1

2014/1/25 1 Zhongguo Liu_Biomedical Engineering_Shandong Univ.

Biomedical Signal processing
Chapter 4 Sampling of Continuous-
Time Signals
Zhongguo Liu
Biomedical Engineering
School of Control Science and Engineering, Shandong University
2
Chapter 4: Sampling of
Continuous-Time Signals
4.0 Introduction
4.1 Periodic Sampling
4.2 Frequency-Domain Representation of
Sampling
4.3 Reconstruction of a Bandlimited Signal
from its Samples
4.4 Discrete-Time Processing of
Continuous-Time signals
3
4.0 Introduction
Continuous-time signal processing can be
implemented through a process of
sampling, discrete-time processing, and
the subsequent reconstruction of a
continuous-time signal.
f=1/T: sampling frequency
| | ( )
,
c
x n x nT
n
=
< < ( ) s rad T
s
/ , 2t = O
T: sampling period
4

=
=
n
nT t ) ( o
4.1 Periodic
Sampling
Continuous-
time signal
T:
sampling
period
5
4.2 Frequency-Domain Representation of Sampling
[ ] ( ) | ( )
c t nT c
x n x t x nT
=
= =
Tsample period; fs=1/T:sample rate
s=2/T: sample rate
( ) ( )

=
=
n
nT t t s o
( ) ( ) ( ) ( ) ( ) ( ) ( )
n
s c c
n
c
x t t nT x t x t s t x nT t nT o o

= =
= = =

( ) ( )
2
s
k
S j k
T
t
o

=
O = O O

( ) ( )
1
c s
k
X j k
T

=
= O O

( ) ( ) ( ) ( ) ( )
( ) ( ) ( ) ( )
1
( )
2
1 2 1
1
*
2
( ) ( )
2
c
s c s
k
s c
c
k
S j X j d
k X j
X j X j S j
d k X j d
T T
u u u
t
t
o u u u o u u u
t
t

= =
= O
= O O = O
O = O O
O
}

} }
Representation of
in terms of
( )
jw
e X
( ) O j X
s
6
Tsample period; fs=1/T:sample rate;s=2/T:sample rate
( ) ( )

=
=
n
nT t t s o
( ) ( )
2
s
k
S j k
T
t
o

=
O = O O

( ) ( )
2
s
k
S j k
T
t
o

=
O = O O

s
jk t
k
n
a e

O
=
=

s(t)T
1
s
jk t
n
e
T

O
=
=

2 ( )
s
jk t F
s
e k to
O
O O
-T
1
t
T 0
( ) s t


/ 2
/ 2
1 1
( )
s
T
jk t
k
T
a t e dt
T T
o
+
O

= =
}
e
0
( ) S jO


2
T
t
2
T
t 2
T
t

7
( ) ( ) ( ) ( ) ( ) ( ) ( )
s c c c
n n
x t x t s t x t t nT x nT t nT o o

= =
= = =

( )
c
k
j Tn
x nT e

=
O
=

( ) ( )
( )
j t
s c
n
X j x nT t nT e dt o

O
=

O =

}
[ ] ( )
c
x n x nT =
( )
( )
j
c
n
j n
X e x nT e
e e

=

DTFT
Representation of in terms of ,
( )
jw
e X ( ) O j X
s
( )
c
X jO
( )
j T
X e
O
=
T e O =
rad
, rad/s
2
s
T
t
O =
( ) ( ) ( )
1
s c s
k
X j X j k
T

=
= O = O O

, rad/s
8
( ) ( )
j j T
X e X e
e O
=
1 2
( )
c
k
j
k
X X j
T T T
e
e
e t

=
| |
| |
=
| |
\ .
\ .

( )
0,
c
if X j
T
t
O = O>
1
( )
c
j
then X X j
T T
e
e
e
| |
=
|
\ .
( ) ( ) ( )
1
s c s
k
X j X j k
T

=
= O = O O

DTFT
e t <
Representation of in terms of ,
( )
jw
e X ( ) O j X
s
( )
c
X jO
Continuous FT
/ T e O=
9
Nyquist Sampling Theorem
Let be a bandlimited signal
with . Then is
uniquely determined by its
samples ,if


The frequency is commonly referred as
the Nyquist frequency.
The frequency is called the Nyquist rate.

