CamScanner

CamScanner
 Hnternnational Islamic University Chittagong
Department of Electrical and Electnonic Engineering

Final Examination Spring-2019 Program: B.Sc. Engg. (EEE.)

course code: EEE 3603 course Title: Digital Signal Processing-I
Time: 2 hours 30 minutes Full Marks: 50
Part A
fAnswer any two questions from the followings; Figures in the right margin indicate full marks.]

1(a). From the expression of z-transform, proof that, the ROC of a causal signal is the 04
exterior of a circle of any radius r2 while the ROC of an anti-causal signal is the
interior of a circle of any radius r 1.

1(b). Determine the z-transform of the signal 03

x(n) = aa(cos usn)u(n)

1(c). Draw the Argand diagram and explain the characteristics of z'" at different region of 83
Argand diagram.

2(a). Define IIR system. Also, Draw the direct tbrm I and canonic form structure for the 05
IIR system whose transfer function is given as

Ht:\-1-;J:-l+2;-r
1+ 0.32-t - 0.72*)
2(b). Consider an FIR system rvith systern function 04
H{z) = 1* 2.BBz-L + 3.40482-2 + L.74z-3 + 0.42-n
Sketch the direct form realization of the system. How many additions and

multiplications instructions are required per output point? Also determine the
number of memory block.
2(c). Why we factorized the system function into second order polynomial instead of 01
first-order polynomial in cascade realization?

3(a). What are the applications of z-transform? Also, Determine the response of the 04

system y(n) : + x(n), to the input signal

fr(n-r) - |f(n-Z)
x(n):o(n1-]o1n-r1.
3(b). For a low-pass RC network (R: lMQ and c : 10--6F), determine the equivalent 0,8

discrete time expressions for the circuit output response y(n), when the input is

x(t): e-2t and the sampling frequency is F, : 50H2.

3(c). Explain linear-phase FIR system with an example. gz

Page 1 of2
 Part B
[Answer any three questions from the followings; Figures in the right margin indicate fuil marks.]

4(a). Write the mathematical expression to find the DFT of a discrete time sequence. 0z
Also, explain the physical meaning of each variable in the expression.
4(b). Distinguish between circular convolution and linear convolution with an example.
0S
Also, develop the expression of circular convolution from the multiplication of two
DFTs.
4(c). Compute the DFT of the 4-point sequence x(n):{0,1,2,3}
03

5(a). "An IIR filter is a recursive system,,-justifl, the statement. 1.5

5(b). What is meant by frequency warping? How could you use the step response of a 3.5
digital filter to measure the performance of that filter?
5(c). What is filter's kernel? Also, define digital fiiter. 05
How do you design all kinds of digital filter by the low pass filter kernel? Explain.

6(a)" Design a single-pole low-pass digital filter with a 3-dB bandwidth of 0.2n,using
the 3.s
bilinear transformation applied to the analog filter H(s) : , where e is the
*
3-dB bandwidrh of analog filter.
6(b). How window improve the performance of a digital filter? Aiso, Compare the role of 4.5
blackman and hamming window to improve the performance of a digital filter.
6(c). Explain the relationship between roll-off and kernel length with necessary figure. z

7(a). Obtain the coefficients of an FIR low-pass filter to meet the specifications given 05
bellow using the window method.

Determine only 5 coefficients.

7(b). For the discrete time sequence x(n):{ 1,1,1,0,0,0,0,0}, Evaluate the g-point DFT 05
using decimation in time FFT Algorithm.

Page? of}
 International Islamic University Chittagong
Department of Electrical and Electronic Engineering

Special Final Examination Autumn-2018 Program: B'Sc. Engg (EEE)

Course Code: EEE 3603 Course Title: Digital Signal Processing
Time: 2 hours 30 minutes Full Marks: 50

Part A
the right margin indicate full nrarks ]
[Answer any Og questions from the followings; figures in

1(a), Expla:''u^ physical significance ofz-transform. 82

l(b). Deter,..ll,u lhe z-transform ofunit step signal. 03

1(c). Show that the ztransfornr oftwo different time domain signals may be identical but 03

their ROC's are different.

