L-3/T -2/EEE Date: 06/03/2018

BANGLADESH UNIVERSITY OF ENGINEERING AND TECHNOLOGY, DHAKA

L-3/T-2 B. Sc. Engineering Examinations 2016-2017

Sub: EEE 311 (Digital Signal Processing I)

Full Marks: 210 Time: 3 Hours

USE SEPARATE SCRIPTS FOR EACH SECTION
The figures in the margin indicate full marks.

SECTION -A
There are FOUR questions in this section. Answer any THREE questions.

1. Two LTI system are connected in parallel as shown in figure below. The input-output
relationships of. the DT systems are also given. For an input of
x(n) = (-lr (u(n)- u(n-2)],
-~-_.- -----_ .. _---

n;r System 1: .
. DTSystem 1
YI(~)= IX(k) x(n) y(n)
k=n-2 .

DT System 2:.
n-4
Y2(n)~ - L>~(k)
'k=n-7
L-J OT System 2

Fig. for Q.l

(a) Determine and plot tlie impulse responses hj(n) and h2(n) of the DT systems 1 and
-----J
2, respectively. Also, determine whether the DT systems 1 and 2 are stable and causal. (10)
(b) Determine and plot the impulse response of the overall system. (5)
(c) Determine the outputy(n) of the overall system by convolution summation. (10)
(d) Find the output y(n) using z-transform. (10)

2. (a) Define poles and zeros of a system and write their significances. Show that a causal
LTI system will be stable provided all its pole lie inside the unit circle. (10)
(b) Describe SchUr-Cohn stability test. Using this test determine the stability of a

system with h(n) = [-!'(-1)" +i+-!.n]u(n). (15)

442
(c) Determine the unit step response of a system governed by
y(n) = 0.9y(n -1)+ 0.81y(n - 2)+ x(n) under the initial condition of y(- 2) = y(-I) = O. (10)

3. (a) Define correlation and its different forms. Show that autocorrelation at zero lag has
the highest value. With examples, describe the use of autocorrelation. (15)
(b) Find correlation between {I 2 3 4} and {4 3 2 I}. (10)
t t
(c) An analog signal xa (t) = sin 4807Z"t + 3 sin 7207Z"t is sampled at 600Hz. Determine (10)
(i) Nyquist sampling rate and folding frequency
(ii) The frequencies in the resulting DT signal x(n) and

(iii) The reconstructed signal if x(n) is passed through an ideal DAC.

Contd P/2
 =2=

EEE 311

4. (a) For the following system, determine its impulse response hen). Consider hen) is the
equivalent of two series-connected systems h](n) and h2(n). If ~(n)=o(n)+o(n-l),
find h2(n). ~ (15)
--------- .-------:-II

~(9i

Fig. for Q. 4(a)

(b) Derive the expression of DTFS. Differentiate DTFS from CTFS. Determine the

Fourier coefficients of the signal x(n) = cos 27l" n+sin 27l"n. (10)
3 5
(c) Determine and sketch the magnitude and phase response of a system governed by
4y(n)=x(n)-2x(n-l)+x(n-2). (10)

SECTION-B
There are FOUR questions in this section. Answer any THREE.

r=oo

5. (a) Given that x[n] = L x[n - rN]. Discrete Fourier series coefficients corresponding (10)
r=-co

to x[n] is X[k] and discrete Fourier transform of x[n] is denoted as X(ej(j)). Show that

,/ X(k] can be expressed as a periodically sampled version of X(ej(j)) between (j) =0

and (j) = 27l".
(b) Let x(n] = 0, n < 0, n > 7 be a real eight-point sequence and let X(k] be its eight-

point DFT. Let w[n] = 0, n < 0, n > 3. be a four-point sequence and let W[k] be its four-

point DFT. If W[k] = X[k]+ X[k + 4]' express w[n] in terms of x[n]. (10)
(c) Two sequences xl[n] and x2[n] are given as follows, Xl [n]= {-I 2 I},

x2[n]={2 0 -I} t (15)

t
(i) Find Xl (k) and X2 (k), the 3 point DFTs of XI [n] and x (n],
2 respectively.

(ii) If v(n] is three point circular convolution between XI (n] and x2 [n], calculate

v[n] from Xl (k) and X2 (k) using an inverse DFT.

Contd P/3
 =3=

EEE 311

6. (a) A causal FIR system has the impulse response, (10)

h(n] = {2 -4 8 8 -4 2}
t
(i) Mention the type of filter and its coefficients.
(ii) Draw the signal flow graph for this filter using three multipliers (maximum).
(b) Fig Q.6(b) shows that the signal flow graph for a causal discrete-time LTI system.
Branches without gains explicitly indicated have a gain of unity. (10)
(i) Determine the difference equation relating x(n] and y(n].
(ii) Determine h(I], the impulse response at n = 1.

------_._----_._ .. -- --. - -_._------_ __

.. .•.,.._.. __ :.....-.

-- ._._e '_" --G-il---~.~CJ!J-------~1

\

(c) Find H(z) for the signal flow graph ofa causal LTI system shown in Fig Q.6(c) (i). (15)
Carefully observe the signal flow graph shown in Fig Q.6( c) (ii). Does the system
corresponding to Fig Q.6(c) (ii) represent the same input-output relationship as the
system corresponding to Fig Q.6(c) (i)? Justify [You do not need to show the H(z) for
the second system]. All branches of the signal flow graphs in this problem have unity
gain unless specifically indicated otherwise.
----~--_.- --
~?_.~-,
--_._- -. . ~~ ---' -":l:-,-,--
l-')- ._' --- --.-- -----.-.--,--.- ..-.
, ' -1
-.. ---~
•..._-~_._---------.---_._--

, , ~ffiJ .' , . ,'

-, '-, - ---.-------~''J-.,--~-b-(cj-f'-)..,-.~-.------'----.:----.----~:;~

-1 ~---~-O-;-f'>~=.-=.
2

=ar--r-~-"
'0~

I ~-:
z-_-'_'
-,~.
:'X.Q-J-~_,,-- --- -< t O?_I ~-'---, ----:..--.-.-----~
I • ' .--.----r;~j-. &..-:~-E(;.)---~~.)-.-------------_.--~,
I

7. (a) What are the two main problems of implementing an ideal FIR LPF? Mention how
to overcome the problems? (8)
(b) What is the problem of using rectangular window? Explain logically why the use of
Hamming window can reduce the problem? (7)
Contd P/4
 =4=

EEE 311
Contd ... O. NO.7

(c) Use the Kaiser window method to design a real-valued type II FIR filter with
generalized linear phase that meets the following specifications: (20)
0.9 < H(eJaJ )< 1.1, 0~ Iwl ~ O.2n,
- 0.06 < H(eJaJ )< 0.06, 0.3n ~ Iwl ~ 0.475n,
1.9 < H(e JaJ
) < 2.1, 0.525n ~ Iwl ~ n

Find Kaiser window function and impulse response of the filter.

