Chapter 6

THE SAMPLING PROCESS

6.4 Poisson’s Summation Formula
6.5 Impulse-Modulated Signals

Copyright
c 2005 Andreas Antoniou
Victoria, BC, Canada
Email: aantoniou@ieee.org

July 14, 2018

Frame # 1 Slide # 1 A. Antoniou Digital Signal Processing – Secs. 6.4, 6.5

Introduction

Various time-domain and frequency-domain relationships exist

between continuous-time and discrete-time signals.

Frame # 2 Slide # 2 A. Antoniou Digital Signal Processing – Secs. 6.4, 6.5

Introduction

Various time-domain and frequency-domain relationships exist

between continuous-time and discrete-time signals.

These relationships are developed by defining a special class of

signals known as impulse-modulated signals which comprise
sequences of continuous-time impulse functions.

Frame # 2 Slide # 3 A. Antoniou Digital Signal Processing – Secs. 6.4, 6.5

Introduction

Various time-domain and frequency-domain relationships exist

between continuous-time and discrete-time signals.

These relationships are developed by defining a special class of

signals known as impulse-modulated signals which comprise
sequences of continuous-time impulse functions.

Impulse-modulated signals are essentially continuous-time signals

but simultaneously they are also sampled signals.

Therefore, on the one hand, they have Fourier transforms and, on

the other, they can be represented by z transforms.

Consequently, impulse-modulated signals can serve as a

mathematical bridge between continuous-time and discrete-time
signals that facilitates the derivations of the various relationships
between the two classes of signals.

Frame # 2 Slide # 4 A. Antoniou Digital Signal Processing – Secs. 6.4, 6.5

Impulse-Modulated Signals

An impulse modulated-signal, denoted as x̂(t), can be

generated by sampling a continuous-time signal x(t) using an
ideal impulse modulator.
Impulse modulator

x(t) ˆ
x(t) (a)

c(t)

Frame # 3 Slide # 5 A. Antoniou Digital Signal Processing – Secs. 6.4, 6.5

Impulse-Modulated Signals Cont’d

An impulse modulator is characterized by the equation

x̂(t) = c(t)x(t)
where c(t) is a carrier given by
∞
X
c(t) = δ(t − nT )
n=−∞

Frame # 4 Slide # 6 A. Antoniou Digital Signal Processing – Secs. 6.4, 6.5

Impulse-Modulated Signals Cont’d

An impulse modulator is characterized by the equation

x̂(t) = c(t)x(t)
where c(t) is a carrier given by
∞
X
c(t) = δ(t − nT )
n=−∞

Hence
∞
X
x̂(t) = x(t) δ(t − nT )
n=−∞

Frame # 4 Slide # 7 A. Antoniou Digital Signal Processing – Secs. 6.4, 6.5

Impulse-Modulated Signals Cont’d

An impulse modulator is characterized by the equation

x̂(t) = c(t)x(t)
where c(t) is a carrier given by
∞
X
c(t) = δ(t − nT )
n=−∞

Hence
∞
X
x̂(t) = x(t) δ(t − nT )
n=−∞

From the properties of the unit impulse function, we get

∞
X
x̂(t) = x(nT )δ(t − nT )
n=−∞

Frame # 4 Slide # 8 A. Antoniou Digital Signal Processing – Secs. 6.4, 6.5

Impulse-Modulated Signals Cont’d

The input and output of an impulse modulator are as follows:

x(t)

x(kT)
(b)

t
kT

× c(t)

(c)
1
t
kT

= ˆ
x(t)

x(kT) (d)

t
kT

Frame # 5 Slide # 9 A. Antoniou Digital Signal Processing – Secs. 6.4, 6.5

Relationship between Impulse-Modulated and
Discrete-Time Signals

Impulse-modulated signals are sequences of continuous-time

impulses.

Frame # 6 Slide # 10 A. Antoniou Digital Signal Processing – Secs. 6.4, 6.5

Relationship between Impulse-Modulated and
Discrete-Time Signals

Impulse-modulated signals are sequences of continuous-time

impulses.
They can be converted to discrete-time signals by replacing
impulses by numbers.

Frame # 6 Slide # 11 A. Antoniou Digital Signal Processing – Secs. 6.4, 6.5

Relationship between Impulse-Modulated and
Discrete-Time Signals

Impulse-modulated signals are sequences of continuous-time

impulses.
They can be converted to discrete-time signals by replacing
impulses by numbers.
On the other hand, impulse-modulated signals can be obtained
from discrete-time signals by replacing numbers by impulses.

