[6] Ex/EE/T/322/79/2010

BACHELOR O F ELECTRICAL ENGINEERING EXAMINATION, 2010

8. a) Find an expression for actual frequency response of a causal
( 3rd Year, 2nd Semester )
FIR digital filter having a symmetric and real impulse
sequence of length M. Assume M to be an odd number. 14 DIGITAL SIGNAL P ROCESSING
Time : Three hours Full Marks : 100
b) The filter coefficients of a 5-tap causal FIR filter is given by
( 50 marks for each Part )
h l = 04 for 0 l 4 .
Use a separate Answer-Script for each part.
i) Find an expression for the frequency response of the
PART - I
filter.
Answer any three questions.
ii) Find the gain of the filter at zero frequency. Two marks reserved for neatness.

iii) Find the gain of the filter at a frequency ws /5, where w s 1. a) Two non-interacting DTLTI systems with impulse
is the sampling frequency. 5+3+3 responses g[n] and h[n] are in cascade. If [n] is the
impulse response of the combination, show that
9. Write notes on any two of the following : 12 12 2
+ + +
( j) [n ] = ( j) g n ( j) h n
n n n
a) Direct realisation of linear-phase FIR digital filters. 4
n = n = n =
b) Contrast enhancement of gray scale images. b) The center of gravity of a sequence x[n] is defined as

c) FIR digital filters for off-line applications. +

nx [ n ]
Cg = n =
+
x [n ]
n =

If X(z) is the z-transform of x[n], derive an expression for

Cg in terms of X(z). 4

c) Fig [A] shows fragments of the amplitude spectrum of a

real discrete-time signal over two separate ranges of
frequency. Draw, with necessary explanations, the
amplitude spectrum over to + radians. 8
[ Turn over
[2] [5]

b) The impulse-invariant transformation is not suitable for

discretizing highpass and bandpass filters. Justify or
correct the above statement with appropriate reasoning. 4

5. Write short notes on any two of the following : 8+8

a) Distortions introduced by practical DACs and their

remedies.

b) Modeling uniform sampling as impulse modulation.

c) FIR and IIR discrete-time LTI systems.

d) Transposed direct form structure for IIR systems.

2. a) The pole-zero diagram in Fig. [B] corresponds to the z-
transform X(z) of a causal sequence x[n]. Sketch the pole-
zero diagram of Y(z), where y [ n ] = x [ n + 3] . Also specify
PART - II

the ROC for Y(z). 8 Answer any two questions.

6. a) Develop an expression for IDFI from DFT. 12

b) Derive the DFT of the sequence x K = [1, 1, 2, 2, 3, 3] and

determine the corresponding amplitude spectrum. 8

c) Compute IDFT of the sequence

X n = 3, ( 2 + j) , 1, ( 2 j) 5

7. a) Develop a radix-z decimation-in-frequency FFT algorithm

from the DFT. Draw the signal flow graph. 20

b) Compare computational loads of DFT and FFT. 5

[ Turn over
[4] [3]

c) Realize the system shown in Fig. [C] using only one unit b) Consider a linear shift-invariant discrete-time system whose
delay element. 7 response to a unit step is
n n
1 1
s [ n ] = u [ n ] + u [n ] + u [ n ]

3 4

Determine the following. 8

i) Difference equation relating the output y[n] and the

input x[n] of the system.

ii) Impulse response of the system, without using any

transform.

iii) Whether or not the system is stable.

3. a) The z-transform of a sequence x[n] is

4. a) Derive the difference equation relating the output and the z 3 2z

X ( z) =
input sequences of a digital lowpass Butterworth filter with z2

the following specifications: Find the possible ROC(s) of X(z). 3

l A DC gain of 0 dB. b) Determine the z-transorms of the following sequence, if it

l The gain should be no less than 2 dB over 0 to 10 exists. Indicate the ROCs and show the location of poles

radians/second, and no greater than 11 dB for and zeros in the z-plane.

frequencies above 15 radians / second. x [ n ] = a n u [ n ] + bn u [ n ] + cn u [ n 1]

l The highest frequency to be processed is 35 radians/ where,
second.
1 1 1
i) a = , b= , c=
Use bilinear transformation prceeded by frequency 4 3 7
prewarping. 1 1 3
ii) a = , b = , c = 6
Derive the frequency warping relation used. 12 8 5 4

[ Turn over