3.

Fourier Analysis of Continuous and

Discrete-Time Signals and Systems

Faculty of Electrical and Computer Engineering

Communication Systems Engineering

9/23/2022 BY Demis 1 Signal and system Analysis

 Content

Introduction to Fourier Analysis

Fourier Series Representation of Periodic Signals

Fourier Transforms and Inverse Transforms

The Sampling Theorem

9/23/2022 BY Demis 2 Signal and system Analysis

 Objectives
Objectives
❖Describe periodic signals using Fourier series representation.
❖Describe and take analysis Fourier transformed periodic and non-
periodic signals
❖ Apply impulse train for sampling

9/23/2022 BY Demis 3 Signal and system Analysis

 3.1 Introduction to Fourier Analysis
We have seen that a signal can generally be represented as a linear
combination of shifted impulse (or sample) functions.
Here, we will see how a signal can be represented as a linear combination of
complex exponential functions, provided certain conditions are satisfied.
Why is this useful?
This is because the output of an LTI system due to a complex exponential
input is the same complex exponential multiplied by a (possibly complex) gain
factor.
A signal for which the system output is a (possibly complex) constant times
the input is referred to as an eigenfunction of the system, and the amplitude
factor is referred to as the system's eigenvalue.
9/23/2022 BY Demis 4 Signal and system Analysis
 3.1 Introduction to Fourier Analysis
If we give a complex exponential as an input to a system, the
output of the system is equal with the input multiplied with a value called the
frequency response for the system.
Mathematically we can express the above statement as follows.
In continuous time we know h(t) the impulse response of the system. We give
the system the input signal x(t) = ejωt. We want to compute the output y(t)
given as:
a function f is an eigenfunction of D
if it satisfies the equation
Df = λf
Where λ is a constant and D is
linear operator

9/23/2022 BY Demis 5 Signal and system Analysis

 3.1 Introduction to Fourier Analysis

Where

H(jω) is a complex constant which represents the frequency response of the

system.
Then the constant H(jω) for a specific value of s(jω) is then the eigenvalue
associated with the eigenfunction ejωt
9/23/2022 BY Demis 6 Signal and system Analysis
 3.1 Introduction to Fourier Analysis
In discrete time: We know h[n] the impulse response of the system. We give
the system the input signal x[n] = ejΩn.We want to compute the output y[n].

9/23/2022 BY Demis 7 Signal and system Analysis

 3.1 Introduction to Fourier Analysis

Where

H(ejΩ) represents the frequency response of the system.

The general notion of describing complicated signals as a function of
frequency is commonly encountered in music.
The study of signals and systems using sinusoidal representations is termed
Fourier analysis.

9/23/2022 BY Demis 8 Signal and system Analysis

 3.1 Introduction to Fourier Analysis
The output of the system is thus a complex sinusoid of the same frequency as
the input, multiplied by the complex number H(jω).
Note that H(jω) is a function of only the frequency ω and not the time t and is
termed the frequency response of the continuous-time system.
Example: the impulse response of the system relating the input voltage to
the voltage across the capacitor in Figure below is derived previously and is:

Find an expression for the frequency response, and plot the magnitude and
phase response.

9/23/2022 BY Demis 9 Signal and system Analysis

 3.1 Introduction to Fourier Analysis

Figure: RC circuit for this example

Solution:

9/23/2022 BY Demis 10 Signal and system Analysis

 3.1 Introduction to Fourier Analysis
The magnitude response is

while the phase response is

By representing arbitrary signals as weighted superposition's of

eigenfunctions, we transform the operation of convolution to multiplication.
To see this, consider expressing the input to an LTI system as the weighted
sum of M complex sinusoids

9/23/2022 BY Demis 11 Signal and system Analysis

 3.1 Introduction to Fourier Analysis
If 𝑒 𝑗𝜔𝑘 𝑡 is an eigenfunction of the system with eigenvalue H(jωk ), then each
term in the input, produces an output term akH(jωk) 𝑒 𝑗𝜔𝑘 𝑡 . Hence, we
express the output of the system as

