BE EXTC D T S P MAY- 2007

By Kiran Talele ( talelesir@yahoo.com )

Q1. (a) Let h[n] be the unit impulse response of a Low Pass Filter with a cutoff
frequency wc, what type of filter has a unit sample response
g[n] = (–1)n h[n]. ? [4]
Solution :
Consider a Low Pass Filter with h[n]=(0.5)n u[n]
z
Then G ( z ) = POLE = 0.5
z − 0.5
g[n] = (–1)n h[n]
g[n] = (-1)n (0.5)n u[n] = (–0.5)n u[n]
z
By ZT, G ( z ) = POLE = – 0.5
z + 0.5
(i) At w=0, z=1
|H(w)|
|H(w)|=0.66 2
(ii) At w=π, z=–1
|H(w)|=2 0.66
0 π

That means g[n] is impulse response of High Pass Filter. ANS

--------------------------------------------------------------------------------------------------------
(b) Assume that a complex multiply takes 1 micro sec and that the amount of
time to compute a DFT is determined by the amount of time to perform
all the multiplications. [8]
(i) How much time does it take to compute a 1024 point DFT directly?
(ii) How much time is required if FFT is used. ?
Solution :
(i) By DFT
Total Complex multiplications = N2 =(1024)2
Time required to compute N2 Multiplications =(1024)2micro sec =1.048576 Sec
(ii) By FFT
Total Complex multiplications = N/2 log2N=(512) log2(1024) = 5120
Time required to compute 5120 Multiplications =(5120)micro sec =0.005120 Sec
--------------------------------------------------------------------------------------------------------
(c) Sequence Xp(n) is periodic repetition of sequence x[n]. What is the
relationship between Ck of Discrete Time Fourier Series of xp[n] and DFT
X[k] of x[n]. [8]
Solution :

talelesir@yahoo.com : 9987030881
☯
 B E EXTC D T S P MAY- 2007 2

Relationship between DTFS and DFT

The fourier series representation of a periodic sequence xp[n] with fundamental period
N is given by,
N −1 2πnk

∑ Ck e
j
xp[n] = N
, − ∞≤n ≤∞
k =0

Where the fourier series coefficients are given by,

− j 2πnk
1
Ck =
N
∑ x p [ n ] e N
, 0 ≤ k ≤ n −1
N −1 N −1 − j 2πnk
By DFT, X [k ] = ∑ x [n] WNnk =
n =0
∑ x [ n] e
n=0
N

By comparing the above equations, If x[n] = xp[n] Then X [k] = N.Ck.

--------------------------------------------------------------------------------------------------------
Q2. (a) A digital filter that is implemented on a DSP chip is described by the
linear constant coefficient difference equation.
y[n] = ¾ y[n–1] – 1/8 y[n–2] + x[n].
In evaluating the performance of the filter, the unit sample response is
measured. The internal storage registers on the chip, however, are not set
to zero prior to applying the input.
Therefore, the output of the filter contains the effect of the initial
conditions, which are y[–1] = –1 and y[–2] = 1.
Determine the responses of the filter for all n ≥ 0 and compare it with the
zero state response. [8]
Solution :

Now, y[n] = ¾ y[n–1] – 1/8 y[n–2] + x[n]

By ZT, Y ( z ) = 34 [z −1Y ( z ) + y (−1)]− 18 [z −2Y ( z ) + z −1 y (−1) + y (−2)]+ X ( z )
Put y[–1] = –1 and y[–2] = 1.
[ ] [
Y ( z ) = 34 z −1Y ( z ) + y (−1) − 18 z −2Y ( z ) + z −1 y (−1) + y (−2) + X ( z ) ]
Y ( z ) = 34 [z −1
] [ ]
Y ( z ) − 1 − 18 z −2Y ( z ) − z −1 + 1 + X ( z )
Y ( z ) = 34 z −1Y ( z ) − 34 − 18 z −2Y ( z ) + 18 z −1 − 18 + X ( z )
Y ( z )(1 − 34 z −1 + 18 z −2 ) = (− 34 + 18 z −1 − 18 ) + X ( z )
(− 78 + 18 z −1 ) X (z)
Y ( z) = +
(1 − 4 z + 8 z ) (1 − 4 z + 18 z − 2 )
3 −1 1 −2 3 −1

