Digital Signal Processing

www.BrainKart.com

Click Here !!! for Digital Signal Processing full study material.

Click Here !!! for other subjects (Anna University)

Click Here !!! for Anna University Notes Android App.

Click Here !!! for BrainKart Android App.

Click Here for Digital Signal Processing full study material.

1. Check whether the signal is Periodic or Aperiodic:

x(t) = 2cos(10t+1) – sin(4t-1)
2𝜋 2𝜋 𝜋
Time period T1= = = sec
𝛺01 10 5
2𝜋 2𝜋 𝜋
Time Period T2= = = sec
𝜋⁄ 𝛺02 4 2
𝑇1 5 2
T= = = sec
𝑇2 𝜋⁄ 5
2
T = 5T1 = 2T2
5𝜋 2𝜋
T= 5 = 2

T = π sec

x(t) = cos 60πt + sin 50πt

2𝜋 1
T1 = = sec
60𝜋 30
2𝜋 1
T2 = = sec
50𝜋1 25
𝑇1 ⁄30 5
T= = =
1
𝑇2 ⁄25 6
T=6T1 = 5T2
1
T = sec
5

x(t) = 3cos 4t + 2sin πt

2𝜋 𝜋
T1 = = sec
4 2
2𝜋
T2 = = 2 sec
𝜋
𝑇1 𝜋
T= = this is not a rational number. So the signal is not periodic.
𝑇2 4

x(n) = cos 2πn

2𝜋 2𝜋𝑚 (Put some small value of m so that N becomes an integer)
N= =
⍵0 2𝜋
N=1

www.BrainKart.com
Click Here for Digital Signal Processing full study material.

x(n) = 𝒆𝒋𝟔𝝅𝒏
2𝜋𝑚 2𝜋𝑚 3
N= =
6𝜋
=3
⍵0
N=1

𝟑𝝅
𝒋(𝟐𝝅 )𝒏 𝒋( )𝒏
x(n) = 𝒆 𝟑 + 𝒆 𝟒

2𝜋𝑚
N1 = = 3m for m = 1; N1 = 3
2𝜋⁄
2𝜋𝑚3 8𝑚
N2 = = for m= 3; N2 =8
3𝜋⁄ 3
4
𝑁1 3
= =N
𝑁2 8
𝑁 = 8𝑁1 = 3𝑁2
8(3) = 3(8) = 24
N = 24

x(n) = 12cos (20n)

2𝜋𝑚 𝜋𝑚
N= =
20 10
For any values of m N is not an integer. So the given signal is aperiodic.

2. Check whether the systems are Time variant/invariant:

a. T[x (n)] = g (n)x (n)
y (n) = g (n) x (n)
Shift the input by k
𝑦1(n) = g (n).x (n-k) ⟶1
Shift the output by k
y (n-k) = g (n-k).x (n-k) ⟶2
Shift in input and shift in output is not equal. So the system is time/shift
variant.

b. T[x(n)] = ∑𝒏𝒌=𝒏𝟎 𝒙(𝒌)

y (n) =∑ 𝑛𝑘=𝑛0 𝑥(𝑘)
www.BrainKart.com
Click Here for Digital Signal Processing full study material.

y (n) = x (𝑛0) + x (𝑛0+1) +. ....... + x (n-1) + x (n)

Shift the output by k
y (n-k) = x (𝑛0) + x (𝑛0+1) + ....... + x (n-k-1) + x (n-k) ⟶1
Shift the input by k
𝑦1 (n) = x (𝑛0) + x (𝑛0+1) +........ +x (n-k-1) + x (n-k) ⟶2
Shift in input and output do not vary. So the system is time invariant.

c. T[x(n)] = 𝒆𝒙(𝒏)
y(n)= 𝑒𝑥(𝑛)
Shift the input by k
y (n) = 𝑒 (𝑛−𝑘)⟶1
Shift the output by k
y (n-k) = 𝑒𝑥(𝑛−𝑘)⟶2
Shift the input and output do not vary. So the system is time invariant.

d. y (n) = x (n)cos 𝝎𝟎n

Shift the input by k
y (n) = x (n-k) cos 𝜔0n⟶1
Shift the output by k
y (n-k) = x (n-k) cos 𝜔0(n-k) ⟶2
Shift in input and output varies. So the system is time variant.

𝟏
e. y(n) = x( )
𝟐𝒏
Shift the input by k
1
y1 (n) = x ( ) ⟶1
2(𝑛−𝑘)
Shift the output by k
1
y (n-k) = x ( ) ⟶2
2(𝑛−𝑘)
The shift in input and output do not vary. So the system is time invariant.

f. y (n) = x (n) + nx (n+1)

Shift the input by k
y1 (n) = x (n-k) + n x (n-k+1) ⟶1
www.BrainKart.com
Click Here for Digital Signal Processing full study material.