1(d). Determine the pole-zero plot for the signal t2

x(n)=a"u1n1,o'O

2(a). e methods of implementing IIR systerrs' 'tz

2(ir). Col -,' unction

. ,1\ sys!.-,,r t'..,, -, -
H(z)=t*2.BBz-1 * 3.4a482-2 + 7'742-3 + 0'"'
cketc,. ,!.- y': , .^.-. .-..r:. additions rnd

nruli . '-!ui mine the nut]1b:'i'

ofnremory block.
2(c). Explain the cascade fonn realization of FIR s)'sterr' Why we factorized the system 04
function into second order polynomial instead of first-order poiynomial in cascact
realization?

3(a). Detc '"se z-transform of

1
=. ,-
X(z)
u.5z_2

lortheROCu. lzl ) 1 (ii) lzl < 0 5 and (iii) 0'5 < lzl < 1
1r.7

3(b). Deterrnine the cascade realizalion ofthe sysien] describcd by the systcm lunetion 05

ro(r - 1,-,)(r -1,-'.)1t + t,.')

tt\z) -
(1 -;' |rr I;i;{r - 1}'7};'-'i[' - t]-; )'-']

-
 Part B
right margin indicate full marks ]
[Answer any {4gg questions frorn the followings; figures in the

4(a). Develop a general expression of linear transfonnation to compute 4-point DFT of a 06

discretetimesequence.Usingthisexpression,computetheDFTofthefollowing
sequence

x(n)={0,1,2,3}

4(b). Write the mathematical expression of Discrete Fourier Transform (DFT) with the 04

explanation of each symbol. To find the Discrete Fourier Transform, compare the
computational complexity of direct DFT rvith FFT.

5(a). Briefly explain the functions ofa digital filter' 03

5(b). Discuss the disadvantages ofdigital filter over analog filter' 0Z

5(c).What''fafilter?Explainthewaysofconvertingthekerneloflowpass05
filter i.,,-,, o.. ,-. - ^. .or.

6(a). Why ideal filter is not physically realizable? 02

6(b). Explain the 1s!at;onshin hetween length of filter kernel and roll-oflrate t2

6(c). Desil lter with o:p=O 3rr and os=0 7n which exhibits a minimum sS

attenuation greater than -50d8 in the stop-band'

7(a). Illustrate the basic butterfly computational structure of DIT-FFT algorithm. From this 04

structufe deteriltille the nulnber of addilions and In u ltiplications required to calculate

N-point FFT.
7(b). Fortl :ime sequence x(n)={1,1,1,0,0,0,0,0}, Evaluate the S-point DirT using 06

decinrallott rrr urr,; FFT Algoritlrrn'

Pzge2 of 2
 International Islamic University Chittagong
Department of Electrical and Electronic Engineering

Special Final Examination Autumn-2018 Program: B'Sc. Engg (EEE)

Course Code: EEE 3603 Course Title: Digital Signal Processing
Time: 2 hours 30 minutes Full Marks: 50

Part A
the right margin indicate full nrarks ]
[Answer any Og questions from the followings; figures in

1(a), Expla:''u^ physical significance ofz-transform. 82

l(b). Deter,..ll,u lhe z-transform ofunit step signal. 03

1(c). Show that the ztransfornr oftwo different time domain signals may be identical but 03

their ROC's are different.

1(d). Determine the pole-zero plot for the signal t2

x(n)=a"u1n1,o'O

2(a). e methods of implementing IIR systerrs' 'tz

2(ir). Col -,' unction

. ,1\ sys!.-,,r t'..,, -, -
H(z)=t*2.BBz-1 * 3.4a482-2 + 7'742-3 + 0'"'
cketc,. ,!.- y': , .^.-. .-..r:. additions rnd

nruli . '-!ui mine the nut]1b:'i'

ofnremory block.
2(c). Explain the cascade fonn realization of FIR s)'sterr' Why we factorized the system 04
function into second order polynomial instead of first-order poiynomial in cascact
realization?