Frame # 6 Slide # 12 A. Antoniou Digital Signal Processing – Secs. 6.4, 6.5

Relationship between Impulse-Modulated and
Discrete-Time Signals Cont’d

ˆ
x(t)

x(kT) (d)

t
kT

x(nT)

x(kT)
(e)

nT
kT

Frame # 7 Slide # 13 A. Antoniou Digital Signal Processing – Secs. 6.4, 6.5

Relationship between Fourier Transform and Z Transform

An impulse-modulated signal is both a continuous-time as well

as a sampled signal, as was stated earlier, and this dual
personality will immediately prove very useful.

Frame # 8 Slide # 14 A. Antoniou Digital Signal Processing – Secs. 6.4, 6.5

Relationship between Fourier Transform and Z Transform

An impulse-modulated signal is both a continuous-time as well

as a sampled signal, as was stated earlier, and this dual
personality will immediately prove very useful.
As a continuous-time signal, an impulse-modulated signal has
a Fourier transform given by
∞
X ∞
X
X̂ (jω) = F x(nT )δ(t − nT ) = x(nT )Fδ(t − nT )
n=−∞ n=−∞
∞
X
= x(nT )e −jωnT
n=−∞

Frame # 8 Slide # 15 A. Antoniou Digital Signal Processing – Secs. 6.4, 6.5

Relationship between Fourier Transform and Z Transform
Cont’d

··· ∞
X
X̂ (jω) = x(nT )e −jωnT
n=−∞

Therefore, from the definition of the z transform we note that


X̂ (jω) = XD (z) jωT

z=e

where
XD (z) = Zx(nT ) (A)

Frame # 9 Slide # 16 A. Antoniou Digital Signal Processing – Secs. 6.4, 6.5

Relationship between Fourier Transform and Z Transform
Cont’d

··· ∞
X
X̂ (jω) = x(nT )e −jωnT
n=−∞

Therefore, from the definition of the z transform we note that


X̂ (jω) = XD (z) jωT

z=e

where
XD (z) = Zx(nT ) (A)

In effect, the Fourier transform of impulse-modulated signal x̂(t) is

numerically equal to the z transform of the corresponding
discrete-time signal x(nT ) evaluated on the unit circle |z| = 1.

Frame # 9 Slide # 17 A. Antoniou Digital Signal Processing – Secs. 6.4, 6.5

Relationship between Fourier Transform and Z Transform
Cont’d

··· ∞
X
X̂ (jω) = x(nT )e −jωnT
n=−∞

Therefore, from the definition of the z transform we note that


X̂ (jω) = XD (z) jωT

z=e

where
XD (z) = Zx(nT ) (A)

In effect, the Fourier transform of impulse-modulated signal x̂(t) is

numerically equal to the z transform of the corresponding
discrete-time signal x(nT ) evaluated on the unit circle |z| = 1.

In other words, the frequency spectrum of x̂(t) is equal to that of

x(nT ).

Frame # 9 Slide # 18 A. Antoniou Digital Signal Processing – Secs. 6.4, 6.5

Example

The continuous-time signal



 0 for t < −3.5 s

1 for −3.5 ≤ t < −2.5



x(t) = 2 for −2.5 ≤ t < 2.5

1 for 2.5 ≤ t ≤ 3.5




0 for t > 3.5


is subjected to impulse modulation.

Find the frequency spectrum of x̂(t) in closed form assuming a
sampling frequency of 2π rad/s.

Frame # 10 Slide # 19 A. Antoniou Digital Signal Processing – Secs. 6.4, 6.5

Example Cont’d

Solution The frequency spectrum of an impulse-modulated signal, x̂(t),

can be readily obtained by evaluating the z transform of x(nT ) on the
unit circle of the z plane.
The impulse-modulated version of x(t) can be expressed as

x̂(t) = δ(t + 3T ) + 2δ(t + 2T ) + 2δ(t + T ) + 2δ(0)

+2δ(t − T ) + 2δ(t − 2T ) + δ(t − 3T )

where T = 1 s.
A corresponding discrete-time signal can be obtained by replacing
impulses by numbers as

x(nT ) = δ(nT + 3T ) + 2δ(nT + 2T ) + 2δ(nT + T ) + 2δ(0)

+2δ(nT − T ) + 2δ(nT − 2T ) + δ(nT − 3T )