Let Y(z)=Yzi(z)+Yzs(z)

talelesir@yahoo.com : 9987030881
☯
 B E EXTC D T S P MAY- 2007 3

(i) To find ZIR

(− 87 + 18 z −1 ) − 78 z 2 + 18 z
Yzi ( z ) = =
(1 − 34 z −1 + 18 z − 2 ) z 2 − 34 z + 18
Yzi ( z ) − 87 z + 18
=
z ( z − 14 )( z − 12 )
Y ( z) A B
By PFE, zi = +
z z − 14 z − 12
Where
Yzi ( z )
( z − 14 )
(i) A = z
z= 1 == [ ] 3
8
4

Yzi ( z )
( z − 12 )
(ii) B = z
z= 1 == [ −5
4
]
2

3⎡ z ⎤ 5 ⎡ z ⎤
Yzi ( z ) = ⎢ ⎥ − ⎢ ⎥
8 ⎣ z − 14 ⎦ 4 ⎣ z − 12 ⎦
n n
3⎛ 1 ⎞ 5 ⎛1⎞
By IZT, y zi [n] = ⎜ ⎟ u[n] − ⎜ ⎟ u[n]
8⎝ 4⎠ 4 ⎝2⎠

(ii) To find ZSR

X ( z) ⎛ z2 ⎞
Yzs ( z ) = = ⎜ ⎟ X ( z)
(1 − 34 z + 18 z ) ⎜⎝ z − 34 z + 18 ⎟⎠
−1 − 2 2

For unit sample response x[n] = δ[n]

∴ X(z)= 1
z2
Yzs ( z ) ==
(z − 14 ) (z − 12 )
Yzs ( z ) z
=
z ( z − 4 )( z − 12 )
1

Y ( z) A B
By PFE, zs = +
z z − 4 z − 12
1

Where
Yzs ( z )
( z − 14 )
(i) A = z z= 1 == [ −1 ]
4
Yzs ( z )
( z − 12 )
(ii) B = z
z= 1 == [ 2 ]
2

⎡ z ⎤ ⎡ z ⎤
Yzs ( z ) = (− 1) ⎢ ⎥ + 2 ⎢ 1⎥
⎣z − 4 ⎦ ⎣z − 2 ⎦
1

n n
⎛1⎞ ⎛1⎞
By IZT, y zs [n] = − ⎜ ⎟ u[n] + 2 ⎜ ⎟ u[n]
⎝4⎠ ⎝2⎠

talelesir@yahoo.com : 9987030881
☯
 B E EXTC D T S P MAY- 2007 4

(iii) To find y[n] = zir + zsr

n n n n
3⎛ 1 ⎞ 5 ⎛1⎞ ⎛1⎞ ⎛1⎞
y[n] = ⎜ ⎟ u[n] − ⎜ ⎟ u[n] − ⎜ ⎟ u[n] + 2 ⎜ ⎟ u[n]
8⎝ 4⎠ 4 ⎝2⎠ ⎝4⎠ ⎝2⎠

−5⎛1⎞
n n
3 ⎛1⎞
y[n] = ⎜ ⎟ u[n] + ⎜ ⎟ u[n] ANS
8 ⎝4⎠ 4 ⎝2⎠
--------------------------------------------------------------------------------------------------------
(b) For following linear shift invariant systems, draw pole zero diagram and
identify the filter type based on pass band. [12]
Z −1 − a * 1 − z −4
(i) H ( z ) = where a < 1 (ii) H ( z ) = A
1 − a z −1 1 + a 4 z −4

Solution : Q2. (b) (i)

Z −1 − a *
H ( z) = where a < 1
1 − a z −1
1 − a* z
H ( z) = Let a = 0.5
z−a
1 − 0.5* z 1 − 0.5 z
H ( z) = =
z − 0.5 z − 0.5

