Chapter 2

Sampling and Quantization

2.1 Find the Nyquist rate and Nyquist interval for each of the following signals:
1. x1 (t) = 3 cos(50πt) + 5 sin(300πt)

2. x2 (t) = 1 + 2 cos(20πt) + 3 sin(30πt)

3. x3 (t) = sinc2 (50πt)
4. x4 (t) = sinc(200πt)sin(150πt)

Solution 2.1 (1) Given that

x1 (t) = 3 cos(50πt) + 5 sin(300πt)
Comparing with
x1 (t) = A1 cos(Ω1 t) + A2 sin(Ω2 t)
we have
Ω1 = 50π and Ω2 = 300π

Maximum frequency present in the signal x(t) is

Ωmax = Ω2 = 300π or fmax = 150

Therefore, the Nyquist rate is given by

Ωs = 2Ωmax = 600π or fs = 2fmax = 300 Hz

The Nyquist interval is given by

1 1
Ts = = sec.
fs 300
(2) Given that
x2 (t) = 1 + 2 cos(20πt) + 3 sin(30πt)

Comparing with
x2 (t) = A0 + A1 cos(Ω1 t) + A2 sin(Ω2 t)

we have
Ω1 = 20π and Ω2 = 30π

Maximum frequency present in the signal x(t) is

Ωmax = Ω2 = 30π or fmax = 15

Therefore, the Nyquist rate is given by

Ωs = 2Ωmax = 60π or fs = 2fmax = 30 Hz

2 Digital Signal Processing

The Nyquist interval is given by

1 1
Ts = = sec.
fs 30
(3) Given that
sin2 (50πt)
x3 (t) = sinc2 (50t) =
(50πt)2
1 1
x(t) = − cos(100πt)
2(50πt)2 2(50πt)2
Comparing with
x(t) = A0 − A1 cos(Ω1 t)
we have
Ω1 = 100π or f1 = 50
Therefore, the Nyquist rate is given by
Ωs = 2Ω1 = 200π or fs = 2f1 = 100 Hz
The Nyquist interval is given by
1 1
Ts = = sec.
fs 100
(4) Given that
x4 (t) = sinc(200t)sin(150πt)
sin(200πt)
= sin(150πt)
200πt
1 1
= cos(50πt) − cos(350πt)
2(200πt) 2(200πt)
Comparing with
x4 (t) = A1 cos(Ω1 t) − A2 cos(Ω2 t)
we have
Ω1 = 50π and Ω2 = 350π
Maximum frequency present in the signal x(t) is
Ωmax = Ω2 = 350π or fmax = 175
Therefore, the Nyquist rate is given by
Ωs = 2Ωmax = 700π or fs = 2fmax = 350 Hz
The Nyquist interval is given by
1 1
Ts = = sec.
fs 350

2.2 The signal z(t) is generated by convolving two band-limited signals x1 (t) and x2 (t), that is,
z(t) = x1 (t) ∗ x2 (t)

where
X1 (Ω) = 0 for |Ω| > 1000π
X2 (Ω) = 0 for |Ω| > 2000π
Impulse-train sampling is performed on z(t) to obtain
∞
X
zs (t) = z(t)δ(t − nTs )
n=−∞
Digital Signal Processing 3

Specify the range of values for the sampling period Ts which ensures that z(t) is recoverable from
zs (t).

Solution 2.2 If
x1 (t) ←→ X1 (Ω)
x2 (t) ←→ X2 (Ω)
then
z(t) = x1 (t) ∗ x2 (t) ←→ Z(Ω) = X1 (Ω)X2 (Ω) = 0 for |Ω| > 1000π

Therefore, the maximum frequency present in the signal z(t) = x1 (t) ∗ x2 (t) is

Ωmax = 1000π

Thus, the Nyquist rate (minimum sampling frequency) is Ωs = 2Ωmax = 2000π, and hence,
2π 2π 1
Ts = = = sec.
Ωs 2000π 1000