8. (a) In the impulse-invariance method of discrete-time IIR filter design, what is meant
by the term "impulse invariance"? What is the major limitation of this method?
Mention a simple way to overcome the limitation. (10)
(b) Consider designing a discrete-time filter with system function H(z) from a
continuous-time filter with rational system function Hc(s) by the transformation

(i) Find the relationship between ill and n. (25)

(ii) Show that if Hc(s) is a rational function with all its poles inside the left-half s-
plane, then H(z) will be a rational function with all its poles inside the unit circle
of the z-plane.
(iii) Find H(z) using the relationship obtained in (i) and considering a
continuous-time Butterworth filter. The discrete-time filter is expected to meet
the following specifications:

0.9 ~ I H(eJaJ )1~ 1, 0~ wI I~ O.2n

1 H(eJaJ )1 ~ 0.2, 0.32n ~ I w I~ n
.'

L-3/T -2/EEE Date: 16/08/2017

BANGLADESH UNIVERSITY OF ENGINEERING AND TECHNOLOGY, DHAKA
I
L-3/T-2 B. Sc. Engineering Examinations 2015-2016
I
Sub: EJEE 311 (Digital Signal Processing I)

Full MarJs: 210 . Time: 3 Hours

USE SEP fRA TE SCRIPTS FOR EACH SECTION
The figures in the margin indicate full marks.
___ .. __ _1 ••••......•.••••••.......•..•.••...................... _ .• ._ _ _ _ ..........•.. _ ••••••••. __ __ .••• __ . •.......•.•..

I SECTION -A
There are FOUR questions in this section. Answer any THREE.
Symbols add abbreviations have their usual meanings.

1.
I
(a) Consider the following system for processing a continuous-time signal with a
discrete-time system: (20)

--- ...--.---.---.-.
-f...-....
------.--.------.-----.---
.. ---''''ri' .--' ----~.-'~~~_.
---
.
I

"~c+r-.
_____________
;-Mf- ~.~_-l:tCi~J~(~~)-:D/A-~~~:~--=
•...•__ ~'_ .•._.R ~__ ._._~._._..l_ ..__. ._~_~~_
_-_. __.-. --~. .

--==7-::- .. --=-....:~------ ..:...£::t a'. ~--.---- .._.-....-.-.-...---.... =- ..

~--

The frequency response (i)fthe discrete-time filter is

2( 1 e-jOJ)
H(e )= jOJ 3-
1 - j(u
1--e
3

If Fs = 4 KHz and Xa (t) = 5 cos 2000 1tt- 2 sin( 30001tt - : ) - 2 sin( 5000nt + : ), find

the output Ya (t ) .

(b) A continuous-time signal xa (t) is composed of a linear combination of sinusoidal

signals of frequenei es F 1 rz, F, Hz, F, Hz and F, Hz. The signal x, (t) is sampled at an

10-kHz rate. The sampledl sequence is then passed through an ideal low pass filter with
a cut-off frequency of 4 kHz, generating a continuous-time signal Ya(t) composed of

three sinusoidal signals of frequencies 350 Hz, 575 Hz, and 815 Hz, respectively. What
are the possible values or h F" F, and F,? Is your answer unique? If not, indicate
another set of possible vallues of these frequencies. (15)

2. (a) Consider the discrete-time system shown in Fig. for Q.2(a). (18)

Contd P12
 =2=

EEE 211

-_.- - _.~.-- .. --- -- - --- ~-_.,.....-- ,._ ..-

-_.~-~-~

(i) Determine its impulse response hen) using a time-domain technique and
explain if the system is stable. Given that y(-l) = 2. .
(ii) Determine a tlealization for its inverse system, that is, the system which
produces x(n) as ak output when yen) is used as an input.
(b) Determine the totll solution yen), for n :::: 0, of the discrete-time system

characterized by the fOlllwing system function: (17)

H(e jOJ)_
- 1+ e-jOJ
1_ 6e,j2OJ

for an input x(n) = 2" u~n) and with initial conditions y(-l) = 1 and y(-2) = -1. Use a
time-domain technique.

3. (a) A causal LTI system has the function (20)

H(z)= l+z-'
l-~z-1
2
Find the input x(n) that WI' ill produce the output.

u(-n-l) y(n)=-- 1(1)" 4

- u(n)--(2) 11

3 4 3 I
(b) Sketch roughly the magnitude IH(w~ of the Fourier transforms corresponding to

the pole-zero patterns of systems given in Fig. for Q.3 (b). (15)
.' /l' .

"c~~~lt--
.- - _.,........
;> P -_..L-\;~'1'..-t=._-._.-
G" ( ~.0 l~
.. - .. _,
PoI e ~
-.. --- ..-- _J. ,
. "",jW1+"'-'-', .,_._ -- .
j

~ -.--- ..-.._ ~ _ x..()-'--q.~~ _ _ ~~~.l~

-7---...... 2.
.~-..-_.....-.. \ /'
, ~~.:\.~ _...........-- --- ---:- __ _ ':l!. -.-.- -.._............•........
__ L __.
, :- ..p>61.-e,-- _".. .. ;._ _ *ctNl?+I'....•
~r(J- _ ,__.._•.., _ _ _ _ ' ''''''''
It)
."--'" --..- --..-.... ._ -.- -F~:.;.--75--- ~._.~'7~~
I

)-__ _..- _ _ _ --l=1:f~j......... .

.•. -- ..••.• - •.•••• - •.......... - ••.••• _. __ .........• c •••• __ •••.....•.•.•.•..•• , .d.~!:W ..\)I.l ..2?~ ..::::::'::=::':~=~':'::::'::'". . , _

Contd P/3
 =3=

EEE 211

4. (a) The system (20)

1
yen) = -yen -1)+ x(n)
2
is excited with the input

x(n)= (1)"
"4 u(n)
Determine the sequences r.u(e), r",,(.e), rXY(.e) and ryy(.e). Use the Z-transform

technique.
(b) The causal system (15)
1
H(z) = N

1 + :L:akZ-k
k=1

is unstable. We modify this system by changing its impulse response hen) to

h/(n)= ),,"h(n)u(n)

(i) Show that by prloperty choosing A we can obtain a new stable system.
(ii) What is the difference equation describing the new system? .

SECTION-B
There are FOUR questions in this section. Answer any THREE questions.

5. (aJ A linear time invariJt system is realized by the flow graph shown in Fig. 5 (aJ. (15)

-?i.C-] ~ - --'-~--~f'-1--=
--------.---~,. -'---,- --:-- -'--1~..~
..::1---_ .._------_._.~~._._-
--- ..----.-- 1" . ---- ---.~-.----.--F;--:~s=-lZ5~.-:
.. --_.--?:...~- --'-f-- .._-:._--- -l.-.-- ..- ..-.--.d----..-....
-:;.._.-:- ------
,l
.- .... ---'-' ~""":""-
.-::..~~--

(i) Write the difference equation relating x[n] and y[n].