Shift the output by k

y (n-k) = x (n-k) + (n-k) x (n-k+1) ⟶2
Shift in input and output varies. So the system is time variant

g. y(n) = sin x (n)

Shift the input by k
y1 (n) = sin x (n-k) ⟶1
Shift the output by k
y (n-k) = sin x (n-k) ⟶2
Shift in the input and output do not vary. So the system is time invariant.
𝐬𝐢𝐧 [(𝒏+𝟏)𝝎]
3. a) Find the Z transform for the signal x (n) =r n
𝐬𝐢𝐧 𝝎
u (n); 0<r<1
Solution:
X (Z) = ∑∞𝑛=−∞ 𝑥 (𝑛) 𝑍−𝑛
sin [(𝑛+1)𝜔]
X (Z) = ∑∞𝑛=−∞ 𝑟𝑛 sin 𝜔
𝑢 (𝑛) 𝑍−𝑛

= ∑∞ 𝑟𝑛 sin [(𝑛+1)𝜔] 𝑍−𝑛=∑∞ 𝑟𝑛 ej (n+1) ω− e−j (n+1) ω 𝑍−𝑛

𝑛=0 sin 𝜔 𝑛=0 sin 𝜔 2𝑗

=∑∞ 𝑟𝑛 ej (n+1) ω 𝑟𝑛 e−j (n+1) ω

𝑍−𝑛 - ∑𝑛=0
∞
𝑛=0 sin 𝜔 𝑍−𝑛
2𝑗 sin 𝜔 2𝑗

=∑∞ 𝑟𝑛 ejnωejω
𝑍−𝑛 - ∑𝑛=0
∞ 𝑟𝑛 e−jnωe−jω
𝑍−𝑛
𝑛=0 sin 𝜔 2𝑗 sin 𝜔 2𝑗

=∑∞ ejω ejnω𝑟𝑛

𝑍−𝑛 - ∑𝑛=0 e e
∞ −jω −jnω𝑟𝑛
𝑍−𝑛
𝑛=0 sin 𝜔 2𝑗 sin 𝜔 2𝑗

=∑∞ ejω ejω𝑟 −jω e−jω𝑟

𝑍−1) 𝑛 - ∑𝑛=0 e
∞
𝑛=0 sin 𝜔 ( ( 𝑍−1) 𝑛
2𝑗 sin 𝜔 2𝑗

ejω 1 e−jω 1 1 ejω(1−e−jω𝑟𝑍 −1)−e−jω (1−ejω𝑟𝑍−1)

= - = [ ]
2jsin 𝜔 1−ejω𝑟𝑍−1 2jsin 𝜔 1−e−jω𝑟𝑍−1 2jsin 𝜔 (1−ejω𝑟𝑍 −1) (1−e−jω𝑟𝑍−1)

1 ejω−r𝑍−1−e−jω+r𝑍−1 1 2jsin 𝜔
=2jsin 𝜔 [ ]= [ ]
1−e−jω𝑟𝑍 −1−ejω𝑟𝑍−1+𝑟2𝑍−2 2jsin 𝜔 1−𝑟𝑍−1 (e−jω+ejω) +𝑟2𝑍 −2
1 1
X (Z) = = 𝑍−2 (𝑍2−𝑟𝑍2 cos 𝜔+𝑟2)
1−𝑟𝑍−12 cos 𝜔+𝑟2𝑍−2

www.BrainKart.com
Click Here for Digital Signal Processing full study material.
𝑍2
X (Z) =
(𝑍2−𝑟𝑍2 cos 𝜔+𝑟2)

𝒁+𝟎.𝟐
3. b) Find the inverse Z transform of X (Z) = , |Z|>1
(𝒁+𝟎.𝟐) (𝒁−𝟏)

Solution:
𝑍+0.2
X (Z) =𝑍2− 0,5𝑍−0.5

X (Z) =Z-1+0.7Z-2+0.85Z-3+0.775Z-4+……
= ∑∞
𝑛=0 𝑥 (𝑛) 𝑍
−𝑛

X (0) =0, x (1) =1, x (2) =0.7, x (3) =0.85, x (4) =0.775, and so on

4. a) Find the 8 point DFT FFT of the sequence x(n)={1,2,3,4}

Solution
𝑊80=(𝑒−𝑗2𝜋/8)0 = 1
𝜋 𝜋
𝑊1=(𝑒−𝑗2𝜋/8)1 = cos - j sin = 0.707-j0.707
8 4 4
𝜋 𝜋
𝑊2=(𝑒−𝑗2𝜋/8)2 = cos - j sin = 0-j(1)= -j
8 2 2
3 3𝜋 3𝜋
𝑊 =(𝑒−𝑗2𝜋/8)3 = cos - j sin = - 0.707-j0.707
8 4 4

www.BrainKart.com
Click Here for Digital Signal Processing full study material.