( ) t X
c
( )
N c
for j X O > O = O 0
( ) t X
c
| | ( ) , 2 , 1 , 0 , = = n nT x n x
c
N s
T
O > = O 2
2t
N
O
N
O 2
10
N N s
O < O O
N N s
O > O O
aliasing frequency
T O = e
2 /
s
O
t
t 2
No aliasing
aliasing

=
O O = O
k
s c s
k j X
T
j X )) ( (
1
) (

=
= O
=
O =
k
c
T s
j
T k j X
T
j X e X
) / ) 2 ( (
1
| ) ( ) (
/
t e
e
e
frequency spectrum of ideal
sample signal
11
4.3 Reconstruction of a Bandlimited Signal
from its Samples
( ) | | ( )

=
=
n
r r
nT t h n x t x
( )
( )
T t
T t
t h
r
t
t sin
=
| |
( )
( )
sin
n
t nT T
x n
t nT T
t
t

=
(

=

( ) ( ) ( )
T j
r r
e X j H j X
O
O = O
Gain: T
12
4.4 Discrete-Time Processing of
Continuous-Time signals
| | ( ) nT x n x
c
=
( )
1 2
jw
c
k
X e
w k
X j
T T T
t

=
=
| |
| |

| |
\ .
\ .

( ) | |
( ) ( )
( )

=
n
r
T nT t
T nT t
n y t y
t
t sin
( ) ( ) ( )
( )

< O
=
O = O
O
O
otherwise
T
e TY
e Y j H j Y
T j
T j
r r
, 0
,
t
( )
jw
H e
13
C/D Converter
Output of C/D Converter


| | ( ) nT x n x
c
=
( )

=
|
|
.
|

\
|
|
.
|

\
|
=
k
c
jw
T
k
T
w
j X
T
e X
t 2 1
14
D/C Converter
Output of D/C Converter

( ) | |
( ) ( )
( )

=
n
r
T nT t
T nT t
n y t y
t
t sin
( ) ( ) ( )
( )

< O
=
O = O
O
O
otherwise
T
e TY
e Y j H j Y
T j
T j
r r
, 0
,
t
( )
,
0,
r
T
H j
T
otherwise
t

O <
O =

15
( )
( )
,
0,
j T
c
H e X j
T
T
t
t
O

O O <

=

O >

4.4.1 Linear Time-Invariant

Discrete-Time Systems
( ) O j X
c
( )
jw
e X
( ) O j Y
r
( )
jw
e Y
( )
jw
e H
( ) ( ) ( )
jw jw jw
e X e H e Y =
( ) ( )
( ) ( )
( )
( )
1 2
j T j T
r r
j T
r c
k
Y j H j H e X e
k
H j H e X j
T T
t
O O

O
=
O = O
| |
| |
= O O
| |
\ .
\ .

16
Linear and Time-Invariant
Linear and time-invariant system behavior
depends on two factors:
First, the discrete-time system must be
linear and time invariant.
Second, the input signal must be
bandlimited, and the sampling rate must
be high enough to satisfy Nyquist
Sampling Theorem.
17
( ) ( ) ( ) ( )
( ) ( )

=
O
O O
|
|
.
|

\
|
|
.
|

\
|
O O =
O = O
k
c
T j
r
T j T j
r r
T
k
j X
T
e H j H
e X e H j H j Y
t 2 1
( )
0 ,
c
If X j for T t O = O >
( )
( ) ( )

> O
< O O
= O
O
T
T
j X e H
j Y
c
T j
r
t
t
, 0
,
( ) ( ) ( ) O O = O j X j H j Y
c eff r
( )
( )

> O
< O
= O
O
T
T
e H
j H
T j
eff
t
t
, 0
,
( )
,
0,
r
T
H j
T
otherwise
t

O <
O =

18
4.4.2 Impulse Invariance
Given:
( )
jw
H e Design:
( ) ( )
( )
,
0,
j T
c eff
H e
T
H j H j
T
t
t
O

O <

O = O =

O >

| | ( )
c
h n Th nT =
| |
h n
( )
c
h nT
( ) O j X
c
( )
jw
e X
( ) O j Y
r
( )
jw
e Y
( )
jw
e H
impulse-invariant version of the continuous-time system
( )
,
c
H jO
19
4.4.2 Impulse Invariance
Two constraints
( )
( )
,
j
c
H e H j T
e
e e t = <
( ) T j H
that such chosen is T
c
t > O = O , 0
1.
2.
| | ( )
c
h n T Th n =
The discrete-time system is called an impulse-
invariant version of the continuous-time system
e t <
| | ( )
c
h n h nT =
( )
1
c
j
X j
T T
X e
e
e
| |
=
|
\ .
| | ( )
c
h n T Th n =
( )
c
j
X X j
T
e
e
e
| |
=
|
\ .
/
C
T t O <