2.3 Let x(n) = {4, 0, 2, −1, −3}. Find and sketch the following signals.
↑

1. The decimated signal xd (n) = x(2n).

n


2. The zero-interpolated signal y(n) = x 2 .
n


3. The step-interpolated signal y(n) = x 2 .

Solution 2.3 Given that

x(n) = {4, 0, 2, −1, −3}
↑

(1) The decimated signal xd (n) is given by

xd (n) = x(2n) = {4, 2, −3}

↑

(2) The zero-interpolated signal is given by

n
y(n) = x = {4, 0, 0, 0, 2, 0, −1, 0, −3}
2 ↑

(3) The step-interpolated signal is given by

n
y(n) = x = {4, 4, 0, 0, 2, 2, −1, −1, −3, −3}
2 ↑

2.4 (a) Describe the sequence of operations required to generate

(b) Let x(n) = {1, 4, 7, 10, 13}. Sketch x(n) and x n − 32 . Use step interpolation where


↑
required.
M


(c) Generalize the results of part (a) to generate x n − L from x(n). Is there any restriction
on M and L

Solution 2.4 (a)

     
interpolate by 3  n  delay by 2 n−2 decimate by 3 3n − 2 2
x(n) −→ x −→ x −→ x =x n−
n → n/3 3 n→n−2 3 n → 3n 3 3
4 Digital Signal Processing

(b) x(n) = {1, 4, 7, 10, 13}

↑
n
y(n) = x = {1, 1, 1, 4, 4, 4, 7, 7, 7, 10, 10, 10, 13, 13, 13}
3 ↑
 
n−2
w(n) = y(n − 2) = x = {0, 0, 1, 1, 1, 4, 4, 4, 7, 7, 7, 10, 10, 10, 13, 13, 13}
3 ↑
   
3n − 2 2
z(n) = w(3n) = x =x n− = {0, 1, 4, 7, 10, 13}
3 3 ↑
(c)
   
interpolate by N  n  delay by M n−M decimate by N Nn − M
x(n) −→ x −→ x −→ x
n → n/N N n→n−M N n → Nn N
 
M
=x n−
N

2.5 Each of the following sinusoids is sampled at fs = 100 Hz. Determine if aliasing has occurred
and set up
an expression for each sampled signal using a digital frequency in the principal range
|f | < 12 .
1. x1 (t) = cos(320πt)
2. x2 (t) = cos(140πt)
3. x3 (t) = sin(60πt)

Solution 2.5 (1)

x1 (t) = cos(320πt) = cos(2πfm t)
Nyquist rate = 2fm = 320
Since fs < 2fm , aliasing will occur. If the signal is sampled at fs = 100 Hz, its equivalent discrete-
time signal is given by
x1 (t)|t=nTs = x1 (nTs ) = x1 (n) = cos(320πnTs )
 
320π
= cos n
100
 
16π
x(n) = cos n
5
 
4π
= cos 4nπ − n
5
 
4π
x(n) = cos n
5
2
The frequency of the discrete-time signal x(n) is fd = 5 cycles/sample.

x2 (t) = cos(140πt) = cos(2πfm t)

(2)
Nyquist rate = 2fm = 140
Since fs < 2fm , aliasing will occur. If the signal is sampled at fs = 100 Hz, its equivalent discrete-
time signal is given by
x2 (t)|t=nTs = x2 (nTs ) = x2 (n) = cos(140πnTs )
 
140π
= cos n
100
 
7π
x(n) = cos n
5
 
3π
= cos 2nπ − n
5
 
3π
x(n) = cos n
5
Digital Signal Processing 5

3
The frequency of the discrete-time signal x(n) is fd = 10 cycles/sample.
(3)
x3 (t) = cos(60πt) = cos(2πfm t)
Nyquist rate = 2fm = 60

Since fs > 2fm , aliasing will not occur. If the signal is sampled at fs = 100 Hz, its equivalent
discrete-time signal is given by

x3 (t)|t=nTs = x3 (nTs ) = x3 (n) = cos(60πnTs )

 
60π
= cos n
100
 
3π
x(n) = cos n
5
3
The frequency of the discrete-time signal x(n) is fd = 10 cycles/sample.

2.6 Suppose that a discrete-time sequence x(n) is bandlimited so that

X(ejω ) = 0 0.3π < |ω| < π

This sequence is then sampled to form the sequence

y(n) = x(nN )

where N is an integer. Find the largest value for N for which x(n) may be uniquely recovered
from x(n).