X(k)={10, -2, -2+2j, -2-2j, -0.414, 2.414, 2.414, -0.144}

4. b) Explain in detail the signal flow graph of DIT radix 2 FFT

The basic computation involves
 Two complex numbers a and b in each computation
 Complex number b is multiplied by a phase factor 𝑊𝑛
𝑘

 The product b 𝑊𝑛𝑘 is added to the complex number a to form new

complex number A
 The product b𝑊𝑛𝑘 is subtracted from the complex number a to form
new complex number B
The basic butterfly or flow gram of DIT radix 2 FFT

In radix 2 FFT, N/2 butterflies per stage are required to represent the
computational process

www.BrainKart.com
Click Here for Digital Signal Processing full study material.

First stage of computation

v11(0)=x(0)+𝑊𝑁0 x(4)
⁄4

v11(1)=x(0)-𝑊𝑁0 x(4)
⁄4

v12(0)=x(2)+𝑊𝑁0⁄x(6)
4

v12(1)=x(2)-𝑊𝑁0 x(6)
⁄4

v21(0)=x(1)+𝑊𝑁0⁄x(5)
4

v21(1)=x(1)-𝑊𝑁0 x(5)
⁄4

v22(0)=x(3)+𝑊𝑁0⁄x(7)
4

v22(1)=x(3)-𝑊𝑁0⁄x(7)
4

Second stage of computation

F1(0)= v11(0)+ 𝑊𝑁0⁄ v12(0)

4

F1(1)= v11(1)+ 𝑊𝑁0 v12(1)

⁄4

F1(2)= v11(0)- 𝑊𝑁0 v12(0)

⁄4

F1(3)= v11(1)- 𝑊𝑁0 v12(1)

⁄4
www.BrainKart.com
Click Here for Digital Signal Processing full study material.
F2(0)= v21(0)+ 𝑊𝑁0 v22(0)
⁄4

F2(1)= v21(1)+ 𝑊𝑁0⁄ v22(1)

4

F2(2)= v21(0)- 𝑊𝑁0⁄ v22(0)

4

F2(3)= v21(1)- 𝑊𝑁0⁄ v22(1)

4

Third stage of computation

X(0)= F1(0)+ 𝑊𝑁0 F2(0)

X(1)= F1(1)+ 𝑊𝑁1 F2(1)
X(2)= F1(2)+ 𝑊𝑁2 F2(2)
X(3)= F1(3)+ 𝑊𝑁3 F2(3)
www.BrainKart.com
Click Here for Digital Signal Processing full study material.

X(4)= F1(4)+ 𝑊𝑁4 F2(4)

X(5)= F1(5)+ 𝑊𝑁5 F2(5)
X(6)= F1(6)+ 𝑊𝑁0 F2(6)
X(7)= F1(7)+ 𝑊𝑁0 F2(7)
Combined butterfly diagram is

5. Design a butter worth filter using bilinear transformation technique

0.8 ≤ |H (𝒆𝒋𝝎)| ≤ 𝟏 ; 𝟎 ≤ 𝝎𝒑 ≤ 𝟎. 𝟐𝝅
|H (𝒆𝒋𝝎)| ≤ 𝟎. ; 0.6𝝅 ≤ 𝝎𝒔 ≤ 𝝅
Solution:
Step 1: draw the filter characteristics

www.BrainKart.com
Click Here for Digital Signal Processing full study material.

Step 2: collect the given parameters 𝜔𝑝 = 0.2 , 𝜔𝑠 = 0.6𝜋

1 1
=0.8 ⇒ 0.8√1 + 𝜀2 = 1 ⇒ √ 2 2 12
= 1+𝜀 ⇒1+𝜀 =( ) ⇒
√1+𝜀2 0.8 0.8

ε = 0.75
1
= 0.2 ⇒ λ = 4.8989
√1+𝜆2

Step 3: Select the transformation technique (BLTT)

2 𝜔𝑝 2 0.2𝜋
Ω = tan ⇒ tan = 0.6498
p
𝑇 2 1 2
2 𝜔𝑠 2 0.6𝜋
Ωs = tan ⇒ tan = 2.7527
𝑇 2 1 2

Step 4: Find the order of filter

𝑙𝑜(𝜆⁄𝜀)
For LPF, N ≥ ⇒
𝑙𝑜(4.8989⁄0.75 )
𝑙𝑜(Ω𝑠⁄Ω𝑝) 𝑙𝑜(2.7527⁄0.6498)

On simplifying
N ≥ 1.299 ⇒ N = 2

Step 5: Butterworth polynomial for order 2, 𝑆2 + √2𝑆 + 1

Step 6: Transfer function of normalized LPF for N=2 (by substituting
the Butterworth polynomial for order 2)
1
|H (jω)| =
𝑆2+√2𝑆+1

Step 7: Denormalize the transfer function

𝑆
For LPF, S→
Ω𝑐
Ω𝑃 0.6498
Ωc = ⇒ (0.75)1⁄2 = 0.7503
𝜀1⁄𝑁

𝑆
S→ 0.7503
1
H(S) = 𝑠 𝑠
[ ]2+√2[ ]+1
0.7503 0.7503

0.5629 0.5629
= 𝑆2+√2(0.7503)+0.5629
= 𝑆2+1.061𝑆+0.5629
www.BrainKart.com
Click Here for Digital Signal Processing full study material.