[1] POLE-ZERO diagram Imaginary

1 − 0.5 z
H ( z ) ==
z − 0.5
ZEROS:
Real
Z1 = 2
0.5 2
POLES :
P1= 0.5

[2] Magnitude Spectrum

1 − 0.5 z
H ( z ) ==
z − 0.5
1. At w=0, z=1 |H(w)|
1
|H(w)|=1 Filter is ALL Pass Filter
2. At w=π, z=–1
|H(w)|=1 0 π

talelesir@yahoo.com : 9987030881
☯
 B E EXTC D T S P MAY- 2007 5

Solution : Q2. (b) (ii)

1 − z −4
H ( z) = A
1 + a 4 z −4
z4 −1
H ( z) = A Let a =0.5
z 4 + a4

[1] POLE-ZERO diagram

z4 −1
H ( z) = A
z 4 + (0.5) 4

ZEROS POLES
z −1 = 0 z + (0.5) = 0
4 4 4
Imaginary
z =1 z 4 = −(0.5)
4 4

z 4 = e j 2πk z 4 = (0.5) e jπ e j 2πk

4

j π2k jπ ( 2 k +1 )
Real
zk = e zk = (0.5)e 4
k=0 , z0 =1 k=0 , z0 = 0.5e
j π4

j π2
k=1, z1 = e k=1, z1 = 0.5e
j 34π

k=2, z2 = e jπ 5π
k=2, z2 = 0.5e j 4

j 32π
k=3, z3 = e j 74π
k=3, z3 = 0.5e

[2] Magnitude Spectrum

z4 −1
H ( z) = A 4 Let A =1 Put z = ejw
z + (0.5) 4

H ( w) =
e j 4w − 1
==
[cos(4w) − 1] + j [sin 4w)]
e + 0.0625
j 4w
[cos(4w) + 0.0625] + j [sin(4w)]

w |H(w)| φ
0 0 0 |H(w)|
0.2π 2.002 0.35 1
0.4π 1.151 -1.000
0.6π 1.151 1.000
0.8π 2.002 -0.353 0 π
π 0 0 Filter is

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

talelesir@yahoo.com : 9987030881
☯
 B E EXTC D T S P MAY- 2007 6

Q3. (a) Consider the following specifications for a Low Pass Filter.
0.99 ≤ ⏐H(e j w) ⏐ ≤ 1.01 for 0 ≤ w ≤ 0.3 π
⏐H(e j w) ⏐ ≤ 0.01 for 0.35 π ≤ w ≤ π

Design a Linear Phase FIR filter to meet these specifications using the
window design method. [10]
Solution :

20log(1.01)
1.01 20log(0.99)
0.99

20log(0.01)
0.01

0 0.5π 0.75π
0 0.3π
0.5 0.35π
0.75π π

(i) Ap = 20log(1.01) – 20log(0.99) == 0.1737 dB

(ii) As = 20log(1.01) – 20log(0.01) == 40.086 dB
(iii) wp = 0.3π (iv) ws = 0.35π (v) Fs = 1 Hz (vi) LPF

(1) Select window function

For Hamming window function As= 53 dB > 40.086 dB(required value)
∴ Select Hamming window function.
(2) Calculate N
C
N ≥
f 2 − f1

(i) For hamming window function C=3.14

(ii) f2 – f1 = fs - fp
wp = 0.3π = 2π fp ∴ fp = 0.15
ws = 0.35π = 2π fp ∴ fs = 0.175

3.14
N≥
0.175 − 0.15
N ≥125.6
Let N = 127 (odd)

talelesir@yahoo.com : 9987030881
☯
 B E EXTC D T S P MAY- 2007 7

(3) Calculate wc
wp + ws
wc = == 0.325π rad
2
--------------------------------------------------------
Step 1. Find Hd(w)
Let Hd(w) = | Hd(w) | φ(w)

Where (i) Magnitude response :

1
⎧ 1 for − wc ≤ ω ≤ wc
Hd(ω) = ⎨
-π -wc 0 wc π ⎩ 0 otherwise

(ii) Phase Response : φ(w) = ej φ

For Linear Phase LPF with symmetric h[n]