Hence XD (z) = Zx(t) = z 3 + 2z 2 + 2z 1 + 2 + 2z −1 + 2z −2 + z −3

Frame # 11 Slide # 20 A. Antoniou Digital Signal Processing – Secs. 6.4, 6.5

Example Cont’d

···
XD (z) = Zx(t) = z 3 + 2z 2 + 2z 1 + 2 + 2z −1 + 2z −2 + z −3

Since the frequency spectrum of an impulse-modulated signal is

given by
X̂ (jω) = XD (e jωT )
we get

X̂ (jω) = XD (e jωT )
= (e j3ωT + e −j3ωT ) + 2(e j2ωT + e −j2ωT )
+ 2(e jωT + e −jωT ) + 2
= 2 cos 3ωT + 4 cos 2ωT + 4 cos ωT + 2

Frame # 12 Slide # 21 A. Antoniou Digital Signal Processing – Secs. 6.4, 6.5

Example

The continuous-time signal

x(t) = u(t)e −t sin 2t

is subjected to impulse modulation.

Find the frequency spectrum of x̂(t) in closed form assuming a
sampling frequency of 2π rad/s.

Frame # 13 Slide # 22 A. Antoniou Digital Signal Processing – Secs. 6.4, 6.5

Example Cont’d

Solution A discrete-time signal can be readily derived from x(t) by

replacing t by nT as
1 j2nT
x(nT ) = u(nT )e −nT sin 2nT = u(nT )e −nT × − e −j2nT


e
2j
1 nT (−1+j2)
− e nT (−1−j2)


= u(nT ) e
2j

Since T = 2π/ωs = 1 s, the table of z transforms gives

 
1 z z
XD (z) = −
2j z − e −1+j2 z − e −1−j2

and after some manipulation

ze −1 sin 2
XD (z) =
z 2 − 2ze −1 cos 2 + e −2

Frame # 14 Slide # 23 A. Antoniou Digital Signal Processing – Secs. 6.4, 6.5

Example Cont’d

···
ze −1 sin 2
XD (z) =
z 2 − 2ze −1 cos 2 + e −2
Since the frequency spectrum of an impulse-modulated signal is
given by
X̂ (jω) = XD (e jωT )
we get

e jω−1 sin 2
X̂ (jω) = XD (e jωT ) =
e 2jω − 2e jω−1 cos 2 + e −2

Frame # 15 Slide # 24 A. Antoniou Digital Signal Processing – Secs. 6.4, 6.5

Poisson’s Summation Formula

As may be expected, the spectrum of a discrete-time signal

must be related to that spectrum of the continuous-time
signal from which it was derived.

Frame # 16 Slide # 25 A. Antoniou Digital Signal Processing – Secs. 6.4, 6.5

Poisson’s Summation Formula

As may be expected, the spectrum of a discrete-time signal

must be related to that spectrum of the continuous-time
signal from which it was derived.
This relationship can be established by using Poisson’s
summation formula.

Frame # 16 Slide # 26 A. Antoniou Digital Signal Processing – Secs. 6.4, 6.5

Poisson’s Summation Formula Cont’d

Consider a signal x(t) with a Fourier transform X (jω).

Poisson’s summation formula states that
∞ ∞
X 1 X
x(t + nT ) = X (jnωs )e jnωs t
n=−∞
T n=−∞

where ωs = 2π/T .

Frame # 17 Slide # 27 A. Antoniou Digital Signal Processing – Secs. 6.4, 6.5

Poisson’s Summation Formula Cont’d

Consider a signal x(t) with a Fourier transform X (jω).

Poisson’s summation formula states that
∞ ∞
X 1 X
x(t + nT ) = X (jnωs )e jnωs t
n=−∞
T n=−∞

where ωs = 2π/T .
If t = 0 and x(t) is a two-sided signal, we have
∞ ∞
X 1 X
x(nT ) = X (jnωs ) (B)
n=−∞
T n=−∞

Frame # 17 Slide # 28 A. Antoniou Digital Signal Processing – Secs. 6.4, 6.5

Poisson’s Summation Formula Cont’d

··· ∞
X
x(t + nT ) =
1 X
∞
X (jnωs )e jnωs t
n=−∞
T n=−∞

If t = 0 and x(t) is a right-sided signal, i.e., x(t) = 0 for

t < 0, then
∞ ∞
X x(0+) 1 X
x(nT ) = + X (jnωs )
2 T n=−∞
n=0

where
x(0−) + x(0+) x(0+)
lim x(t) = =
t→0 2 2

Frame # 18 Slide # 29 A. Antoniou Digital Signal Processing – Secs. 6.4, 6.5

Poisson’s Summation Formula Cont’d

··· ∞
X
x(t + nT ) =
1 X
∞
X (jnωs )e jnωs t
n=−∞
T n=−∞

If t = 0 and x(t) is a right-sided signal, i.e., x(t) = 0 for

t < 0, then
∞ ∞
X x(0+) 1 X
x(nT ) = + X (jnωs )
2 T n=−∞
n=0

where
x(0−) + x(0+) x(0+)
lim x(t) = =
t→0 2 2
Note: In Fourier analysis, the value of a time function at a
discontinuity is always taken to be the average of the left and
right limits (see textbook).