Step 8: convert analog filter into respective digital filter using BLTT
2 1−𝑧−1
Replace S→ ( ) , T =1
𝑇 1+𝑧−1

0.5629
H (z) = 2 1−𝑧
−1
2 1−𝑧−1
2( + 1.061(2)( )+0.5629
1+𝑧 −1) 1+𝑧−1

0.5629 (1+𝑧−1)2
=
4(1−𝑧−1)2+2.122(1−𝑧−1)(1+𝑧−1)+0.5629(1+𝑧−1)2

0.5629 (1+2𝑧−1+𝑧−2)
= 4(1−2𝑧−1+𝑧−2)+2.122(1−𝑧−2)+0.5629(1+2𝑧−1+𝑧−2)

0.5629+1.1258𝑧−1+0.5629𝑧−2
H (Z) = 6.6849−6.8742𝑧−1+2.4409𝑧−2
0.5629+1.1258𝑧−1+0.5629𝑧−2
H (Z) = 6.6849[1−1.0283𝑧−1+0.3651𝑧−2]

Step 9: final transfer function

0.0842+0.1684𝑧−1+0.0842𝑧−2
H (Z) =
1−1.0283𝑧−1+0.3651𝑧−2

6. Design a HPF, using hamming window with a cut off frequency

1.2rad/sec and N=9.
Solution:

Given: N=9 ⍵c=1.2 rad/sec

For HPF desired frequency response is
H (⍵) = 𝑒−𝑗⍵𝛼 ; −𝜋 ≤ ⍵ ≤ −𝜔𝐶 𝑎𝑛𝑑 𝜔𝐶 ≤ ⍵ ≤ 𝜋
d {
0; 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
𝑁−1 9−1
α= 2 ⇒α= ⇒α=4
2
H (⍵) = 𝑒−𝑗⍵4 ; −𝜋 ≤ ⍵ ≤ −1.2 𝑎𝑛𝑑 1.2 ≤ ⍵ ≤ 𝜋
d {
0; 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
1 𝜋
hd (n) = ∫−𝜋 𝐻𝑑 (𝜔)𝑒𝑗𝜔𝑛𝑑𝜔 [Hd (ej⍵) =Hd (𝜔)]
2𝜋

www.BrainKart.com
Click Here for Digital Signal Processing full study material.
1`
hd (n) = −1.2 1` �� −𝑗⍵4 𝑗⍵𝑛
∫ 𝑒−𝑗⍵4𝑒𝑗⍵𝑛𝑑⍵ + ∫ 𝑒 𝑒 𝑑⍵
2𝜋 −𝜋 2𝜋 1.2

1` −1.2 1` 𝜋
= ∫−𝜋 𝑒𝑗⍵(𝑛−4)𝑑𝜔 + ∫ 𝑒𝑗(𝑛−4)𝑑𝜔
2𝜋 2𝜋 1.2

1` 𝑒𝑗(𝑛−4) 1` 𝑒𝑗(𝑛−4)
= 2𝜋 { (𝑛−4)} + 2𝜋
{
(𝑛−4)
}

1 1
=2𝜋(𝑛−4) {𝑒−1.2(𝑛−4) − 𝑒−𝜋𝑗(𝑛−4)}+ 2𝜋(𝑛−4)
{𝑒1.2(𝑛−4) − 𝑒𝜋𝑗(𝑛−4)}

1 𝑒−𝑗1.2(𝑛−4)−𝑒𝑗1.2(𝑛−4) 𝑒𝑗𝜋(𝑛−4)−𝑒−𝑗𝜋(𝑛−4)
= (𝑛−4) {{ 2𝑗
}+ {
2𝑗
}}

1
= 𝜋2(𝑛−4) {𝑒−𝑗1.2(𝑛−4) − 𝑒𝑗1.2(𝑛−4) + 𝑒𝑗𝜋(𝑛−4) − 𝑒−𝑗𝜋(𝑛−4)}
1
hd (n)= { 𝑖𝑛[𝜋(𝑛 − 4)] − 𝑠𝑖𝑛[1.2(𝑛 − 4)]}
(𝑛−4)