φ= − ( N2−1 ) w = – αw

∴φ(w) = e−j α w

⎧⎪ e − j α w − wc ≤ ω ≤ wc
By substituting, H d ( w) = ⎨
⎪⎩ 0 otherwise
where α = 63 and wc = 0.325π

Step 2. Find hd[n]

π
1
By Inverse DTFT, hd [n] =
2π ∫ H ( w)e jnw dw
−π
wc
1
hd [n] =
2π ∫
− wc
(e − j α w )e jnw dw

wc
1
hd [n] =
2π ∫
− wc
e j ( n −α ) w dw

wc
1 ⎡ e j ( n −α ) w ⎤
hd [n] = ⎢ (n − α ) j ⎥
2π ⎣ ⎦ − wc

talelesir@yahoo.com : 9987030881
☯
 B E EXTC D T S P MAY- 2007 8

1 ⎡ e j ( n − α ) w c − e − j ( n − α ) wc ⎤
hd [ n ] = ⎢ ⎥
π (n − α ) ⎣ 2j ⎦

w c sin( n − α ) w
h d [n ] = c
π (n − α )w c where α =63 and wc = 0.325π

------------------------------------------------------------------------------------------
Step 3. Find h[n]
Linear phase FIR filter with impulse response h[n] is given by,

h[n] = hd[n] w[n]

⎡ ⎛ sin( n − 63 ) 0 . 325 π ⎞⎤ ⎡ ⎛ 2π n ⎞ ⎤ for 0 ≤ n ≤ 126
h [ n ] = ⎢ 0 . 325 ⎜⎜ ⎟⎟ ⎥ ⎢ 0 . 54 − 0 . 46 cos ⎜ 126 ⎟ ⎥
⎣ ⎝ ( n − 83 ) 0 . 325 π ⎠⎦ ⎣ ⎝ ⎠⎦
----------------------------------------------------------------------------------------------------
(b) Use the Bilinear Transformation to design a Discrete Time Chebyshev High
Pass filter with an equiripple passband with [10]

0 ≤ ⏐H(e j w) ⏐ ≤ 0.1 for 0 ≤ w ≤ 0.1 π

0.9 ≤ ⏐H(e j w) ⏐ ≤ 1.0 for 0.3 π ≤ w ≤ π
Solution :

1.0 20 log(1.0)
20 log(0.9)
0.9

0.1 20 log(0.1)

0
0 0.1π 0.3π π
0 0.5ππ 0.3π
0.1 0.75π π

(i) Ap = 20 log(1.0) – 20 log(0.9) == 0.9151 dB

(ii) As = 20 log(1.0) – 20 log(0.1) == 20 dB
(iii) wp = 0.3π (iv) ws = 0.1π (v) Fs = 1 Hz (vi) HPF

Step-1 : Design Analog Chebyshev High Pass filter

(1) Calculate Ω p, Ω s

2 ⎛ wp ⎞ 2 ⎛ 0 .3 π ⎞
Ωp = tan ⎜ ⎟= tan ⎜ ⎟ = 1019.05 rad sec
T ⎝ 2 ⎠ 1 ⎝ 2 ⎠
1000
2 ⎛ ws ⎞ 2 ⎛ 0.1π ⎞
Ωs = tan ⎜ ⎟ = tan ⎜ ⎟ = 316.77 rad sec
T ⎝ 2 ⎠ 1 ⎝ 2 ⎠
1000

talelesir@yahoo.com : 9987030881
☯
 B E EXTC D T S P MAY- 2007 9

(2) Calculate Filter order N.

