the
system BIBO stable? Explain!

(b) A system is minimum-phase if it has a causal and stable inverse. Is H1 (z) minimum-phase?
Explain!

(c) Assume that the maximum of the magnitude response satisfies maxθ |H1 (ejθ )| = 1. Deter-
mine the system function H1 (z).

(d) For H1 (z), design a causal and stable equalizer with system function H2 (z) such that the
magnitude response of the cascade H(z) = H1 (z)H2 (z) identical to the constant 1, i.e., H(z)
is an allpass with |H(ejθ )| = 1.

(e) In the figure below the equalizer H2 (z) comes after the system H1 (z). Could you also use
H2 (z) as a predistorter in front of H1 (z) without changing H(z)? Explain.

x[n] H1 (z) H2 (z) y[n]

5
Problem 2 (30 Points)

Consider the following multirate system:

x[n] v[m] w[n]


xc (t) C/D ↓3 ↑3 H(ejθ ) y[n]

fs

The frequency response of the ideal lowpass filter H(ejθ ) is given by:

H(ejθ )
3

−π −π/3 π/3 π θ

The continuous-time signal xc (t) = cos(2πf0 t) with frequency f0 = 200 Hz is sampled at a rate
fs = 0.8 kHz in order to obtain the discrete-time signal x[n].

(a) Determine the discrete-time signal x[n]. Can you find an alternative choice for f0 yielding
the same discrete-time signal x[n]?

(b) Sketch the Fourier transforms of the signals x[n], v[m] und w[n] for the input signal xc (t).
Sketch at least two perods of the spectra.

(c) Is it true that y[n] = x[n] holds? If not, sketch the frequency response of an ideal filter
H(ejθ ) such that y[n] = x[n] holds.

(d) Repeat tasks (b) and (c) for a sampling frequency of fs = 1.6 kHz. How does the new
sampling frequency affect the result?

(e) Which requirement has f0 to fulfill for a given sampling frequency such that y[n] = x[n]
holds?

You might also like