3(a). Detc '"se z-transform of

1
=. ,-
X(z)
u.5z_2

lortheROCu. lzl ) 1 (ii) lzl < 0 5 and (iii) 0'5 < lzl < 1
1r.7

3(b). Deterrnine the cascade realizalion ofthe sysien] describcd by the systcm lunetion 05

ro(r - 1,-,)(r -1,-'.)1t + t,.')

tt\z) -
(1 -;' |rr I;i;{r - 1}'7};'-'i[' - t]-; )'-']

-
 Part B
right margin indicate full marks ]
[Answer any {4gg questions frorn the followings; figures in the

4(a). Develop a general expression of linear transfonnation to compute 4-point DFT of a 06

discretetimesequence.Usingthisexpression,computetheDFTofthefollowing
sequence

x(n)={0,1,2,3}

4(b). Write the mathematical expression of Discrete Fourier Transform (DFT) with the 04

explanation of each symbol. To find the Discrete Fourier Transform, compare the
computational complexity of direct DFT rvith FFT.

5(a). Briefly explain the functions ofa digital filter' 03

5(b). Discuss the disadvantages ofdigital filter over analog filter' 0Z

5(c).What''fafilter?Explainthewaysofconvertingthekerneloflowpass05
filter i.,,-,, o.. ,-. - ^. .or.

6(a). Why ideal filter is not physically realizable? 02

6(b). Explain the 1s!at;onshin hetween length of filter kernel and roll-oflrate t2

6(c). Desil lter with o:p=O 3rr and os=0 7n which exhibits a minimum sS

attenuation greater than -50d8 in the stop-band'

7(a). Illustrate the basic butterfly computational structure of DIT-FFT algorithm. From this 04

structufe deteriltille the nulnber of addilions and In u ltiplications required to calculate

N-point FFT.
7(b). Fortl :ime sequence x(n)={1,1,1,0,0,0,0,0}, Evaluate the S-point DirT using 06

decinrallott rrr urr,; FFT Algoritlrrn'

Pzge2 of 2
 {natermational Islarnic Uxriversity Chittagong
Departmexrt of Electrical and Electronlc trngineering

Final Examination Spring-20l8 Program:B.Sc. Engg. (EEE)

Course Code: EEE-3603 Course Title: Digital Signai Processing I
Tirne: 2 hours 30 minutes Full Marks: 50
Figures in the right margin indicate full marks.

Part A
fAnsrver afiy ttro questions from the foilowings]

1(a). Define the following terms:

i. Trivial and non-trivial pole
ii. All pole system and all zero system
l(b). Using long division, determine the inverse z-transform of
: \\'hen RoC: zi>
X1z)
,*"+."r*. I

1-(c). Determine the z-transform and pole-zero plot of the signal

x(n): f])i'tnt
And also explain u'hether it is stable or unstable?

2{a). Discuss the major factors that influence the choice of a specific structure 3
realization.
2(b). What is the signihcance of the name "direct form structure realization"? 3
Dei,ermine the direct form Ii realization for a third.order IIR transfer
function.

Hlzt : a.2Bz2 -O.37gz+O.04

usz3+oszz+arr z-a,
2{c). Draw the cascade realization structures for the s-vstem described by the 4
system function
(1-z-t 1t
"\-/
n\z)- /
(r-i,-')(r-*,-,)
3(a). What is system function? Briefly discuss about the pole-zero cancellations. 2,5
3(b). Determine the response of the system y(n): fV(n-fl -iV@-Zl + x(n), to the 4
-
input signal x(n) : O1n1 -
lOln-r1.
3(c). Why a linear time-invariant system is BIBO stable if and only if the ROC of 3.5
the system function includes the unit circle? Explain.

Page 1 of2
 Part B
followings']
[Ansrver any \hLeg questions from the

4(a).WhatistheclifferencebelweenFouriertransformandDFT?1 5
4(b). Explain the symmetry and periodicity properlies of the phasor factor wx
using N : 4' Find the matrix W+ and using this matrix
compute the 4-point
DFT of the sequence
x(n) : cos-
rln

4(c). What is butterfly graph? Given that x(n): {0 1

2 3}' Find X(k) using 4
DIT FFT algorithm.