1
⎡ 10 As 10 − 1 ⎤
−1
2
cosh ⎢ 10 Ap / 10 − 1 ⎥
N = ⎣ ⎦
⎡ Ωs ⎤
cosh − 1 ⎢ ⎥
⎣Ωp ⎦
1
⎡ 10 2 − 1 ⎤
−1
2
cosh ⎢ 10 0 . 09151 − 1 ⎥
N = ⎣ ⎦
⎡ 1019 . 05 ⎤
cosh − 1 ⎢
⎣ 316 . 77 ⎥⎦
N=

Let N=

(3) Calculate Normalized HPF

(4) Calculate De-normalized HPF

Step-2 : Design Digital Chebyshev High Pass filter

Ooooooooooooooooooo

--------------------------------------------------------------------------------------------------------
Q4. (a) Consider the sequence x[n] = δ[n] + 2 δ[n-2] + δ[n-3]
(i) Find four point DFT of x[n]. [4]
(ii) If y[n] is four point circular convolution of x[n] with itself, find y[n] and
the four point DFT Y[k]. [4]
(iii) With h[n] = δ[n] + δ[n-1] + 2 δ[n-3] Find four point circular convolution of
x[n] with h[n]. [4]

Solution : Q4. (a) (i) To find four point DFT of x[n]

Given x[n] = δ[n] + 2 δ[n-2] + δ[n-3] = { 1 , 0, 2, 1 }

↑
N −1
By DFT, X[k] = ∑ x[n] WNnk
n= 0

X [ K ] = ⎡ wN
0 0
wN 0
wN wN0 ⎤ ⎡ x[ 0 ] ⎤
⎢ 0 ⎥ ⎢
⎢ wN w1N 2
wN 3
wN ⎥ ⎢ x[1] ⎥⎥
⎢ 0 6 ⎥
⎢ wN wN2 4
wN wN ⎥ ⎢ x[ 2 ] ⎥
⎢ wN
0
w3N 6
w9N ⎥⎦ ⎢ ⎥
⎣ wN
⎣ x[ 3 ] ⎦

talelesir@yahoo.com : 9987030881
☯
 B E EXTC D T S P MAY- 2007 10
⎡ 1 1 1 1 ⎤ ⎡ 1 ⎤
⎢ 1 − j −1 j ⎥ ⎢ 0 ⎥
X [k ] = ⎢ ⎥ ⎢ ⎥
⎢ 1 −1 1 −1 ⎥ ⎢ 2 ⎥
⎢ ⎥ ⎢ ⎥
⎣ 1 j −1 − j ⎦ ⎣ 1 ⎦

⎡ 4 k =0 ⎤
⎢ −1 + j ⎥ ANS
X [k ] = ⎢ ⎥
⎢ 2 ⎥
⎢ ⎥
⎣ −1 − j ⎦

Solution : Q4. (a) (ii) To find Y[k]

y[n] = x[n] ⊗ x[n]
By Convolution property of DFT,
Y[k] = X[k] X[k]
⎡ 4 ⎤ ⎡ 4 ⎤ ⎡ 16 k = 0⎤
⎢ −1 + j ⎥⎥ ⎢⎢ −1 + j ⎥⎥
== ⎢ − 2j ⎥ ANS
Y [k ] = ⎢ ⎢ ⎥
⎢ 2 ⎥ ⎢ 2 ⎥ ⎢ 4 ⎥
⎢ ⎥ ⎢ ⎥ ⎢ ⎥
⎣ −1 − j ⎦ ⎣ −1 − j ⎦ ⎣ 2j ⎦
To find y[n]:
2π
N −1 −j
1
∑ Y [k ] W − nk 1 =e
By Inverse DFT, y[n] = N
where W N N
N k =0

y[ n ] = 1 ⎡ w N0 w N0 w N0 w N0 ⎤ ⎡ Y [0 ] ⎤
⎢ ⎥ ⎢ Y [1 ] ⎥⎥
N 0 −1 −2 −3
⎢ w w w w ⎥
N N N N ⎢
⎢ w 0
w −2
w −4
w −6 ⎥ ⎢ Y [2] ⎥
⎢ ⎥
N N N N
−3 −6 −9 ⎢ ⎥
⎢⎣ ⎥⎦
0
w N w N w N w N ⎣ Y [3] ⎦

