Ferhat Abbas University – Sétif1 1 ﺟﺎﻣﻌﺔ ﻓﺮﺣﺎت ﻋﺒﺎس – ﺳﻄﻴﻒ
Faculty of Technology ﻠﻴﺔ اﻟﺘﻜﻨﻮﻟﻮﺟﻴﺎɠ
Department of Eletronics فوﻧﻴﻚﻗﺴﻢ اﻻﻟﻜ
Division of Telecommunications ﺷﻌﺒﺔ اﻻﺗﺼﺎﻻت اﻟﺴﻠﻜﻴﺔ واﻟﻼﺳﻠﻜﻴﺔ
Level: M1 U.Y : 2024/2025
Specialty: Networks and Telecommunications Matter : Advanced Signal Processing
Tutorial 3
Exercise 1
Consider
y[n] − cos(ω0 )y[n − 1] + (1/4)y[n − 2] = x[n − 2] − (1/2) cos(ω0 )x[n − 3] ,
which represents a causal LTI system.
a) Compute the z-transform of the system H(z).
b) Sketch the pole-zero plot for the system for ω0 = 0, ω0 = π/3, and ω0 = π/2.
c) For ω0 = π/3 or ω0 = 0, compute and sketch the magnitude response |H(ω)| for −2π ≤ ω ≤ 2π.
d) For ω0 = π/3 or ω0 = 0, is this filter a low pass, band-pass, high pass, band stop, or all pass filter?
Exercise 2
1 − (2)e−jω
Consider H(ω) = 5 , which represents a causal LTI system.
1 − (1/2)e−jω
a) Determine the difference equation for this system.
b) Sketch the pole-zero plot for the system.
c) Sketch the magnitude response |H(ω)| for −2π ≤ ω ≤ 2π.
d) Compute and sketch the phase response ∠H(ω) for −2π ≤ ω ≤ 2π.
e) Compute the output from input x[n] = 2 + 4 cos( π2 n).
Exercise 3
1. For the following pole-zero plots, specify if they correspond to an FIR filter or an IIR filter, specify
if each filter is stable, specify if the filter has a linear phase, and specify if the filter’s inverse is
stable (i.e., the filter invertible).
2. Match the above pole-zero plots with the associated magnitude fequency responses.
3. Match the pole-zero plots with the associated impulse responses.
Exercise 4
Design an FIR linear-phase, digital filter approximating the ideal frequency response
π
1, for |ω| ≤
Hd (ω) = 6
0, π
for < |ω| ≤ π
6
(a) Determine the coefficients of a 25-tap filter based on the window method with a rectangular window.
(b) Determine and plot the magnitude and phase response of the filter.
Exercise 5
During the transmission of a digital signal (sampled at a frequency of 2.5KHz), it was affected by a
noise localised in the frequency band: 350Hz - 550Hz. We want to eliminate the noise by using an FIR
filter with a transition band def=100Hz. Design this filter:
a) Using the windowing method. We want an attenuation ( A 20 log10 20 dB ).
- Draw H f ideal.
- Determine h(n) and the filter order N.
- Calculate h(0), h(1)= h(-1).
- Plot approximately H f .
b) Using the frequency sampling technique.
- Determine N.
- Give the values of H(k) and its plot.
- Determine h(n).
- Compare these two design techniques.
Exercise 6
1. Show that the impulse response h(n) of an ideal high-pass filter with cutoff frequency fc= 1/4 can be expressed as :
Hd (f)
sin( /2)
− ∀ ≠0
ℎ( ) = 1
1
for =0 f
2
-1/2 -1/4 1/4 1/2
Can this filter be implemented as a RIF filter?
This ideal filter is approximated by a causal linear-phase RIF filter of length N as small as possible, using the
windowing method. This filter must comply with the frequency template shown in Figure : transition band of
1/10; low-band attenuation greater than 30 dB. How do I go about it?
- .
High-Pass Filter
2. The previous windowing is replaced by a symmetrical rectangular windowing on 11 points.
- Calculate the coefficients of the synthesized filter.
- Draw the structure of this filter.
- Draw the template in which its frequency response fits, specifying the width of the transition band
and the attenuation of the low band.
- Show that if two filters like this one are cascaded (one after the other), the resulting system respects the template
in question 2.