Now, n=0 to N-1 and here n= 0 to 8

1
hd (0) = {𝑠𝑖𝑛(0 − 4) − 𝑠𝑖𝑛1.2(0 − 4)}
(0−4)
1
hd (0) = (0-0.9961) = 0.0792
−12.5663
1
hd (1) = {sin π (1-4) – sin1.2 (1-4)}
(1−4)
1
hd (1) = (0-0.4425) = 0.0478
−9.4247
1
hd (2) = {sin π (2-4) – sin1.2 (2-4)}
(2−4)
1
hd (2) = (0+0.6754) = -0.1075
−6.2831
1
hd (3) = {sin π (3-4) – sin1.2 (3-4)}
(3−4)
1
hd (3) = (0+0.9320) = -0.2966
−3.1415
1
hd (4) = sin π (4-4) – sin1.2 (4-4)} = ∞ (infinity)
(4−4)
So applying L’hospital rule,
At n=α=4
1
hd (4) = {𝑠𝑖𝑛𝜋 (𝑛 − 4) − 𝑠𝑖𝑛1.2 (𝑛 − 4)}
(𝑛−4)
𝑠𝑖𝑛(𝑛−4).1 1 𝑠𝑖𝑛 1.2(𝑛−4)
= (𝑛−4)
- 𝜋 (𝑛−4)

www.BrainKart.com
Click Here for Digital Signal Processing full study material.
1 𝜋−1.2
hd (4) = 1 - = =0.6180
𝜋 𝜋
1
hd (5) = {sin π (5-4) – sin1.2 (5-4)}
(5−4)
1
hd (5) = (0-0.9320) = -0.2966
𝜋
1
hd (6) = {sin π (6-4) – sin1.2 (6-4)}
(6−4)
1
hd (6)= (0-0.6754) = -0.1075
2𝜋
1
hd (7) = {sin π (7-4) – sin 1.2 (7-4)}
(7−4)
1
hd (7)= (0-0.4425) = 0.0478
3𝜋
1
hd (8) = sin π (8-4) – sin 1.2 (8-4)}
(8−4)
1
hd (8) = (0+0.9961) = 0.0792
4𝜋
Now, to find ⍵H (n)
2𝑛
0.54 − 0.46𝑐𝑜𝑠 ;0≤𝑛≤𝑁−1
⍵H (n) ={ 𝑁
0; 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
⍵H (0) = 0.54 - 0.46 cos (0) = 0.08
𝜋
⍵H (1) = 0.54 - 0.46cos ( ) = 0.2147
4
𝜋
⍵H (2) = 0.54 - 0.46cos ( ) = 0.54
2
𝜋
⍵H (3) = 0.54 - 0.46cos ( ) = 0.8652
8
⍵H (4) = 0.54 - 0.46cos (π) = 1
10𝜋
⍵H (5) = 0.54 - 0.46cos ( ) = 0.8652
8
12𝜋
⍵H (6) = 0.54 - 0.46cos ( ) = 0.54
8
14𝜋
⍵H (7) = 0.54-0.46cos ( ) = 0.2147
8
⍵H (8) = 0.54 - 0.46cos (2π) = 0.08
h (n)= hd (n) ⍵H (n)
h (0)= hd (0) ⍵H (0) = 0.0792 x 0.08 = 6.335x10-3
h (1)= hd (1) ⍵H (1) = 0.0478 x 0.2147 = 0.01006
h (2)= hd (2) ⍵H (2) = -0.1075 x 0.54 = =0.05799
h (3)= hd (3) ⍵H (3) = -0.2966 x 0.8652 = -0.2566
h (4)= hd (4) ⍵H (4) = 0.6180 x 1 = 0.6180
h (5)= hd (5) ⍵H (5) = -0.2966 x 0.8652 = -0.2566
h (6)= hd (6) ⍵H (6) = -0.1075 x 0.54 = =0.05799
h (7)= hd (7) ⍵H (7) = 0.0478 x 0.2147 = 0.01006
www.BrainKart.com
Click Here for Digital Signal Processing full study material.

h (8)= hd (8) ⍵H (8) = 0.0792 x 0.08 = 6.335x10-3

Magnitude response of N= odd,
𝑁−1
|H (w)| = h ( ) + ∑𝑁−12
𝑁−1 − 𝑛 𝑐𝑜𝑠 ⍵
2ℎ
2 𝑛=1 ( 2 ) n
= h (4) + ∑4𝑛=1 2 ℎ (4 − 𝑛) 𝑐𝑜𝑠 ⍵n
= h(4)+ [2h(3)cos ⍵ +2h(2)cos 2⍵+2h (1) cos 3⍵ + 2h (0) cos 4⍵]
= 0.6180 + {2(-0.2566) cos ⍵ + 2(-0.05799) cos 2⍵ +
2(0.01006) cos 3⍵ + 2(6.335x10-3) cos 4⍵}
= 0.6180 + {-0.5132cos ⍵ – 0.11598 cos 2⍵ +
0.02012 cos 3⍵ +0.12672 cos 4⍵}
Transfer function,
H (z) = ∑𝑛=0
𝑁−1 ℎ (𝑛)-n