⎡ 1 1 1 1 ⎤ ⎡ 16 k = 0⎤
1 ⎢ ⎥ ⎢ − 2 j ⎥
⎢ 1 j −1 − j ⎥ ⎢ ⎥
y[n ] = 4
⎢ 1 −1 1 −1 ⎥ ⎢ 4 ⎥
⎢ ⎥ ⎢ ⎥
⎣ 1 − j −1 j ⎦ ⎣ 2 j ⎦
⎡ 5 n = 0 ⎤
⎢ 4 ⎥ ANS
y[n ] = ⎢ ⎥
⎢ 5 ⎥
⎢ ⎥
⎣ 2 ⎦

---------------------------------------------------------------------------------------------------------------
Solution : Q4. (a) (iii) To find circular convolution of x[n] with h[n].
Given x[n] = δ[n] + 2 δ[n-2] + δ[n-3] = { 1 , 0, 2, 1 } Length L = 4
↑

h[n] = δ[n] + δ[n-1] + 2 δ[n-3] = { 1 , 1, 0, 2 } Length M = 4

↑

y[n] = x[n] ⊗ h[n]

∞
y[n] = ∑ x[m]
m =0
h[((n − m))]

y[ n ] = ⎡1 2 0 1 ⎤ ⎡1 ⎤ = ⎡2 n = 0⎤
⎢1 1 2 0 ⎥ ⎢0 ⎥ ⎢5 ⎥
⎢ ⎥ ⎢ ⎥ ⎢ ⎥ ANS
⎢0 1 1 2 ⎥ ⎢2 ⎥ ⎢4 ⎥
⎢ ⎥ ⎢ ⎥ ⎢ ⎥
⎣2 0 1 1 ⎦ ⎣1 ⎦ ⎣5 ⎦
--------------------------------------------------------------------------------------------------------

talelesir@yahoo.com : 9987030881
☯
 B E EXTC D T S P MAY- 2007 11

(b) Explain the POLE ZERO locations for Type-I, Type-II, Type-III, and Type-IV
Linear Phase FIR filter. [8]
Solution :
Case-1 : When h[n] is Symmetric with N EVEN [ Type-2 Linear Phase Filter ]
⎛1⎞
H ( z ) = z −( N −1) H⎜ ⎟
⎝Z ⎠
⎛ 1 ⎞
H ( z ) = z − ODD H⎜ ⎟
⎝Z ⎠
(1) At Z = –1 w = π (2) At Z = 1 w = 0

H (−1) = (−1) −ODD H (− 1) H (1) = (1) −ODD H (1)

H (−1) = (−1) H (− 1) H (1) = (1) H (1)

H (−1) = − H (− 1)
H (1) = H (1)
So, H(-1) = 0
H ( z ) z = −1 = 0 No definite ZERO exists at z = 1
That means there exists definite ZERO
at z = –1

Case-2 : When h[n] is Symmetric with N ODD [ Type-1 Linear Phase Filter ]
⎛1 ⎞
H ( z ) = z − ( N − 1) H⎜ ⎟
⎝Z ⎠
⎛ 1 ⎞
H (z) = z − EVEN H⎜ ⎟
⎝Z ⎠
(1) At Z = –1 w = π (2) At Z = 1 w = 0

H (−1) = (−1) − EVEN H (− 1) H (−1) = (−1) − EVEN H (− 1)

H (1) = (1) H (1)
H ( −1) = (1) H (− 1) H (1) = H (1)
H (1) = H (1)
No definite ZERO exists at z = –1 No definite ZERO exists at z = 1

Case-3 : When h[n] is Anti-Symmetric with N EVEN [ Type-4 Linear Phase Filter ]
⎛1⎞ ⎛1⎞
H ( z ) = − z − ( N −1) H⎜ ⎟ H ( z ) = − z − ODD H⎜ ⎟
⎝Z⎠ ⎝Z ⎠
(1) At Z = –1 w = π (2) At Z = 1 w = 0
H (1) = − (1) −ODD H (1)
H (−1) = − (−1) −ODD H (− 1)
H (1) = − H (1)
H ( −1) = − ( −1) H (− 1)
H ( −1) = H (− 1)
So, H(1) = 0 H ( z ) z =1 = 0