(ii) What is the traAsfer function of the system?
(iii) In the rea1izJtion of Fig. 5 (a), how many real multiplications and real
additions are requi~ed?
(iv) Is it possible Jo reduce the number of storage registers by using a different
structure? I
I
(b) Draw the signal flow graphs for the direct form II and transposed direct form II
implementations of the UTI system with transfer function. (12)
5 -I 1 -2
1 +-z +-z
H(z) = 6 6
1 -I 1 -2
1--z --z
2 2
Contd P/4
 =4=

EEE 211
Contd ... Q. NO.5

(c) Why is the flow graph shown in Fig. 5(c) noncomputable? (8)
I
-"-ac;r I
_._._----~- --_._-----_._._._-~.- .--_ .._~...._._-----_.~----;

. ._M......-__ -----_._--_.----
-.'--- ..Po- -:,
From the transfer function of the network, obtain a flow graph that is computable.

. I I

6. (a) For two sequences, xl[n] = {I 00 I}, x2[n] = {2 I}, show that v[n] = Ig[n~4r],
Itt ~
where v[n] is the 4-pt circular convolution between Xl [n] and x2[n] and g[n] the linear
convolution between xl[h] and x2[n]. . (10)
(b) For two sequences, xt[n] = {I 0 2} and x2[n] = {1 03}, (18)

(i) Using 3-PtDFj'

Itt
find X,[k] and X,[k] in terms of W: ~ exp
[-J'2Jrkn]
N .

(ii) Calculate x[n] = xl[n] Q) x2[n].

(c) Using the IDFT, find x[O] from X[k]. (7)

7. (a) A discrete-time filtel; with generalized linear phase that meets specifications of the
following form: (15)
IH(ejW ) ~ 0.01, 0 ~ Iml ~ 0.25n
0.951:; IH(e jW
) ~ 1.05, 0.35n ~ Iml ~ 0.6n
IH(ejw) ~ 0.01, 0.65n ~ Iml ~ n

(i) Determine the minimum length of the impulse response and the value of the
Kaiser Window pdrameter P for a filter that meets the specifications.
(ii) Determine the belay of the filter.
(b) Find the ideal imp~lse response hd[n] to which the Kaiser Window should be
applied and also fmd thel aetnal response h[n]. (10)
(c) A multiband discrete-time filter with generalized linear phase has the following
I
magnitude response as shown in Fig. 7(c). (10)

~~\-&,i I - -_.--~--======---------..-._-_.
__._---....._....-..'.l
, -------.-G1r '_';'
""
__'_ _ .e- • .tUl._'1.CcJ__ 1

; 0
~ _-_ .._._- ..--_ ..(A~- --.-.----.---- _._'- -.----- ..- --.--- -..- - ..-- --- - .
i
~--.--._.---6-
..
--.~--,-. ~--_.b
Find hmb[n].

Contd P/5
•

=5=

EEE 211

8. (a) A discrete-time LPF is to be designed by applying impulse invariance method to the

continuous time Butterwrrth filter with the following specifications . (12)

0.89125 ~ IH(e 1,jlll

~ ~ 0 ~ lcol ~ 0.21t
IH(ej(d~~ 0.17783, 0.31t ~ lcol ~ 1t

for the Butterworth LPF, the magnitude squared function IS of the form,

IH,Unt = 1 2N I
l+U~J
(i) Calculate the filter order N.

(b) How is H(Z) obtaine! from H(S)? (13)

(c) Mention the names and draw the shapes of commonly used windows. Explain
briefly the reasons behind generation of ripples in the windowing method for FIR filter

design. (10)
L-3/T -2/EEE Date: 01108/2016
BANGLADESH UNIVERSITY OF ENGINEERING AND TECHNOLOGY, DHAKA
L-3/T-2 B. Sc. Engineering Examinations 2014-2015

Sub: EEE 311 (Digital Signal Processing I)

Full Marks: 210 Time: 3 Hours
The figures in the margin indicate full marks.
USE SEP ARA TE SCRIPTS FOR EACH SECTION

SECTION -A
There are FOUR questions in this Section. Answer any THREE questions.

1. (a) A causa11inear time-invariant system is described by the difference equation (15)

y[n]-5y[n -1 ]+6y[n - 2]= 2x[n -1]

(i) Determine the homogenous response of the system.

(ii) Determine the impulse response of the system.
(iii) Determine the step response of the system.
(b) For the system in Figure Q. No. 1(b), determine the outputy(n] when the input x(n] is

jOJ 1 IlUl < ~

bIn] and H(dj is an idea110w pass filter, H (e ) = 0 (20)
{ ~ < IlUl ~ Jr

7' .... .. )~ ., (~l~U.l

X(h3-J:i.J~ EI)" '" :JG>1-

~ __ .. =,~)==I-~~~~~~-I.',-"
~ __~. .~i_-~-HC..~J_' ..{i~~~W-.
--=r I
-2/r
IJ- -X
cI:J=_~.~..
-1Y2,. 7'2.- K
1
2/t
'. :tV--
'"----- --_._------------~---------- --~---------
2. (a) A linear time-invariant (LTI) system is described by the input-output relation

y[n]= x[n]+ 2x[n -1 ]+x[n - 2]' (18)

I
I '
\ .

(i) Is this a stable system?

(ii) Determine Frequency response of the system, H(dj.
(iii) Plot the magnitude and phase of the frequency response.

(iv) Consider a new system, HI (ejOJ )= H(ej(OJ+7r)). Determine h[n].

(b) Consider the cascade ofLTI discrete - time system shown in Figure Q. No. 2(b). (17)

_~.a:a-L.T-~JI:t-~-r
'-'~'-'--C-, -.- -'----~-. ,.-~. -----~---~_.-

kf..~!lLTttt~ f--~
~cr.J- F~{b)
" H J (te."~) ,

The first system, HI elOJ = ( . ) {1o IlUl < 0.5Jr .

0.5Jr ~ IlUl < Jr

The second system; y[n] = w[n]- w[n -1]

The input to the system is x[n] = cos(0.6Jrn)+35[n-5]+2

Determine the output y[n] and Y(ejOJ

).

Contd P/2

1.
. "r,
"
 =2=

EEE 311

3. (a) The input to an LTI system is x(n]= (~r u(n]+ 2nu(-n -1].