Solution 2.6 (See Section 2.12) Define a discrete-time sampled signal xs (n) which is equal to the
multiplication of the original signal x(n) with a sampling signal (an impulse train) s(n), i.e.,

xs (n) = x(n)s(n)
where
∞
X
s(n) = δ(n − kN )
k=−∞
Therefore, we have
∞
X
xs (n) = x(n)s(n) = x(n) δ(n − kN )
k=−∞

∞
X
xs (n) = x(kN )δ(n − kN )
k=−∞

Using the multiplication property, we obtain

Z
jω jω 1jω
Xs (e ) = X(e ) ∗ S(e ) = X(ejθ )S(ej(ω−θ) )dθ
2π 2π

The DTFT of the sampling signal s(n) (an impulse train) is

∞
2π X
S(ejω ) = δ(ω − kωs )
N
k=−∞
6 Digital Signal Processing

2π
where ωs = N is the sampling frequency. Therefore,
Z ∞
1 2π X
Xs (ejω ) = X(ejθ ) δ(ω − θ − kωs )dθ
2π 2π N
k=−∞
N −1 Z
1 X
= X(ejθ )δ(ω − θ − kωs )dθ
N 2π
k=0
N −1
1 X 
Xs (ejω ) = X(ejθ )

N θ=ω−kωs
k=0
N −1
1 X  j(ω−kωs ) 
Xs (ejω ) = X e
N
k=0

Now consider the decimated or down sampled signal (See Section 2.12.1) y(n) = xs (nN ) =
x(nN ). The DTFT of the decimated signal y(n) is given by
∞
X
Y (ejω ) = y(n)e−jωn
n=−∞
X∞
= xs (nN )e−jωn
n=−∞
m
Let m = nN , or equivalently n = N, we can write
∞
m
X
Y (ejω ) = xs (m)e−jω N
m=−∞
N −1
ω 1 X  j (ω−2πk) 
Y (ejω ) = Xs (ej N ) = X e N
N
k=0

Y (e ) is periodic with period 2π. It is evident that we can recover X(ejω ) from Y (ejω ) if
jω

ωo N ≤ (2π − ωo N )
π
N≤
ωo
π
N≤
0.3π
10
N≤
3

2.7 Find two different continuous-time signals that will produce the sequence
x(n) = cos(0.15πn)
when sampled with a sampling frequency of 8 KHz.

Solution 2.7 Given that x(n) = cos(0.15πn) = cos 2π 0.15

2 n = cos(2πfd n) and fs = 8KHz.

f0
fd =
fs
f0 = fd fs
0.15
f0 = 8 = 6KHz
2
The aliased frequency components of frequency f0 sampled at fs is given by
fk = f0 + kfs , k = ±1, ±2, · · ·
f1 = 6 + 8 = 14KHz fork = 1
Digital Signal Processing 7

Therefore the two different continuous-time signals that will produce the sequence

x(n) = cos(0.15πn)

when sampled with a sampling frequency of 8 KHz are given by

x1 (t) = cos(2πf0 t) = cos(12πt)

x2 (t) = cos(2πf1 t) = cos(28πt)
2.8 An analog signal
x(t) = sin(480πt) + 3 sin(720πt)
is sampled 600 times per second.
(a) Determine the Nyquist sampling rate for x(t).
(b) Determine the folding frequency.
(c) What are the frequencies, in radians, in the resulting discrete-time signal x(n)?

Solution 2.8 (a) Given that

x(t) = sin(480πt) + 3 sin(720πt)
Comparing with

x(t) = A1 sin(Ω1 t) + A2 sin(Ω2 t)

we have

Ω1 = 480π and Ω2 = 720π

Maximum frequency present in the signal x(t) is

Ωmax = Ω2 = 720π or fmax = 360

Therefore, the Nyquist rate is given by

Ωs = 2Ωmax = 1440π or fs = 2fmax = 720 Hz

(b) The folding frequency is given by

fs 720
ffold = = = 360 Hz
2 2
(c) Given that fs = 600. The resulting discrete-time signal is given by

x(n) = x(t)|t=nTs = x(nTs ) = sin(480πnTs ) + 3sin(720πnTs )

   
480π 720π
= sin n + 3sin n
600 600
   
4π 6π
x(n) = sin n + 3sin n
5 5
   
4π 4π
= sin n + 3sin 2nπ − n
5 5
     
4π 4π 4π
x(n) = sin n − 3sin n = −2sin n
5 5 5