5(a). Define digital filter. Discuss the five main steps involved in the design of 4

digital filters.
An FIR BandPass filter is to be designed to meet the following frequenoy 2
5(b).
response sPecifi cations:
e Passband 0'18 - 0'33 (normalized)
* Transition width 0.04 (nonnalized)

' Stopbanddeviation 0'001

e Passbanddeviation 0'05
Sketch the tolerance scheme for the filter'
5(c). The following transfer functions replesent two different filters meeting 4

identical amplitude-frequency response specifications:

(i) :
Hr(z) (b0 + bl z'1 +bzz-Z) / (1 + a1Z-1+a2z-1) and
(ii) Hz(z) : 2,1o1=on(t<)z-k
For each f,tlter,
(p) State whether it is an FIR or iIR filter'
(q)Representthefilteringoperationinablockdiagramformandwritedown
the difference equation.
(r) Determine and comment on the computational and storage requirements'

5(a). Briefly discuss about filter implementation'

1'5

6(b). What is frequency warping? In the case of bilinear transfotmation, show

that 4'5
is ron-linear'
the lrequency relation between time and frequency domain
6(c). For lhe analog transform function 4
L
H(s):
G+ 1Xi+ A
DetermineH(z)usingimpulseinvarianttechnique.AssumeT:1s'
1(a). Determine the IDFT of X(k)
:{3,2+i,1,2-i)' 3
1
Zini. What are the common filtering objectives?
specifications' 6
zi.j. Design a low pass FIR with 11 coefficients for the following
t Passband frequency edge = 0'25 KHz

Use Hamming window.

Page2 of2
 International Islamic University Chittagong
Department of Electrical and Electronic Engineering
Final Assessment of Spring-2021 Program: B.Sc. Engg. (EEE)
Course Code: EEE-3603 Course Title: Digital Signal Processing I
Time: 5 hours (Writing - 4 hours 30 Full Marks: 50 (Written 30 + Viva-20)
minutes + 30 minutes submission time)

[Answer each of the questions from the followings; Figures in the right margin indicate full marks.
Answer script must be submitted through online method within 5 hours from starting time.]

Important Note: Compare your ID with the syntax ET-ALBCMN and find the value of A, B, C, L, M
and N. You have to use the values A, B, C, L, M and N in different questions to assume the parameters
related to the question. For example, if your matric ID is ET-143021, you have to assume A=1, B=3,
C=0, L=4, M=2 and N=1.
Course Outcomes (COs) of the Questions
To learn about mathematical representation of analog signals in digital domain, manipulate
CO1
signals using analytical techniques and familiarize with discrete time signal & systems.
To interpret the information of discrete time signals by means of frequency domain analysis
CO2 using mathematical tools such as Z-transform, Discrete Fourier Transform (DFT), Fast
Fourier Transform etc.
CO3 To design & realize the responses of discrete-time systems like FIR and IIR Filter etc.

Bloom’s Levels of the Questions

Letter Symbols R U App An E C
Meaning Remember Understand Apply Analyze Evaluate Create

1(a). Determine the pole zero plot of the signal x(n). Choose x(n) according to CO2 Ap 02
condition from the following table-
Table-1
N even N odd
𝑀+3 𝑛 𝑁+1 𝑛
𝑥(𝑛) = 𝑛 ( ) 𝑢(𝑛) 𝑥(𝑛) = 𝑛 ( ) 𝑢(𝑛)
𝑁+2 𝑀+1

Also, draw an approximate plot of x(n).

1(b). Determine the inverse z-transform of the following expression by power CO2 Ap 04
series expansion/ Partial fraction expansion method:
1
𝑋(𝑧) =
1 − 𝐴𝑧 −1 + 𝐵𝑧 −2
Now, find the value of A, B, method of solution and signal type from
Table-2.
Table-2
Condition Assumption Condition Assumption
M is even A=(L+M+10)/5 M is odd A=(L+N+10)/5
B=(M+N-10)/10 B=(M-N-10)/10