(20)

(i) Find the system function H(z) of the system. Plot the poles and zeros of H(z) and
indicate the Region of convergence.
(ii) Find the impulse response h(n] of the system.
(iii) Write the difference equation that characterizes the system.
(iv) Is the system stable? Is it causal?
(b) let x(n] be the sequence with the pole-zero plot shown in Figure Q. No. 3(b). Sketch

the pole-zero plot for, y(n] = (~) n x(n]. Specify the region of convergence of Y(z). (15)
-----
~
_._' _'.~"'ft~

i '-Pc-fi~C2-t*3q,-)-

4. (a)
--1 ~ ..._--.-.-.---

~r".J ~H ~[".1 £H.0'1-j

For the system in Figure Q. No. 4(a), given that Hd(z) is causal, stable and non- minimum
phase, given by

Choose Hc(z) so that it is stable, causal and the magnitude of overall effective frequency

response is unity. (15)

(b) The system for changing the sampling rate by a non-integer factor is given by Figure

L = up sampling factor = 2. M = down sampling factor = 3

Contd P/3

1
('..
•
_ )
 =3=

EEE 311
Contd ... Q. No. 4(b)
Sketch

(i) IX(e j
()) ~

(ii) IXe (ej(V ~

(iii) IX (e
i
j
()) ~

(iv) IX d (e j
()) ~

Where, x[n] is obtained from xc(t), xc(jQ) is given in Figure Q. No. 4(b)-2.

SECTION-B
There are FOUR questions in this Section. Answer any THREE questions.
Symbols have their usual meanings.

5. (a) What are the two main problems in designing ideal low pass filter? Briefly describe

the principle of Windowing method of FIR filter design. (10)

(b) What is the limitation of the impulse invariance method of IIR filter design? How the

principle used in the bilinear transfer method can overcome the limitation? (10)

(c) Fig for Q. No. 5(c) shows the signal flow graph for a causal discrete time LTI system (15)

(i) Find H(z) = Y(z) .

X(z)
(ii) Determine h[l], the impulse response at n = 1.
.- -- - ~ -----
. -----:-,
,
,

i 1-
X[!)) ~[nJ-.
-I
~c; -1:
-I 2
,_._------
-I
~
4

_____ F_~31a.No. ;(e)

6. (a) Consider a finite-length sequence x[n] oflength N. Two finite-length sequences xl[n]
and x2[n] oflength 2N are constructed from x[n]. In Fig. for Q. No. 6(a), x[n]' xl[n] and
x2[n] are shown (solid line is used to suggest the envelope of the sequence values. The
N-point DFT of x[n] is denote by X[k] and the 2N point DFTs of xl[n] and x2[n] are

denoted by X1[k] and X2[k], respectively. (15)

Contd P/4
 =4=

EEE 311
Contd ... Q. No. 6(a)
(i) Specify how X2[k] can be obtained if X[k] is given.
(ii) Show that X[k] can be obtained by decimating XI[k] by 2.

--' ---------'--------------------------'-------~

(b) Given that xI[n] = [ 1 0 1 2] (20)

t
x2[n] = [ 1 0 1]

t
Yl[n] =xd«n-2))4], Y2[n] =xd«n+l))4],
(i) Find YI[k] and Y2[k]
(ii) IfYc[n] = YI[n] @ Y2[n], find Yc[k] and using Yc[k] find yc[n] .

7. (a) For the system shown in Fig. for Q. No. 7(a), for the basic filter block H(ejOl) it is
given that

h[n]={ h[-n] , -Lsns.L

o ,otherwise

H( dOl) satisfies the following approximation error specification: (15)

(1- ° 1) s IH(e)l1J ~ ~ (1+ °1), 0 S OJ S OJ p

- 02 S IH(e)l1J ~ s ° 2, OJs S OJ S 7r

(i) Determine whether the overall impulse response g[n] is FIR and symmetric.
(ii) Find the approximation error specification for G(dOl).

.
~[nJ .
t.rnJ
Heeiw) .... 'Wlt\} 14 ce?"')
2.

Contd P/5

\,
J
\
 =5=

EEE 311
Contd ... Q. No.7
(b) By using Kaiser Window method design Type-I FIR filter with generalized linear

phase that meets the following specifications: (20)

jaJ
0.9 < JH(e ~ < 1.1, 0::;;levi::;; 0.27l"

- 0.06 < IH(ejaJ ~ < 0.06, 0.37l"::;; levi::;; 0.4757l"

1.9 < IH(ejaJ ~ < 2.1, 0.5257l" ::;;levi::;; 7l"

(i) Determine appropriate values of ripple 8, ~ L1co and filter length.

(ii) Find ~[n] and h[n].

8. (a) The system function of a discrete-time system is

Assume that this discrete-time filter was designed by the impulse-invariance method with
h[n] = 3hc(3n), where l\;(t) is real. Find the system function Hc(s) of a continuous time

filter that could have been the basis for the design. Is your answer unique? (10)
. (b) Consider designing a discrete-time filter with system function H(z) from a continuous
time filter with rational system function Hc(s) by the transformation

(i) Show that the j n axis of the s-plane is mapped to the unit circle of z-plane.

(ii) Find the mapping between n and co.

(iii) Consider the following specifications for the desired discrete-time filter

0.8::;; IH(ejaJ ~ ::;;1, 0::;;levi::;; 0.257l"

IH(ejaJ~::;;0.15, 0.47l"17l"1::;;7l"

using the transformation specified at the beginning find H(z). (25)

L-3/T-2/EEE Date: 27/0112016
BANGLADESH UNIVERSITY OF ENGINEERING AND TECHNOLOGY, DHAKA
L-3/T-2 B. Sc. Engineering Examinations 2013-2014

Sub: EEE 311 (Digital Signal Processing I)

Full Marks: 210 Time: 3 Hours
The figures in the margin indicate full marks.
USE SEP ARA TE SCRIPTS FOR EACH SECTION

SECTION -A
There are FOUR questions in this section. Answer any THREE.
Symbols bear their usual meanings.

1. (a) Consider the following difference equation (18)

y(n] -7y(n-1] + 10y(n - 2] = 2x(n] + 5x(n - 1]
Determine the step response of the system which is described by the above difference

equation. Consider y[-l] = 14 , y[_ 2] = 49 .

5 25
(b) Using convolution sum property find y(n] when (17)
x(n] = 28(n + 1] + 8(n] + 28(n - 2]

~TnJ.:.J~~~,[n]~.~=n~~;I~J=Hn=3}.J
•.•.•••
~~.._ ••••.
~••~.rnJ!
-~,~..'''----'~'._._... ,.~ . I
"

==."",~,~,~~~.=.~
••,~.~~,~
..~~~8,'~~£
...-'~.~.~~~_~2
•.'~~'-,'.'._
••-.'.'
•.-_".",~
__-.-_.:'~
••,-,-.,
--'=~__~

2. (a) An LTI system is described by the following input-output relationship (20)

y(n] = x(n] + 2x(n - 1] + x(n - 2]
Find (i) H (~(j); (ii) Group delay (iii) Sketch the magnitude and phase of the frequency

response. (iv) Find y(n] when x[n] = cos

,
(1m3 + 1t).
6

(b) For x[n]=2n u[n-4] findX(~(j).