=h (0)z-0+ h (1)z-1+ h (2)z-2+ h (3)z-3+ h (4)z-4+ h (5)z-5+ h (6)z-6+

h (7)z-7+ h (8)z-8
= 6.335x10-3 z-0+ 0.01006 z-1 - 0.05799 z-2 - 0.2566 z-3 +0.6180 z-4
- 0.2566 z-5 –0.05799 z-6 +0.01006 z-7 + 6.335x10-3 z-8
=6.335x10-3(z-0+z-8) + 0.0101(z-1+z-7)- 0.05799(z-2+z-6)-
0.2566(z-3+z-5)+ 0.6180(z4)

7. a) Derive the steady state output noise power quantization of input

data.
Quantization step size – q
𝑅
Q= 𝑏 (for 2’s complement)
2

𝑅
Q=2𝑏−1 (for sign magnitude and 1’s complement)

Let x (n) =Unquantised sample of the signal.

www.BrainKart.com
Click Here for Digital Signal Processing full study material.

xq (n) =quantised sample of the signal.

e (n) =quantization error.
e (n) = xq (n)-x (n)
−𝑞 𝑞
Range of e (n) = to
2 2

Mean value or expected value of error signal=E {e}

1 𝑞/2
∫−𝑞/2
E {e} =𝑞 𝑞
𝑒de
2
−(−2))

𝑞
1 𝑒2
= [ ] 2
−𝑞
𝑞 2
2

1 2 𝑞2
= [ − ]-
2𝑞 4 4

E {e} =0
Variance of error signal 𝜎e2=E {e2}-E2 {e}
E {e} =0 so E 2 {e} =0
1 𝑞/2
E {e} =𝑞 ∫ 𝑞 −𝑞/2
𝑒2de
2
−(−2)

1 𝑒3 𝑞⁄
2
= 𝑞[ 3
]−𝑞
⁄2

1 3 −𝑞3
=3𝑞 [ 8− ]
8

1 2𝑞3
=3𝑞[ 8
]

𝜎2 𝑞2 𝑅
e = Where q= 𝑏
12 2
𝑅2
∴ 𝜎 2= 𝑅22−2𝑏
e
22𝑏12
12= 12
2−2𝑏
When R=2, 𝜎 2=
e
3

Steady state noise power due to the quantization error signal

www.BrainKart.com
Click Here for Digital Signal Processing full study material.

h (n) ⟶impulse response

y’ (n) ⟶response
y’ (n) =xq (n)*h (n)
=[x (n) +e (n)]*h (n)
=[x (n)*h (n)] + [e (n)*h (n)]
=y (n) + (𝑛)
Where y (n) =x (n)*y (n)
(n)=e (n)*h (n)
(n)=output noise power or steady state output noise power (variance)
due to quantization error signal
Steady state output noise power due to quantization
error=𝜎2 2=𝜎2 ∑∞ ℎ2(n)
𝑒𝑜 𝑒 𝑛=0

1
=𝜎𝑒2 2𝜋𝑗 ∮𝑁 (Z) H (Z-1) Z-1dz
= 𝜎2 ∑𝑁 𝑟𝑒[H (Z) H (Z-1) Z-1] |z=pi
𝑒 𝑖=1

Pi=P1, P2,. ...... PN.

7. b) Explain the characteristics of a limit cycle oscillation with respect to
the system described in the equation
y (n) =0.95 y (n-1) +x (n). Determine the dead band.
Solution:
y (n) = 0.95y (n-1) +x (n)
y’ (n) =Q [0.95y’ (n-1)] + x (n)
Assume
www.BrainKart.com
Click Here for Digital Signal Processing full study material.

y’ (n) =0 for n<0

x (n) = 0.75 𝑓𝑜𝑟 𝑛 = 0
{
0 𝑓𝑜𝑟 𝑛 ≠ 0

When n=0
y’ (n) =Q [0.95y’ (-1)] +x (0)
=Q [0.95x0] +0.75
=0.75
=0.11002
When n=1
y’ (n) =Q [0.95y’ (1-1)] +x (1)
=Q [0.95x0.75] +0
=Q [0.7125]
𝑡𝑜 𝑏𝑖𝑛𝑎𝑟𝑦 𝑎𝑑𝑑 𝑠𝑖𝑔𝑛 𝑏𝑖𝑡 𝑟𝑜𝑢𝑛𝑑 𝑜𝑓𝑓 𝑡𝑜 4 𝑏𝑖𝑡𝑠
.7125→ . 101102 → 0.101102 →

𝑐𝑜𝑛𝑣𝑒𝑟𝑡 𝑡𝑜 𝑑𝑒𝑐𝑖𝑚𝑎𝑙 𝑒𝑥𝑡𝑟𝑎𝑐𝑡 𝑠𝑖𝑔𝑛 𝑏𝑖𝑡

0.6875 ← +. 10112 ← 0.10112
0.7125x2=1.425 .1011
0.425x2=0.85
1x2-4=0.0625
0.85x2=1.7
1x2-3=0.125
0.7x2=1.4
0x2-2=0
0.4x2=0.8
1x2-1=0.5
y’ (1) =0.6875
When n=2
y’ (2) = Q [0.95y’ (2-1)] +x (2)
=Q [0.95x0.6875] +0

www.BrainKart.com
Click Here for Digital Signal Processing full study material.