N is even Power Series Expansion N is odd Partial fraction

method expansion method
𝑴≥𝑵 Signal is causal 𝑴<𝑵 Signal is non-causal

Page 1 of 3
2(a). Assume a linear phase (if N is even) or, non-linear phase (if N is odd) FIR CO2 An 04
system having (L+M+N) number of samples (if M+N<7) or (L+M+N-7)
number of samples (if M+N>=7) in the impulse response such as
ℎ(𝑛) = {𝑏0 , 𝑏1 , 𝑏2 , … … … , 𝑏𝐿+𝑀+𝑁−1 }
Assume the values of 𝑏0 , 𝑏1 , 𝑏2 , … … … , 𝑎𝑛𝑑 𝑏𝐿+𝑀+𝑁−1randomly from the
set of numbers {𝐴, 𝐵, 𝐶, 𝐿, 𝑀, 𝑁}. You have to use every number from the
set at least once and as many times necessary to assume the samples of
h(n).
After assuming the mentioned FIR system, answer the following questions-
i) Sketch the direct form realization of the system.
ii) How many additions and multiplications are required are
required per output point? Also determine the number of
memory block.
2(b). Determine the direct form-I realization of the system described by the CO2 E 02
system function
(1 + 𝑎𝑧 −1 )(1 + 𝑎∗ 𝑧 −1 )
𝐻(𝑧) =
(1 − 𝑏 𝑧 −1 )(1 − 𝑐𝑧 −1 )
Where, ‘a’ is a complex number and assume, a=(M+N) +(L×j).
b and c are real constants. Assume, b=(L+M)/2 and c=(L+N)/2

3(a). Assume, a discrete time causal signal 𝑥(𝑛) = {𝐿, 𝑀, 𝑀 − 𝑁, 𝑁} CO1 Ap 04

Now, answer to the following questions-
i) Compute the DFT of the discrete-time signal x(n).
ii) Compute the FFT of the discrete-time signal x(n). [Hints: Draw the
butterfly diagram for the 4-point FFT and compute the FFT of x(n) using
this diagram]
iii) From the above two computations justify the statement, “DFT and FFT
do the same task, however, FFT is more computationally efficient than
DFT”.
3(b). Compute the linear convolution of the following two sequences by circular CO1 Ap 02
convolution-

4. Design a band-pass/band-stop FIR filter with the following specifications: CO3 C 06

Lower Cutoff frequency= x Hz
Lower transition width= 𝑥 ′ Hz
Upper Cutoff frequency= y Hz
Upper transition width= 𝑦 ′ Hz
Passband ripple= PR dB
Stopband attenuation = SA dB
Sampling rate= 8000 Hz
Find the value of 𝒙, 𝒙′ , 𝒚, 𝒚′ , 𝑷𝑹, 𝑺𝑨 and filter type from the Table-3
based on the value of L, M and N. Calculate only 11 samples of impulse
response of the filter symmetrically including center of symmetry.

Page 2 of 3
 Table-3
Condition Assumption Condition Assumption
M is even Band-stop filter M is odd Band-pass filter
N is even 𝑥 = 1000 + (𝑀 + 𝑁) ∗ 20 N is odd 𝑥 = 1100 + (𝑀 − 𝑁) ∗ 20
𝑥 ′ = 400 + (𝑀 + 𝑁) ∗ 30 𝑥 ′ = 500 + (𝑀 − 𝑁) ∗ 30
𝑦 = 2100 + (𝑀 + 𝑁) ∗ 25 𝑦 = 2300 + (𝑀 − 𝑁) ∗ 25
𝑦 ′ = 500 + (𝑀 + 𝑁) ∗ 15 𝑦 ′ = 600 + (𝑀 − 𝑁) ∗ 15

𝑴>𝑵 PR= 0.04 𝑴<𝑵 PR= 0.02

SA=40 𝒐𝒓, SA=60
𝑴=𝑵

Exception: If you find any parameter ( 𝑥, 𝑥 ′ , 𝑦, 𝑦 ′ ) is 0 in your case, you have to

assume a non-zero value for that parameter which is appropriate to solve the
problem.

5(a). Design a digital IIR low-pass/high-pass filter using Bilinear Transformation CO3 C 04
method of specifications:
Cutoff frequency= x Hz
Sampling rate= Fs Hz
Find the value of x, 𝑭𝒔 and filter type from the Table-4 based on the value
of L, M and N (find L, M and N from your matric ID).
Table-4
Condition Assumption Condition Assumption
M is even High-pass filter M is odd Low-pass filter
N is even 𝑥 = (𝑀 + 𝑁 + 𝐿) ∗ 3 𝐻𝑧 N is odd 𝑥 = (𝑀 + 𝑁 + 𝐿) ∗ 2
𝐹𝑠 = (𝑀 + 𝐿) ∗ 10 𝐻𝑧 𝐹𝑠 = (𝑁 + 𝐿) ∗ 10