If S(ejW )= X (ej(w + 1t)), find 8(n]. (15)

3. (a) An LTI system has the following system transfer function (20)
1
H(z) = 1+z-

(1-~z-1)(1+~z-1)

(i) Find h(n] considering stable and causal system.

(ii) Find the z-transform X(z) of an input x(n] that will produce the following output:

y[n] = - -1 ( - -l)n u[n] - -4 2 n u[- n -1]

3 4 3
Contd P/2
 =2=
EEE 311
Contd ... Q. No.3

(b) The pole-zero diagram shown in Fig. for Q. No. 3(b) corresponds to the z-transfonn X(z)
of a causal sequence x[n]' Sketch the pole-zero diagram of Y(z), where y[n] = x[-n + 3].
Specify the region of convergence for Y(z). (15)
.-:-":-~~ ---:-- rn-:-. -- ~

1 ~'

4. (a)
.,. ------- -~- -- - -- -~-~-- ~ .•.... ~

.. :~n~..-k
-_._- --- .... .. __ ..
~.- -_ --~----~~-' -
..
J;~,,~.N.D. 4@).
For the above system, given that H(z) is causal, stable but non-minimum phase and C(z)
~ ..•.

associated with H(z) is given by (17)

P3 = PI;* P4 = P2*
How should Hd(Z) be chosen so that it is stable an~ causal and so that the magnitude of
the overall effective frequency response is unity.
(b) . -xc~t)'~-- e~;--~[~f~(,i")-5[;~TJ2.Jr--._~-,,>[Y)Tl
Fs.--. _.. . .. yawn ~OJVj>lij'

[X~(JJ2.)I_ IH(~W}I

... . _._----<~-..I_-+-~
-~ -3J 0 ~ Jl~ -St - fl(2. 0 Trf,- r l<)
_- .• - ... _.J::. .. __ ~ .. ,.~~-----, ---..---"'-.~'- ..• -.. -~
Fj..:~~__
~Q,..~o. 1(h},
In the system shown in Fig. for Q. No. 4(b), discrete-time signal x[n] is obtained from
continuous time signal xc(t) that follows magnitude spectrum shown in the figure. s[n] is
obtained by filtering x[n] using a discrete-time filter with magnitude spectrum shown in
the figure. Finally y[n] is obtained by downsampling s[n] with a factor 2. Sketch
w
(i) IX(e.iw)l, (ii) IS(e.iw)1 and (iii) Y( e.i ). (18)
Contd P/3
•

=3=

EEE 311
SECTION -B
There are FOUR questions in this section. Answer any THREE.

5. (a) Consider the sequences (12)

x](n) = {1, 2, 3, 4} and sen) = {l, 0, 0, O}
t t
Determine the sequence x2(n) such that
s(k) = X](k) X2(k)
where s(k), X](k) and X2(k) are 4-point DFTs.
(b) Determine the Fourier transform X(co) of the signal (12)
x(n) = {1, 2, 3, 2, 1, O}
t
Using the expression ofX(co), compute the 6-point DFT V(k) of the signal
u(n) = {3, 2, 1,0, 1, 2}
t
(c) Explain briefly how filtering of long data sequences is performed using 'over1ap~save'
method. (11)

6. (a) Why does spectral leakage occur during frequency analysis using the DFT? How can
you reduce spectral leakage? (12)
(b) Consider xa(t) be an analog signal with bandwidth B = 3 KHz. You need to do an
N = 2m point DFT of the signal with a resolution less than or equal to 50 Hz. Determine
(i) the minimum sampling rate, and (ii) the minimum number of required samples. (12)
(c) Obtain the direct form II realization for the system given by (11)

y(n) = ~ y(n -1) _! y(n - 2) +x(n) +! x(n -1)

4 8 3

7. (a) The impulse response of an FIR filter is given by (10)

hen) = {0.5, 1, 1, 0.5}
t
Explain why the filter with the given impulse response is unsuitable as a highpass filter.
(b) A highpass FIR filter is characterized by the following impulse response coefficients: (10)
hen) = {0.127, -0.026, -0.237, 0.017, 0.434}
t
Write down the coefficients of an equivalent lowpass filter with the aid of "frequency
shifting" property of the Fourier transform.
(c) Obtain the first three coefficients of an FIR lowpass filter that meets the following
specification using the window method. (15)
passband edge freq : 1.5 kHz
transition width: 0.5 kHz
stopband attenuation: > 50 dB
sampling frequency : 8 kHz

Contd P14
•
=4=

EEE 311

8. (a) In a certain signal processing application, the input signal is contaminated by a 50 Hz

component and its harmonic at 150 Hz. The signal is sampled at 500 Hz. Find the transfer

function of a simple digital filter to remove the interference and its harmonic. (10)
(b) Determine using the impulse invariant method, the transfer function of the digital
equivalent of a resistive-capacitive (Re) lowpass filter. Assume a sampling frequency of

150 Hz and a cutoff frequency of30 Hz. (10)

(c) A digital filter is required to remove baseline wander and artefacts due to body movement

in a certain biomedical application. The filter is required to meet the following requirements: (15)
passband: 1 - 128 Hz
stopband: 0 - 0.5 Hz
passband ripple: 3 dB
stopband attenuation: 20 dB
sampling frequency: 256 Hz
Determine the order of a suitable IIR filter and its transfer function H(z). The transfer
function of prototype lowpass filter is given by
1
order 1 : H(s)=-
s+1

order 2 : H(s)= __ I __
s2+.J2s+1

1
order 3 : H(s)= (2 )
(s+l) s +s+1
 , ~,
/} _:8'V.
.•.
~\"''V

L-3rr -2/EEE Date: 15/12/2014

BANGLADESH UNIVERSITY\ OF ENGINEERING AND TECHNOLOGY, DHAKA
L-3/T-2 B.Sc. Engineering Examinations 2012-2013

Sub: EEE 311 (Digital Signal Processing I )

Full Marks: 210 Time: 3 Hours
The figures in the margin indicate full marks.
Symbols have their usual meaning.
USE SEPARATE SCRIPTS FOR EACH SECTION

SECTION-A
There are FOUR questions in this Section. Answer I any THREE.
.

1. (a) Two LTI systems are connected as shown with their I/O relationships. (25)

~2..C~
(i) Determine and plot hl(n) and h2(n). F1'1« Q.., No, i~)
(ii) Justify the stability of the systems.
(iii) Justify the causality of the systems.
(iv) Find and plot hen) of overall system.
(v) Find yen) with x(n) = (_1)" [U(n) - U(n-2)] . _
(b) Define quantization and write its necessity. With an example, describe aliasing, its

causes and remedy. (10)

2. (a) Define correlation of LTI system. Describe how correlation is related with output.
Describe how signal buried in noise can be determined using autocorrelation. (18)
(b) Find the condition of stability of an LII system in terms of impulse response. Find the
condition of stability of a system having (17)
hen) = a"U(-n) + b" U(n-l).