That means there exists definite ZERO

No definite ZERO exists at z = –1 at z = 1

talelesir@yahoo.com : 9987030881
☯
 B E EXTC D T S P MAY- 2007 12

Case-4 : When h[n] is Anti-Symmetric with N ODD [ Type-3 Linear Phase Filter ]
⎛1⎞ ⎛1⎞
H ( z ) = − z − ( N −1) H⎜ ⎟ H ( z ) = − z − EVEN H⎜ ⎟
⎝Z⎠ ⎝Z⎠

(1) At Z = –1 w = π (2) At Z = 1 w = 0

H (−1) = − (−1) − EVEN H (− 1) H (1) = − (1) − EVEN H (1)

H ( −1) = − H (− 1) H (1) = − H (1)

So, H(-1) = 0 So, H(1) = 0 H ( z ) z =1 = 0

H ( z ) z = −1 = 0 That means there exists definite ZERO
at z = 1
That means there exists definite ZERO
at z = –1
--------------------------------------------------------------------------------------------------------

Q5. (a) The unit sample response of An FIR filter is , [10]

⎧ αn 0 ≤n≤6
h[n] = ⎨
⎩ 0 Otherwise
(i) Draw the direct form implementation of this system.
(ii) Determine system function and use this to draw a flow graph that is cascade
of FIR system with IIR system.
(iii) For both of this implementations, determine the number of multiplications
and additions required to compute each output value and the number of
storage registers that are required.

Solution : Q5. (a)(i) Direct form implementation

⎧ αn 0 ≤n≤6
h[n] = ⎨ x[n] y[n]
⎩ 0 Otherwise
–1
z
By ZT,
∞
H ( z ) = ∑ h[n] z −n α
n= −∞ z–1
6
H ( z ) = ∑ α n z −n z–1
n =0
z–1
( )
6
H ( z) = ∑ α z −1 n
z–1
n =0

1 − (α z ) −1 7 z–1
H ( z) = z–1
1 − α z −1 -α7
1 − α 7 z −7
H ( z) =
1 − α z −1

talelesir@yahoo.com : 9987030881
☯
 B E EXTC D T S P MAY- 2007 13

Solution : Q5. (a)(ii) Cascade form implementation

1 − α 7 z −7
H ( z) =
1 − α z −1

H ( z) = [1−α 7 −7
z ] ⎡⎢ 1 ⎤
⎥
⎣ 1−α z
−1
⎦
Let H ( z ) = H1 ( z ) H 2 ( z )

x[n] y[n]
z–1 z–1

z–1
α
z–1

z–1

z–1

z–1

z–1
-α7

Solution : Q5. (a)(iii)

Sr No Parameter Direct Form Cascade form

1 Number of 2 2
Mltiplications
2 Number of Additions 2 2
3 Number of Delay Blocks 7 7

-------------------------------------------------------------------------------------------------------
(b) Draw a lattice filter implementation for the all POLE filter. [10]
1
H ( z) =
1 − 0.2 z + 0.4 z −2 + 0.6 z −3
−1

And determine the number of multiplications, additions and delays required to

implement the filter. Compare this structure to a direct form realization of H(z)
in terms of multiplies and adds and delays.
Solution :
1 1
H ( z) = −1 −2 −3
== where N= 3
1 − 0.2 z + 0.4 z + 0.6 z AN ( z )
All pole system
For N = 3, A3(z) = 1 – 0.3z–1 + 0.4 z–2 + 0.6 z–3

talelesir@yahoo.com : 9987030881
☯
 B E EXTC D T S P MAY- 2007 14

(i) To find k3
A3(z) = 1 – 0.3z–1 + 0.4 z–2 + 0.6 z–3 ∴ k3 = 0.6

B3(z) = 0.6 + 0.4 z–1 – 0.3 z–2 + z–3

(ii) To find k2
A3 ( z ) − k3 B3 ( z )
A2 ( z ) =
1 − k32

A2 ( z ) =
(1 − 0.3z −1
) (
+ 0.4 z −2 + 0.6 z −3 − (0.6) 0.6 + 0.4 z −1 − 0.3 z −2 + z −3 )
1 − (0.6) 2