5(b). Design (find transfer function, draw block diagram and approximate CO3 C 02
frequency response plot) a resonator/notch filter. Assume, parameters based
on the instructions of Table-5.
Table-5
Condition Assumption Condition Assumption
M is even Resonator M is odd Notch filter
Allowed frequency= Frequency to be removed=
(M+N+L)*3 Hz (M+N+L)*3 Hz
Sampling frequency= 150 Hz Sampling frequency= 150 Hz
N is even 𝑀+𝐿 N is odd 𝑁+𝐿
𝜆= 𝜆=
50 50

6. Viva-voce: The viva-voce will be held at university according to the routine published by 20
dept. Submitted assignment will be justified by viva-voce.

Page 3 of 3
 International Islamic University Chittagong
Department of Electrical and Electronic Engineering

Final Assessment Autumn-2020 Program: B.Sc. Engg. (EEE)

Course Code: EEE-3603 Course Title: Digital Signal Processing-I
Time: 5 hours (Writing - 4 hours 30 Full Marks: 50 (Written 30 + Viva/Viva-Quiz-20)
minutes + 30 minutes submission time)
[Answer each of the questions from the followings; Figures in the right margin indicate full
marks. Answer script must be submitted through online method within 5 hours from starting
time. Also, write down the Q. Set on the front page of your answer script]

Q. Set-A
1(a). Determine the z transform of CO2 E 02
1 5
𝑥(n) = (2)n u(n)+ (2)𝑛 𝑢(−𝑛 − 1).
Also draw ROC and pole zero location of X(z).

1(b). Determine the inverse z transform of X(z) = 1 CO2 E 02

1−𝑧 −2
When (a) ROC │z│>1

(b) ROC │z│< 1.

1(c). Find the z transform of CO2 E 02
(a) x(n) = 𝑎𝑛 𝑢(𝑛) ∗ 𝑢(𝑛)
𝑛
(b) x(n) = sin 2

2(a). Write down the advantages of Linear Phase FIR system. CO2 R 01
2(b). Determine the cascade form realization of the system described by the CO2 E 02
system function
1 1
H(z) = 1 − 𝑧 −1 + 𝑧 −2 + 2𝑧 −3
5 7

2(c). Determine the parallel realization with second order and first order CO2 E 03
subsections for a LTI system described by the system function
1+z−1 +3z−2
H(z)=1+.25z−1−.1z−2

3(a). Determine the circular convolution by using graphical method of two CO2 E 02
sequences
X1(n)= {4,1,5,1} and X2(n)= {4,5,6,7}
↑ ↑
3(b). Compare between DFT and DTFT. CO2 R 02
3(c). Compute the linear convolution of the following two sequences by circular CO2 E 02
convolution x(n) ={1,4,5,8}, h(n)={h1,h2,h3}.
↑ ↑
Assume the values of h1, h2 &h3 .

4(a). Explain the spectral inversion and spectral reversal technique for low- CO3 E 02
pass to high-pass filter conversion.
4(b). Design a FIR low-pass filter having length 7 and ꞷc=1radian/sample. CO3 C 03
Use Hanning window for the design process.

Page 1 of 2
4(c). Why an ideal low pass filter is not physically realizable? CO3 R 01
5(a) Design a low pass filter with cut-off frequency of .3𝜋 using bilinear CO3 C 03
transformation applied to the analog filter
ꞷ
𝐻(𝑠) = 𝑠+ꞷ𝑎
𝑎
where ꞷ𝑎 is the frequency of the analog filter and T=1s.
5(b). “Shifting the passband of DC notch we can design Hum removal filter” CO3 E 03
– Explain the above statement.
6. Viva/Viva-Quiz: The time of viva/viva-quiz will be declared in google
classroom.