3. (a) Describe poles and zeros in z-domain, and their significance.' Describe pole-location
and time domain behavior of a system with a suitable example. (17)
(b) Describe SchUr-Cohn stability test. Using this test, find the condition of stability for a
/ 1
system haying H(z) = -I -2
(18)
l+alz +a2z

4. (a) Consider the following causal system: (18)

r->
yi) := J' S-
.~ :2- nib
f4..C~)
Contd P/2
 =2=

EEE 311
Contd ... Q. No. 4(a) ?

(i) Determine the system function and impulse response ifH(z)lz=' = 1

(ii) Is the system stable?

(iii) Sketch the possible implementation of the system and write corresponding
difference equation.
(b) Define different forms of realization structure and describe the steps of converting

one form to other. With justification, draw both forms for an FIR MA system. (17)

SECTION -B
There are FOUR questions in this Section. Answer any THREE.
_ ex>
5. (a) Consider x[n]= Lx[n-rN], where a finite-length signal x(n] = 0 except
r=-<x>

0::::;n::::; N -1. show that N equally spaced samples of X(eJ(j)), Fourier transform of x(n],

can be expressed as . i[k]~ discrete Fourier series coefficients of :X[n], Le., (10)

(b)

x,[n]=[~a be d.e 0 oJ
x2[n]= [1e 0 0 0 abc J
The above two eight-point sequences have discrete Fourier transforms X,[k]and X2[k],
respectively. Determine the relat.ionship between Xl [k] and X 2 [k]. (7)

x[nl=cos(n7r), n=O, 1,2,3

(c) 2 (18)
h[n]~2n n=O, 1, 2, 3
(i) Calculate X(k], four point DFT ofx(n]
(ii) Calculate H(k], four point DFT ofh(n]
(iii) Calculate y[n] = x(n] @ h(n] (four point circular convolution)
(iv) Also calculate y(n] using IDFT.

6. (a) Derive the general expression of frequency response H(eiW) for Type II FIR linear

phase filter. (8)

(b) A causal linear time-invariant discrete-time system has following system function. (15)

Contd .; P/3
 =3=

EEE 311
Contd ... O. No. 6(b)
(i) Find expressions for a minimum-phase system H)(z) and an all-pass system Hap(z);
such that H(z) = H)(z) Hap(z)
(ii) Between two group delays corresponding to H1(z) and H(z), which one is lower

and why?
(iii) Find expressions for a different minimum-phase system H2(z) and a generalized
linear-phase FIR system Hlin(Z),such that H(z) = H2(Z) Hlin(Z),
1 .-1
1+- z
(c) H(z) = 2 (12)
. (1- 2 z-1{1 --e 4 -}; z-I)
4 }; z-IJ(1 --e
33.

(i) Sketch the pole-zero plot of c(z) = H(z)H*( (X. )

(ii) Sketch the pole-zero plot of all possible systems which are both causal and stable

and have the same c(z) found in (i).

(iii) Sketch the pole-zero plot of a minimum phase system which has the same c(z) I
~-

found in (i) and also sketch the magnitude spectrum of this system.

7. (a) Explain with necessary graphical representations the reason behind the generation of
ripples in the resulting filter when a rectangular window is used in windowing method of
FIR filter design. Mention the advantages of using Hamming window instead of using

rectangular window . (10)

. 1- 0.01 ~ IH(ej/l) 1~ 1 +0.01, 0 ~ Iml ~ 0.4Jl'

(b) IH(ej/l) ) $; 0.03 O.5Jl' $; Iml $; 0.65Jl'

0.250 - O.005 $; IH(ej/l) ) $; 0.250 + 0.005, 0.7 Jl' $; Iml $; Jl'

A multi-band FIR filter is to be designed using the Kaiser window method based on the

above specifications. Find (i) Filter length, (ii)

;
P for Kaiser window,
..
(iii) impulse
response of the desired filter, (iv) expression of Kaiser window function w(n] and

(v) impulse response of the designed filter. (25)

8. (a) What is the main problem of impulse invariance method of IIR filter design? Show
that by using bi-linear transform method of IIR filter design, causal stable continuous

type filter can map into causal stable discrete time filter.

eontd P/4

-\
 '.'

;. =4=
EEE 311 .'
Contd ... O. No.8
(b) "

Consider that xc(t) is bank-limited signal with a band limit of 2 kHz. x[n) is generated

from xc(t) using sampling frequency Fs = 24 K samples/sec and then S[11] is obtained by
.;.

down sampling x[n] with a factor 2. Design discrete-time IIR filter using bi-linear
transform method using the following specifications.

O.75 ~ IH(ejaJ
) ~ 1, 0 ~ Iml ~ OJ p

IH(ej(i) ~~ 0.25, ms ~ IOJI ~ 1r

Select illp such that the frequency characteristic of xc(t) 'is preserved and use
, .
\
.l1w = ills - illp = 0.37t. (25)
i
L-3/T-2/EEE Date: 12/05/2014
BANGLADESH UNIVERSITY OF ENGINEERING AND TECHNOLOGY, DHAKA
L-3/T-2 B. Sc. Engineering Examinations 2011-2OJ2

. Sub: EEE 311 (Digital Signal Processing I)

Full Marks: 210 Time: 3 Hours

USE SEP ARA TE SCRIPTS FOR EACH SECTION
The figures in the margin indicate full marks .
.._ _ __ ._ __ _ , __ _ _ _.,,_.,..,.._ ._M __ . .._~. ._ ..
m •••• _ •••• " ••• ' •• __ ••••• _ •••••••••• _ ••••••• -_ ••••••• __ ••••••••••• _-_._ .•••••••••••••••• __ •••

'" ;~.

SECTION-A
There are FOUR questions in this section. Answer any THREE.

1. (a)A continuous time signal xe(t)==COS(2ITFJt)+COS(27TFi) is sampled at Fs

samples/sec and resulting discrete-time signal is x[n] == A cos(27if1'z).Find the value of Fs

that will provide f corresponding to F I and in this case also find the value of A. Consider

FI == 2000 Hz and F2 == 8000 Hz. (10)

(b) Consider the discrete-time system as shown in ~ig. for Q. No. 1(b). (10)

LTI
s1stuYI-1

The first system is described by

I (l) 1< 0.5 IT
0.5 IT :::;1w 1< IT.

The second system is described by y[n]=~{n]-v[n-2]. The input to this system is

x[n] = cos(O.8nn)+o(n -3]+ 5. Determine the output y[n].

(c) Derive the condition of causality of an LTI system. (7)
(d) Determine whether the system T {x(n]} == ex[n"3] is (i) causal, (ii) stable, and (iii) LTI. :(8)

2. (a) (12)

-2..