=Q [0.653125]
𝑡𝑜 𝑏𝑖𝑛𝑎𝑟𝑦 𝑎𝑑𝑑 𝑠𝑖𝑔𝑛 𝑏𝑖𝑡 𝑟𝑜𝑢𝑛𝑑 𝑡𝑜 4 𝑏𝑖𝑡𝑠
+.653125 → +. 1010012 → 0.1010012 →
𝑡𝑜 𝑑𝑒𝑐𝑖𝑚𝑎𝑙 𝑒𝑥𝑡𝑟𝑎𝑐𝑡 𝑠𝑖𝑔𝑛 𝑏𝑖𝑡
+.625 ← +.1010 ← 0.1010

.653125x2=1.30625
.1010
.30625x2=0.6125
0x2-4=0
.6125x2=1.225
.225x2=0.45 1x2-3=1⁄8

.45x2=0.9 0x2-2=0
.9x2=1.8 1x2-1=1⁄2
=0.1010012 y’ (2) =0.625
When n=3
y’ (3) = Q [0.95y’ (3-1)] +x (3)
=Q [0.95x0.625] +0
=Q [0.59375]
𝑡𝑜 𝑏𝑖𝑛𝑎𝑟𝑦 𝑎𝑑𝑑 𝑠𝑖𝑔𝑛 𝑏𝑖𝑡 𝑟𝑜𝑢𝑛𝑑 𝑡𝑜 4 𝑏𝑖𝑡𝑠
+.59375 →+. 100112 → 0.100112 →
𝑡𝑜 𝑑𝑒𝑐𝑖𝑚𝑎𝑙 𝑒𝑥𝑡𝑟𝑎𝑐𝑡 𝑠𝑖𝑔𝑛 𝑏𝑖𝑡
+.625 ← +.1010 ← 0.1010

System enters the limit cycle when n=2

2−𝑏
Dead band=± here b=5 |a|=0.95
1−|𝑎|

2−5
=± =± 0.625 = [+0.625, -0.625]
1−0.95

www.BrainKart.com
Click Here for Digital Signal Processing full study material.

1. State the sampling theorem

The sampling frequency must be at least the highest frequency present in
the signal
F≥2fm

2. Define a signal
A signal is defined as any physical quantity that varies with time, space or
any other independent variable.

3. define a system
A system is defined as an entity that manipulates one or more signals to
accomplish a function, therefore producing new signal.
x(t) input signal y(t) output signal
System

4. What is the condition for stability?

A system is said to be stable if it produces a bounded output for every
bounded input (BIBO).
The system which does not satisfy this condition is an unstable
system.
Condition for Stability:
∞
∫−∞ ∣ ℎ(𝑡) ∣ 𝑑𝑡 < ∞

5. State the superposition theorem

The response to a weighted sum of a signal be equal to the
corresponding weighted sum of the outputs of the system to each of the
individual input signal,
A system is said to be linear if and only if
T [a1 x1(n) + a2x2 (n)] = a1T [x1(n)] + a2T [x2(n)]

www.BrainKart.com
Click Here for Digital Signal Processing full study material.

A system is said to be non-linear if it doesn’t obey superposition

principle.
6. Define Z transform.
Z transform converts difference equations into algebraic equations thereby
simplifying the analysis of discrete time systems
Definition
The Z transform of a discrete time sequence x (n) is defined as
X (Z) =∑𝑛=−∞
∞ 𝑥 (𝑛) 𝑍−𝑛

7. What is bilateral Z transform?

The Z transform of a discrete time sequence x (n) is defined as
X (Z) =∑𝑛=−∞
∞ 𝑥 (𝑛) 𝑍−𝑛
If the sequence x (n) exists for n in the range -∞ to ∞ the above equation
represents two sided or bilateral Z transform

8. What is unilateral Z transform?

If the sequence exists only for n≥0 then the equation changes to
X (Z) =∑𝑛=0
∞ 𝑥 (𝑛) 𝑍−𝑛

Which is called one sided Z transform or unilateral Z transform

9. Define inverse Z transform.

Inverse Z transform of X (z) is
1
X (n) = ∮ 𝑋 (𝑍) 𝑍𝑛−1𝑑𝑍
2𝜋𝑗 𝑐

10. Define Region of convergence.

www.BrainKart.com
Click Here for Digital Signal Processing full study material.