Page 2 of 2
 International Islamic University Chittagong
Department of Electrical and Electronic Engineering

Final Examination Spring-2020 Program: B.Sc. Engg. (EEE)

Course Code: EEE-3603 Course Title: Digital Signal Processing I
Time: 5 hours (Writing - 4 hours 30 Full Marks: 50 (Written 30 + Viva/Viva-Quiz-20)
minutes + 30 minutes submission
time)
[Answer each of the questions from the followings; Figures in the right margin indicate
corresponding course outcome (CO), bloom’s level and full marks respectively. Answer script
must be submitted through online method within 5 hours from starting time. Also, write
down the Q. Set on the front page of your answer script]
Q. Set-EE
***In the question you will get many cases where you were asked to assume/take a parameter
or figure. It is expected that your assumption will be different from others as well as from book
or lecture materials. Correctness of your answer and dissimilarity of assumption with others
are two main key points of getting good marks.

1(a). Suppose, z is a complex number and z=a+ib. CO3 Ap 02

Where, a=1st digit (from right) of your ID.
b= 2nd digit (from right) of your ID.
Place z in argand diagram and draw an approximate plot of the real
part of 𝑧 −𝑛 .
1(b). Evaluate the convolution sum of following two discrete time causal CO3 E 03
signals using convolution theorem of z-transform.
𝑥(𝑛) = {𝑥0 , 𝑥1 , 𝑥2 , 𝑥3 }
𝑦(𝑛) = {𝑦0 , 𝑦1 , 𝑦2 , 𝑦3 }
Assume, the sample values from any numbers between -5 to 5. Every
sample value should be different.
1(c). The z-transform of a causal signal x(n) is given below CO3 Ap 01
1
𝑋(𝑧) =
1 − 𝑎𝑧 −1
1𝑠𝑡 𝑑𝑖𝑔𝑖𝑡 𝑜𝑓 𝐼𝐷−2𝑛𝑑 𝑑𝑖𝑔𝑖𝑡 𝑜𝑓 𝐼𝐷
Where, 𝑎 = 2
Draw the time-domain behavior of the signal x(n).

2(a). Suppose, you have converted an analog signal into digital with CO4 Ap 01
sampling rate Fs (where, 𝐹𝑠 = ∑ 𝑎𝑙𝑙 𝑑𝑖𝑔𝑖𝑡𝑠 𝑜𝑓 𝑦𝑜𝑢𝑟 𝐼𝐷). The
duration of analog signal is 2 s. If you want to transform this time
domain digital signal into frequency domain by a microprocessor, how
much time will require? Suppose, the transformation method is DFT
and your processor needs 1 ns time to execute an
addition/multiplication.

Page 1 of 3
2(b). Assume, any two discrete time sequence 𝑥1 (𝑛) and 𝑥2 (𝑛) having 4 CO3 E 05
samples in each sequence. Using these sequences proof that
multiplication of DFT’s of two sequences is equivalent to the circular
convolution of the two sequences in the time domain. [Remember that
correctness of your answer and dissimilarity of assumption with others
are two main key points of getting good marks.]

3(a). Determine a cascade form realization for the following discrete time CO4 E 04
system with system function
𝐻(𝑧) = 𝑎 + 𝑏𝑧 −1 + 𝑐𝑧 −2 + 𝑑𝑧 −3 + 𝑒𝑧 −4 + 𝑓𝑧 −5 + 𝑔𝑧 −6
You have to take the values (a to f) from your matric ID comparing
with the expression ‘ET-abcdef’. ‘g’ is a constant and assume, the
value of ‘g’, any value between 2 to 5.

3(b). Determine the direct form-I realization of the system described by the CO4 E 02
system function
(1 − 𝑎𝑧 −1 )(1 + 𝑎∗ 𝑧 −1 )
𝐻(𝑧) =
(1 − 𝑏 𝑧 −1 )(1 − 𝑐𝑧 −1 )
Where, ‘a’ is a complex number.
b and c are real constants.
You have to assume the values of a,b, and c before implementation of
the system.