Frame # 18 Slide # 30 A. Antoniou Digital Signal Processing – Secs. 6.4, 6.5

Poisson’s Summation Formula Cont’d
1.0

0.5

x(t)
0

−0.5

−1.0
0 T 2T 3T 4T 5T
(a)

1.5

1.0
|X(jω)|

0.5

0
ω
−3ωs −2ωs −ωs 0 ωs 2ωs 3ωs

4.0

2.0
arg x(jω), rad/s

−2.0

−4.0
ω
−3ωs −2ωs −ωs 0 ωs 2ωs 3ωs
(b)

Frame # 19 Slide # 31 A. Antoniou Digital Signal Processing – Secs. 6.4, 6.5

Spectral Relationship between Discrete-Time and
Continuous-Time Signals

Given a continuous-time signal x(t) with a Fourier transform

X (jω), then from the frequency-shifting theorem we have
x(t)e −jω0 t ↔ X (jω0 + jω)

Frame # 20 Slide # 32 A. Antoniou Digital Signal Processing – Secs. 6.4, 6.5

Spectral Relationship between Discrete-Time and
Continuous-Time Signals

Given a continuous-time signal x(t) with a Fourier transform

X (jω), then from the frequency-shifting theorem we have
x(t)e −jω0 t ↔ X (jω0 + jω)
From Poisson’s summation formula, we get
∞ ∞
X
−jω0 nT 1 X
x(nT )e = X (jω0 + jnωs )
n=−∞
T n=−∞

where ωs = 2π/T and if we now replace ω0 by ω, we deduce

the important relationship
∞ ∞
X 1 X
x(nT )e −jωnT = X (jω + jnωs )
n=−∞
T n=−∞

Frame # 20 Slide # 33 A. Antoniou Digital Signal Processing – Secs. 6.4, 6.5

Spectral Relationship Cont’d

··· ∞ ∞
X 1 X
x(nT )e −jωnT = X (jω + jnωs )
n=−∞
T n=−∞

It was shown earlier that

∞
X
X̂ (jω) = XD (e jωT
)= x(nT )e −jωnT
n=−∞

∞
1 X
and hence X̂ (jω) = XD (e jωT ) = X (jω + jnωs )
T n=−∞

Therefore, the frequency spectrum of the impulse-modulated signal

x̂(t) is numerically equal to the frequency spectrum of discrete-time
signal x(nT ) and the two can be uniquely determined from the
frequency spectrum of the continuous-time signal x(t), namely,
X (jω).

Frame # 21 Slide # 34 A. Antoniou Digital Signal Processing – Secs. 6.4, 6.5

Spectral Relationship Cont’d

As is to be expected, X̂ (jω) is a periodic function of ω with

period ωs since the frequency spectrum of discrete-time
signals is periodic.

Frame # 22 Slide # 35 A. Antoniou Digital Signal Processing – Secs. 6.4, 6.5

Spectral Relationship Cont’d

As is to be expected, X̂ (jω) is a periodic function of ω with

period ωs since the frequency spectrum of discrete-time
signals is periodic.
To check this out, we can replace jω by jω + jmωs in
∞
1 X
X̂ (jω) = X (jω + jnωs )
T n=−∞

to obtain
∞
1 X
X̂ (jω + jmωs ) = X [jω + j(m + n)ωs ]
T n=−∞
∞
1 X
= X (jω + jn0 ωs ) = X̂ (jω)
T
n0 =−∞

Frame # 22 Slide # 36 A. Antoniou Digital Signal Processing – Secs. 6.4, 6.5

Spectral Relationship Cont’d

For a right-sided signal, x(t) = 0 for t ≤ 0−, and hence the

impulse-modulated signal assumes the form
∞
X
x̂(t) = x(nT )δ(t − nT )
n=0

where x(0) ≡ x(0+).

Frame # 23 Slide # 37 A. Antoniou Digital Signal Processing – Secs. 6.4, 6.5

Spectral Relationship Cont’d

For a right-sided signal, x(t) = 0 for t ≤ 0−, and hence the

impulse-modulated signal assumes the form
∞
X
x̂(t) = x(nT )δ(t − nT )
n=0

where x(0) ≡ x(0+).