A2 ( z ) = 1 − 0.0937 z −1 + 0.625 z −2 ∴ k 2 = 0.625

B2(z) = 0.625 – 0.0937 z–1 + z–2
(ii) To find k!,
A2 ( z ) − k2 B2 ( z )
A1 ( z ) =
1 − k22

A1 ( z ) =
(1 − 0.0937 z −1
+ 0.625 z −2 ) − (0.625)(0.625 − 0.0937 z −1 + z −2 )
1 − (0.625) 2

A1 ( z ) = 1 − 0.0576 z −1 ∴ k1 = –0.0576

Lattice Realization Diagram of All POLE IIR filter :

f3[n] f2[n] f1[n]

x[n] fo[n] y[n]
+ + +

–k3 k3 –k2 k2 –k1 k1

+ Z–1 + Z–1 + Z–1

g3[n] g2[n] g1[n]
[ ]
Sr No Parameter Direct Form Lattice form
1 Number of 3 6
Mltiplications
2 Number of Additions 2 6
3 Number of Delay Blocks 3 3

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

talelesir@yahoo.com : 9987030881
☯
 B E EXTC D T S P MAY- 2007 15

Q6. (a) Determine 8 point DFT for a continuous time signal x[t]= sin(2πFt) with FS = 50
Hz using DIF algorithm. [10]
Solution :
x[t]= sin(2πFt) Let F=1 Hz
⎛ 2πn ⎞ ⎛ 2πn ⎞ ⎛ nπ ⎞
By sampling, x[n ] = sin ⎜ ⎟ = sin ⎜ ⎟ = sin ⎜ ⎟
⎝ Fs ⎠ ⎝ 50 ⎠ ⎝ 25 ⎠
⎛ nπ ⎞
x[n ] == sin ⎜ ⎟
⎝ 25 ⎠
x[n ] == { 0, 0.1253, 0.2486, 0.3681, 0.4817,0.5877, 0.6845, 0.7705 }
↑
2π
N −1 −j
X [k ] = ∑ x[n] WNnk where (i) N = 8 (ii ) W N1 =e N
n =0

X[k] = X[0]+ X[1] WNk + X[2] WN2k + X[3] WN3k + X[4] WN4k + X[5]WN5k + X[6] WN6k + X[7] WN7k
X[k] = 0.1253 WNk + 0.2486 WN2k + 0.3681WN3k + 0.4817WN4k + 0.5877 WN5k + 0.6845 WN6k + 0.7705 WN7k
==================
k=0, X[0] = 3.2664
k=1, X[1] =0.1253 WN1 + 0.2486 WN2 + 0.3681WN3 + 0.4817WN4 + 0.5877 WN5 + 0.6845 WN6 + 0.7705 WN7
X[1] =
k=2, X[2] =0.1253 WN2 + 0.2486 WN4 + 0.3681WN6 + 0.4817WN8 + 0.5877 WN10 + 0.6845 WN12 + 0.7705 WN14
X[2] =
k=3, X[3] =0.1253 WN3 + 0.2486 WN6 + 0.3681WN9 + 0.4817WN12 + 0.5877 WN15 + 0.6845 WN18 + 0.7705 WN21
X[3] =
k=4, X[4] =0.1253 WN4 + 0.2486 WN8 + 0.3681WN12 + 0.4817WN16 +0.5877 WN20 + 0.6845 WN24+0.7705 WN28
X[4] =

k=5, X[5] =
k=6, X[6] =
k=7, X[7] =
--------------------------------------------------------------------------------------------------------
(b) With the help of block diagram, explain architecture of TMS320C5X series of
processor. [10] Soln : Refer DTSP-Dec-2004 Q7 (b)(i)
--------------------------------------------------------------------------------------------------------
Q7. (a) Explain the Goertzel Algorithm. [6] Soln : Refer DTSP-May-2005 Q7 (b)
(b) Write short note on applications of DCT. [6]
(c) Explain briefly any one method of long data filtering. [8]
Soln : Refer DTSP-May-2005 Q7 (c)
--------------------------------------------------------------------------------------------------------

talelesir@yahoo.com : 9987030881
☯