"" ..-
tor\
-••.......... ~ ,

Contd P/2
 =2=

EEE 311
Contd ... Q. No. 2(a).

y[n]. (10)

a.,,~ [H (eiw) J
,
J.:"

-1L

Contd P/3

.' ..,:
 =3=

EEE 311
Contd ... Q. No. 2(c)

Consider the Z-transform shown in Fig. for Q. No. 2(c) corresponding to a system
function H(z). (i) Mention all possible ROCs, (ii) For each ROC, comment on causality
and stability (iii) Write the expression of causal-stable h(n] (if any).

3. (a) Consider the following system (Fig. for Q. No. 3(a» (18)

~[nJ

,., 0" .~ ••• ,_ ••••• _ •••• _~;

T': 2.T
__ ., ",.'_ ~,_ •• ,.

In the. above system,

and IXcCjn~ 1s shown in the figure. Given that the original sampling rate Os

(corresponding to T in the .figure) is three times of ON (Le. Os = 30N)' Sketch

. (i) Ix(e jOJ

1, (ii) IW(e 1, (iii) IV(e 1, and (iv) 1';; (jn)
jOJ jOJ

(b) When the input to an LTI system is (17)

the corresponding output is

yIn] = sGJ uln]-s(~ J uln]

Using Z-transform method, find the impulse response h(n] of the system and determine

whether or not the system is stable and/or causal.

Contd P/4
 =4=

EEE 311

4. (a) Draw the signal flow graph for a causal FIR system with impulse response (10)

h[n] = ~ 1 - 2 5 -2 1 2 J

Note that your design must ensure minimum number of multiplier.

(b) In a decimator systerh what type of filter" is generally used and why? With a
simplified block diagram explain the process of changing the sampling rate of a signal
•
x[n] from Fs to 0.8 Fs. (10)
(c) Find the step response of an initially-relaxed causal LTI system described by
following difference equation using time-domain technique . (15)
y[n] = 5y[n -1]- 6y[n - 2]+ 2x[n -1].

SECTION -B
There are FOUR questions in this section. Answer any THREE.

5. (a) The transfer function H(z) of a causal LTI system has the pole-zero configuration
shown in Fig. fpr Q. 5(a). It is also known that H(z) = 6 when z = 1. (12)

Determine the response of the system to the input signal x(n), obtained from sampling
the continuous signal .
x(t) = 50 + lOcos (2 0tlt )+ 30cos (407lt )
at a sampling frequency Fs = 40 Hz.
(b) Consider a causal sequence x(n) with the z-transform (11)

Find the values of a so that anx(n) .be a real, minimum-phase sequence.

Contd P/5
 =5=

EEE 311
Contd •.. Q. No.5

(c) Suppose that a stable nonminimum-phase LTI system, H(z) is cascaded with a

compensating system Hc(z) as shown in Fig. for Q. S(c) . (12)

..,,-----(;;jZ} - .
i>----------j
xf--n)---:-7\ \-Ie~\ w(-<i) fc(2) 1:
'1_.:- - - - -l.

r~.-jo ~ sec)
(i) How should Hc(z) be chosen so that it is stable and causal, and the magnitude
of the'ovenin effective frequency response is unity?
(ii) Assume that
H(z) = (1- O.8ejO,31l' z,-l )(1- O.8e-jO,31l' z-l Xl-1.2e jO,71l' Z-l )(1_1.2e-jo. 7
1l' z-l )

Find Hc(z) and G(z) for this case.

6. (a) Fig. for Q. 6(a) shows a sequence x(n) for which the value of x(3) is an unknown

constant c. Let (11)

Xl (k)= X(k )ej21l'3k/5,

where X(k) is the five-point, DFT ofx(n). The sequence xl(n) plotted in Fig. for Q. 6(a)
is the inverse DFT OfXl(k). What is the value ofc?

2 "~1'~2. z.,C-r\) ..~

'.-...
_'
...
_"-'.'
_._'1.L\'~'~' ...•
~ -' ,. -.... .. _
.,.- " - _ .

-, 0 1 s~6.,..---n-*
2.-,..1-_ ..•;•....... ..

tb) Consider the finite-length sequence (13)

x(n) = 2o(n)+o(n-l)+o(~-3).
(i) Compute the four-pointDFT X(k)

(ii) Compute the four-point inverse DFT of Y(k)= X2 (k) to obtain a sequence yen)

forn=O,I,2,3.
(iii) If N-point DFTs are used in the previous two steps, how should you choose N

so that y(n) = x(n)* x(n) for 0 $ n $ N -1 ?

Contd : P/6
 =6=

EEE 311
Contd •.. Q. No.6

(c) An input signal x(n) has been distorted by an FIR system with the impulse response (11)

h(n) = o(n)-.lo(n - no)

-; 2
where no is a positive integer. We wish to recover the original signal x(n) by processing

the distorted output. Suppose that we use an FIR filter of length N in an attempt to
implement the inverse filter, and let the N-point DFT of the FIR filter be
1
G(k) = H(k)' k = 0, 1, ..., N-1.

It might appear that the FIR filter d(k) = ~(k) implements the inverse filte~ perfectly.

Briefly explain the fallacy in this argument.

7. (a) Derive the necessary conditions for a realizable digital filter to have a linear phase

characteristic. Also explain the advantages of filters with such a characteristic. (12)
(b) A highpass linear phase FIR filter is characterized by the following impulse response

coefficients: (8)
h(n) = {0.0575,-0.0544,-02955,0.5673,-02955,-0.0554,0.0575}
Determine the coefficients of an equivalent lowpass filter.

(c) A linear phase bandpass digital filter is required to meet the following specification: (15)
passband 12-16 kHz
transition width 3 kHz
sampling frequency 48 kHz
passband ripple 0.01 dB
stopband attenuation >50 dB
Assume that the coefficients of the filter are to be calculated using window method.
Determine the following issues for the filter:
(i) type of window
(ii) the number of filter coefficients
(iii) the expression of ideal bandpass filter hD(n) for -a:::<n<oc

(iv) the coefficients of the required bandpass filter h(n) = hD(n)a>(n) for -3 ~ n~3
(v) the necessary modification for causal -implementation of the obtained filter
h(n).

Contd P17
..•

=7=

EEE 311

8. (a) An analog signal is contaminated by a 50 Hz signal and its harmonic at 150 Hz.
Assume that the contaminated signal is sampled at 1 kHz. Find the transfer fUnction of a

simple digital filter to remove the interference and its harmonic.

(b) A requirement exists to simulate, in a digital computer, an analog filter with the

following transfer functiob:

(10)

H(s) = 0.5
s(s + 0.5)
Obtain the transfer function 'of the digital filter using the impulse i~variant method .

. Assume a sampling frequency of 1 (normalized).

(c) The specification of a certain highpass filter is given below: (15)

passband' 125-250 Hz
stopband 0-50 Hz
passband ripple 3 dB
stopband attenuation 20 dB
sampling frequency 500 Hz
Assuming that the filter has a Butterworth response, determine the transfer function H(z)

, of the filter using the bilinear z-transform.