The Fourier transform of the signal is given by
∞
X
X̂ (jω) = x(nT )e −jωnT = XD (e jωT )
n=0

Frame # 23 Slide # 38 A. Antoniou Digital Signal Processing – Secs. 6.4, 6.5

Spectral Relationship Cont’d

For a right-sided signal, x(t) = 0 for t ≤ 0−, and hence the

impulse-modulated signal assumes the form
∞
X
x̂(t) = x(nT )δ(t − nT )
n=0

where x(0) ≡ x(0+).

The Fourier transform of the signal is given by
∞
X
X̂ (jω) = x(nT )e −jωnT = XD (e jωT )
n=0

Thus Poisson’s summation formula gives

∞
x(0+) 1 X
X̂ (jω) = XD (e jωT ) = + X (jω + jnωs ) (C)
2 T n=−∞

Frame # 23 Slide # 39 A. Antoniou Digital Signal Processing – Secs. 6.4, 6.5

Spectral Relationship Cont’d

··· ∞
jωT x(0+) 1 X
X̂ (jω) = XD (e )= + X (jω + jnωs ) (C)
2 T n=−∞

By letting jω = s and e sT = z, Eq. (C) can be expressed in

the s domain as
∞
x(0+) 1 X
X̂ (s) = XD (z) = + X (s + jnωs )
2 T n=−∞

where X (s) and X̂ (s) are the Laplace transforms of x(t) and
x̂(t), respectively.

Frame # 24 Slide # 40 A. Antoniou Digital Signal Processing – Secs. 6.4, 6.5

Spectral Relationship Cont’d

··· ∞
jωT x(0+) 1 X
X̂ (jω) = XD (e )= + X (jω + jnωs ) (C)
2 T n=−∞

By letting jω = s and e sT = z, Eq. (C) can be expressed in

the s domain as
∞
x(0+) 1 X
X̂ (s) = XD (z) = + X (s + jnωs )
2 T n=−∞

where X (s) and X̂ (s) are the Laplace transforms of x(t) and
x̂(t), respectively.
This relationship will be used in Chap. 11 to design digital
filters based on analog filters.

Frame # 24 Slide # 41 A. Antoniou Digital Signal Processing – Secs. 6.4, 6.5

Example

Find X̂ (jω) if x(t) = cos ω0 t.

Solution From the table of Fourier transforms (Table 6.2), we have

X (jω) = F cos ω0 t = π[δ(ω + ω0 ) + δ(ω − ω0 )]

Hence Poisson’s summation formula, i.e.,

∞
jωT 1 X
X̂ (jω) = XD (e )= X (jω + jnωs )
T n=−∞

gives
∞
π X
X̂ (jω) = [δ(ω + nωs + ω0 ) + δ(ω + nωs − ω0 )]
T n=−∞

Frame # 25 Slide # 42 A. Antoniou Digital Signal Processing – Secs. 6.4, 6.5

Example Cont’d

^
| X( jω)|

−2ωs −ωs −ω0 ω0 ωs 2ωs

(a)

Frame # 26 Slide # 43 A. Antoniou Digital Signal Processing – Secs. 6.4, 6.5

Example
Find X̂ (jω) if x(t) = u(t)e −t .

Solution From the table of Fourier transforms (Table 6.2),

1
X (jω) = F[u(t)e −t ] =
1 + jω
Since x(t) = 0 for t < 0 in this case, we need to use the second form of
Poisson’s summation formula, i.e.,
∞
x(0+) 1 X
X̂ (jω) = XD (e jωT ) = + X (jω + jnωs )
2 T n=−∞

The initial-value theorem of the Laplace transform gives

s
x(0+) = lim [sX (s)] = lim =1
s→∞ s→∞ 1+s
∞
1 1 X 1
and hence X̂ (jω) = +
2 T n=−∞ 1 + j(ω + nωs )

Frame # 27 Slide # 44 A. Antoniou Digital Signal Processing – Secs. 6.4, 6.5

Example Cont’d

1.0

0.8 |X( jω)|

0.6

0.4

0.2

0
−30 −20 −10 0 10 20 30

^
3 |X( jω)|
1 1 |X( jω)| 1 |X( jω− jω )|
T |X( jω+ jωs)| T s
T
2

0
−30 −20 −10 0 10 20 ω 30
−ωs ωs ωs ωs
−
2 2
(b)

Frame # 28 Slide # 45 A. Antoniou Digital Signal Processing – Secs. 6.4, 6.5

 This slide concludes the presentation.
Thank you for your attention.

Frame # 29 Slide # 46 A. Antoniou Digital Signal Processing – Secs. 6.4, 6.5