ROC is the region where Z transform converges. From definition of Z

transform it is clear that Z transform is an infinite power series
11. What is zero padding? Why it is needed?
Appending zeros to a sequence in order to increase the size or length
of the sequence is called zero padding.
During convolution when two input sequences are of different size
then they are converted to equal size by zero padding.

12. What is sectioned convolution?

In linear convolution of two sequences if one sequence is very much
longer, the longer sequence is sectioned(splitted) into smaller
sequence equal to the size of the smaller sequence and then the
convolution is performed. The output sequences obtained are finally
combined to get the overall output sequence this technique is called
sectioned convolution

13. What is radix 2 FFT?

The radix 2 FFT is an efficient algorithm for computing N point DFT of
a N point sequence. In radix 2 FFT the N point sequence is decimated
into 2 point sequences and the 2 point DFT for each decimated
sequence is computed. From the result of 2 point DFT the 4 point DFTs
are computed. From the 4 point DFTs the 8 point DFTs are computed
and so on until we get N point DFT

14. How many complex multiplications and additions are involved in DFT
and FFT?
In FFT
𝑁
Complex multiplications involved - log2N
2
www.BrainKart.com
Click Here for Digital Signal Processing full study material.

Complex additions involved –N log2N

In DFT
Complex multiplications involved – N2
Complex additions involved –N (N-1)

15. What is decimation in time radix 2 FFT?

The DIT radix 2 FFT is an efficient algorithm for computing DFT. In
DIT the time domain N point sequence is decimated into 2 point
sequences. The result of 2 point DFTs are used to compute 4 point
DFTs. The two numbers of 2 point DFTs are combined to get 4 point
DFT. the result of 4 point DFTs are used to compute 8 point DFTs. Two
numbers of 4 point DFTs are combined to get an 8 point DFT. This
process is continued until we get N point DFT.

16. What is phase factor or twiddle factor?

The complex number WN is called phase factor or twiddle factor. WN

representsa complex number 1⎿-2π/N or e-j2π/n. It also representsthe
Nth root of unity.

17. What are the basic elements used to construct the block diagram?
 Adder
 Multiplier
 Delay unit

18. List the types of structures for realizing IIR?

 Direct form 1
 Direct form 2
 Cascade form
 Parallel form

19. What is the advantage of Direct Form1 over Direct Form2?

www.BrainKart.com
Click Here for Digital Signal Processing full study material.

In direct form 2 structure the number of delay elements requires is

exactly half that of DF1, when the number of poles and zeros are equal.
Hence it requires less memory.

20. What are the difficulties in cascade realization?

 Decision of pairing poles and zeros.
 Deciding the order of cascading the first and second order sections.
 Scaling multipliers should be provided between individual sections to
prevent the filter variables from becoming too large or too small.

21. What is the advantage in cascade and parallel realization?

During digital implementation the filter coefficients are quantized.
This may change the values of poles. This can be minimized by using
cascade and parallel realization.

22. What is Gibb’s phenomenon?

In FIR filter design by Fourier series method or rectangular window method,
the infinite duration impulse response is truncated to finite duration impulse
response. The abrupt truncation of impulse response introduces oscillations
in the passband and stopband. This effect is called Gibb’s oscillation.

23. What are the steps involved in FIR filter design?

1. Choose the desired (ideal) frequency response Hd (w).
2. Take IFT of Hd (w) to get hd (n).
3. Convert the infinite duration hd (n) to finite duration sequence h (n).
4. Take z-transform of h (n) to get the transfer function H (z) of the FIR
filter.

1. Multiply H(z) by z-(N-1)/2 to convert the noncausal transfer function to

a realizable causal FIR filter transfer function.

24. Write the procedure for designing FIR filters using window technique.
1. Choose the desired frequence response of the filter Hd(w).
2. Take IFT of Hd(w) to obtain the desired impulse response hd(n).
3. Choose a window sequence w(n) and multiply hd(n) with w(n) to
convert the infinite duration impulse response to finite duration impulse
response h(n).
www.BrainKart.com
Click Here for Digital Signal Processing full study material.

4. Take z-transform of h(n) to find the transfer function H(z) of the

filter.

(i) The mean value of rounding error signal is zero

(ii) The variance of the rounding error signal is least

25. What is limit cycle?

In recursive systems, when the input is zero or some non-zero
constant value, the non linearities due to finite precision arithmetic
operation may cause periodic oscillations in output. During periodic
oscillations the output will oscillate between a finite positive and negative
value the output becomes constant for increasing n. Such oscillation is called
limit cycles.
26. Define dead band
In a limit cycle the amplitudes of the output are confined to a
range of values called dead band of the filter
2−𝑏 −2−𝑏 +2−𝑏
Dead band = ± =[ , ]
1−|𝑎| 1−|𝑎| 1−|𝑎|
b- Number of bits (including sign bit)

www.BrainKart.com