4(a). What is the kernel of a filter? Explain the ways of converting the CO5 U,C 03
kernel of low pass filter into high pass filter.
4(b). Why frequency scale is leveled 0 to 0.5 in the frequency response CO5 An 02
curve of the digital filter?
4(c). Following is the impulse response of a low pass filter. Suppose, the CO5 Ap 01
cutoff frequency of the filter is 0.2, find the impulse response of high
pass filter for the cutoff frequency 0.3.
h(n)= [0 0.05 -0.07 -0.01 -0.08 0.1 0.6 0.7 0.6 0.1 -0.08 -0.01
-0.07 0.05 0]

5(a). Design a FIR low-pass filter using following constraints CO5 C 05

- Passband edge ripple, 𝜔𝑝 : Assume, any value between
0.2𝜋 𝑡𝑜 0.35𝜋
- Stopband edge ripple, 𝜔𝑠 : Assume, any value between
0.6𝜋 𝑡𝑜 0.8𝜋
- 𝜔𝑠 > 𝜔𝑝
Minimum stopband attenuation: -60 dB
5(b) Draw the approximate frequency response of your designed filter in CO5 Ap 01
problem 5(a). If you subtract the impulse response obtained in 5(a)
from 𝛿(𝑛) what will be the frequency response?

Page 2 of 3
6. Viva/Viva-Quiz: The time of viva/viva-quiz will be declared in google 20
classroom.

Page 3 of 3
 International Islamic University Chittagong
Department of Electrical and Electronic Engineering

Final Evaluation, Autumn-2019 Program: B.Sc. Engg. (EEE)

Course Code: EEE 3603 Course Title: Digital Signal Processing
Time: 12 hours Full Marks: 40
[In the question you will get many cases where you were asked to assume/take a parameter or
figure. It is expected that your assumption will be different from others as well as from book or
lecture materials. Correctness of your answer and dissimilarity of assumption with others are
two main key points of getting good marks.]

SET-EE

1(a). “Z-transform does not uniquely specify the signal in the time domain CO2 E 03
without ROC.”
Justify the above statement with an example. It is expected that your
example will be different from others.
1(b). Explain the physical significance of Z-n that is used in the expression of CO2 An 04
z-transform. Take any arbitrary value of z (z must have a real part and
imaginary part) and plot the oscillation associated with the real part of
Z-n.
Note: There are infinite of ways to choose a value of z from complex
plane. It is expected that your choice will be unique from others.
1(c). Give an example of a finite duration signal which has ROC: entire z- CO2 Ap 01
plane except z=0 and z=∞. Also find the z-transform of that signal.

2(a). Assume the transfer function of a linear phase FIR system. Then sketch CO3 Un 03
the direct form realization of that system. How many additions and
multiplications instructions are required per output point? Also
determine the number of memory block.
2(b). Assume the transfer function of an IIR system having four or more CO3 Un 05
poles and three or more zeros. Then sketch the cascade form structure
of that system. How many additions and multiplications instructions
are required per output point? Also determine the number of memory
block.

Page 1 of 2
3. Assume any two discrete time sequence 𝑥1 (𝑛) and 𝑥2 (𝑛) having 4 CO2 E 08
samples in each sequence. Using these two sequences proof that
multiplication of DFT’s of two sequences is equivalent to the circular
convolution of the two sequences in the time domain. Remember that
correctness of your answer and dissimilarity of assumption with others
are two main key points of getting good marks.
[Hints: Suppose, 𝑥1 (𝑛) = {1,2,1,2} and 𝑥2 (𝑛) = {0,9,1,4}.
Calculate 𝑋(𝑘) = 𝑋1 (𝑘)𝑋2 (𝑘). Now using IDFT find x(n) from X(k).
Again find the circular convolution of two sequence 𝑥1 (𝑛) and 𝑥2 (𝑛).
This result must be equal to x(n).]

4(a). Which filter is useful at very low frequency between analog and digital CO3 R,Un 04
filter? Write some applications of digital filter in your own words.
4(b). How could you use the step response of a digital filter to measure the CO3 An 04
performance of that filter?

5(a). Design a FIR low-pass filter with ωp and ωs (assume the value of ωp CO3 C 06
and ωs such that 0.2𝜋 ≤ 𝜔𝑝 ≤ 0.8𝜋, 0.2𝜋 ≤ 𝜔𝑠 ≤ 0.8𝜋 and 𝜔𝑠 > 𝜔𝑝 ).
You must use a window function to reduce Gibbs phenomenon. Choice
any window function among Bartlett, Hanning, Hamming and
Blackman window.
5(b). Draw the approximate frequency response of your designed filter in CO3 Ap 02
problem 5(a). If you multiply the impulse response obtained in 5(a)
with (-1)n what will be the frequency response?

Page 2 of 2