-----------------------------------

, ,
..

.:j
L-3/T-2/EEE Date: 18/07/2011
BANGLADESH UNIVERSITY OF ENGINEERING AND TECHNOLOGY, DHAKA
L-3/T-2 B. Sc. Engineering Examinations 2009-2010

Sub: EEE 311 (Digital Signal Processing I)

Full Marks: 210 Time: 3 Hours
USE SEP ARA TE SCRIPTS FOR EACH SECTION

SECTION -A
There are FOUR questions in this Section. Answer any THREE.
If otherwise sated, symbols have their usual meanings.
The figures in the margin indicate full marks.

1. (a) The impulse response of a discrete-time LTI system is given by

I ifn=O
hen) = -1 if n = 1
{
. 0 otherwise

(i) let x(n) and yen) be the input and output, respectively of this system. Find the real
numbers a, p, and y such that yen) = ax(n+ 1) + P x(n) + yx(n -1) (5)
(ii) Is the system causal? Justify your answer. (3)
(iii) Find the output yen), when the input of the system is given by, (7)

I if n=-I
x(n) = 1 if n = 0
{
o otherwise
(b) There are two kinds of particles inside a nuclear reactor. In every second an a particle
splits into eight p particles and at the same time a p particle splits into an a particle and
two p particles.
(i) Determine the difference equations that describe the number of particles in the
reactor at a given second. (10)
(ii) If there is a single a particle in the reactor at time n = 0 second, how many paticles
are there altogether at time n = 100 second? (10)

2. (a) Design a causal discrete-time LTI system such that if the input IS

x(n) = (1'2)n u(n) -

( 4'1)( '21 )n-l u(n -1) then the output is yen) = (1'3)n u(n).

(i) Determine the impulse response hen) and the system function H(z). (7)
(ii) Sketch the pole-zero plots and ROC ofH(z), if the system is stable. (4)
(iii) Find the difference equation that characterizes this system. (3)
(iv) Realize the system with a minimum requirement of memory. (6)

Contd P/2
 =2=
EEE 311
(b) Consider the system shown in Fig. for Q. 2(b) for processing a continuous-time signal
with a discrete-time system. (15)

r I r I
I
----_
i
.. --_.
•..
--'X-tY\~-
~ H (cu) .
~H'Y\~-'-
1>/A
.. i
I

l' Y
..-
AID y
---
y
--~o.('ty\
I

\ XaH)--
j -
i .l1";
"''''

'''_j~
__Ts I \
;

The frequency response of the discrete-time system is

The sampling frequencies of the AID and DIA converters are 2 kHz and 5 kHz,

respectively. Find the output y,( t) when the input signal x, (t) ~ Sin( IOOO1lt + : )-

3. (a) The system function of a communication channel is given by (10)

H(z) = (1- O.gej0.41l z-J »- O.ge- jO.41l z-J 1-1.5e t jO.61l z-'J )(1-1.5e- jO.61l z~J)

Determine the system function Hc(z) of a causal and stable compensating systm so that
the cascaded interconnection of the two systems has a flat magnitude response.
(b) The discrete time system shown in Fig. for Q. 3(b) is characterized by the following
input-output relationshiQs: _
= ---
£Jffl)l

':-". i
I
----------.:...------------- .. 1
_____ ._ h~.:.-::ful' f).,. 3(k:» i
===-------~-----------:--------~
System 1: yen) = x(n -1) + x(n) + x(n+ 1)
System 2: yen) = v(n -1) + yen) + v(n+ 1)
k=oo
Consider that the input to this system is the periodic signal x(n) = Lo(n - 6k)
k=-oo

(i) Determine the Fourier series coefficients ofx(n) denoted as ak for all k. (5)
(ii) Determine and plot several periods ofy(n) and wen). (8)
(iii) Determine the Fourier series coefficients ofy(n) denoted as bk for all k. (8)
(iv) Determine the Fourier series coefficients of wen) denoted as Ck in terms of bk for all
k. (4)
Contd P/3
 =3=
EEE 311

4. (a) Let Xm(ro) (m = 1,2,3) be the DTFT of the discrete-time signal xm(n). Ifx3(n) = x[(n)
1 Jr
x2(n), then prove that X3 (OJ) =- IX[ (A)X2 (OJ - A)dA (7)
27T -Jr

(b) Consider a discrete-time LTI system described by the following difference equation:

yen) = -x(n + 1) + 2x(n) -x(n -1)

(i) Find and plot the magnitude of the frequency response function, i.e., IH(ro)l. (8)
(ii) A discrete-time input signal x(n) is obtained by sampling an analog signal (8)

xa(t) = 1 + cos 3t + 2 cos( 6t+ :)witha sampling period Ts =:. Does any aliasing

happen? Plot the magnitude spectrum of the DTFT of the sampled signal x(n) denoted

as IX( ro)1.
(iii) Plot the magnitude spectrum of the corresponding sampled output signal yen)
denoted as IY(ro)l. (6)
n=oo
(iv) Determine the numerical value of ~]y(n)12 (6)
n=-oo

SECTION-B
There are FOUR questions in this Section. Answer any THREE.
The questions are of equal value.

5. (a) An LTI system with frequency response H(ro) is excited with the periodic input

00

x(n) = I.t5(n - kN)

k=-oo

IfY(k) is the N-point DFT ofy(n), 0 =:;n =:;N -1, then find the relation between Y(k) and
H(ro).
(b) Describe overlep-add method of filtering of long data sequence using DFT. Discuss
the implications of finite data record in frequency analysis using DFT. '
6. (a) Define DFT of an arbitrary sequence x(n) and find its relations with Fourier series
coefficients and also with Z-transform.
(b) Show that circular convolution between 2-sequences is equivalent to multiplication of
their DFTs. Find N-point circular convolution of the sequences,

. 27T 27T
x[ ()
n =Slll-n and x2 (n) = cos-n for 0 =:;n =:;N -1
N N

Contd P/4
•

=4=
EEE 311
7. (a) Discuss windowing in calculating FIR filter coefficients. Using window method, find
the first 5 coefficients of an FIR BPF to meet the following specifications:
Pass band: 150 - 250 Hz
Transition width: 50 Hz
Stop band attenuation: 60 dB
Sampling frequency: 1 kHz

(b) Describe the alternation theorem to determine FIR fiiter coefficients by optimum
equiripple method.

8. (a) "Analog filters can be transformed into IIR type more easily than into FIR ones"-

Explain. Convert a simple analog filter with H(S) = ~ into equivalent IIR digital
S-p
filter by impulse invariant method.

(b) Deduce the relations between s - and Z - parameters to convert an analog filter into its
equivalent IIR digital filter by bilinear Z transformation method. Convert the analog filter
1
with transfer function H (s) = s + O~ into digital IIR filter using BZT method.
(s+O.I) +16

The digital filter is to have a resonant frequency of (Or = n/2