The operation of convolution, b(t) * x(t), becomes multiplication, akH(jωk)

because x(t) is expressed as a sum of eigenfunctions.
This property is a powerful motivation for representing signals as weighted
superpositions of complex sinusoids.
If x(t) is a continuous-time signal of fundamental period T, we represent x(t)
by the FS
9/23/2022 BY Demis 12 Signal and system Analysis
 3.1 Introduction to Fourier Analysis
where ω0 = 2π/T is the fundamental frequency of x(t).
The frequency of the kth sinusoid is kω0 , and each sinusoid has a common
period T.
A sinusoid whose frequency is an integer multiple of a fundamental frequency
is said to be a harmonic of the sinusoid at the fundamental frequency.
Thus, 𝑒 𝑗𝑘𝜔0 𝑡 is the kth harmonic of. 𝑒 𝑗𝜔0 𝑡 .
Similarly for periodic DT signal x[n]

9/23/2022 BY Demis 13 Signal and system Analysis

 3.1 Introduction to Fourier Analysis
We seek weights or coefficients A[k] such that x[n] and x(t) are good
approximations to x[n] and x(t), respectively.
The DTFS and FS coefficients to be given in the following sections minimize
the MSE.

9/23/2022 BY Demis 14 Signal and system Analysis

 3.2 Fourier Series Representation of Periodic Signals
i. Fourier series representation of CT periodic signals
Let x(t) be a CT periodic signal with fundamental period T.
Then if certain conditions (Dirichlet conditions) are satisfied, we can
represent x(t) as a Fourier series (FS) (an infinite sum of complex
exponentials):

where the (possibly complex) Fourier coefficients {X[k]} are given by

9/23/2022 BY Demis 15 Signal and system Analysis

 3.2 Fourier Series Representation of Periodic Signals

The FS coefficients are known as a frequency-domain representation of x (t)

because each FS coefficient is associated with a complex sinusoid of a different
frequency.
The above equations are referred to as the synthesis and analysis equations
respectively.
Note:The Dirichlet conditions are as follows:
Condition 1: Over one period, x(t) must be absolutely integrable; that is,

9/23/2022 BY Demis 16 Signal and system Analysis

 3.2 Fourier Series Representation of Periodic Signals

Condition 2. In any finite interval of time, x(t) is of bounded variation; that

is, there are no more than a finite number of maxima and minima during any
single period of the signal.

Figure:- signal that violates condition 2

Condition 3. In any finite interval of time, there are only a finite number of
discontinuities.
An example of a function that violates Condition 3 is illustrated in Figure
below.
9/23/2022 BY Demis 17 Signal and system Analysis
 3.2 Fourier Series Representation of Periodic Signals

Figure:- signal that violates condition 3 and decreases by a factor of 2

Convergence is very important in the FS representation of a signal.
Notes also:
Using Euler’s relationship, we can also express the FS representation as an
infinite sum of sine and cosine terms.

9/23/2022 BY Demis 18 Signal and system Analysis

 3.2 Fourier Series Representation of Periodic Signals

Given x(t), we can determine {𝑋[𝑘]}

𝑘=−∞ ; conversely, we can

reconstruct x(t) from {𝑋[𝑘]} . 𝑘=−∞

{𝑋[𝑘]}
𝑘=−∞ give the frequency-domain description of the signal and

are called its spectral coefficients.

A periodic signal x(t) of fundamental period T has components at
frequencies 0, ±2π/T, ±4π/T, . ., i.e. at multiples of the fundamental
frequency ω0 = 2π/T or f0 = 1/T.
The component at frequency n f0 is called the nth harmonic.
Conjugate symmetry property: If x(t) is real, then X[-k] = X[k]∗.

9/23/2022 BY Demis 19 Signal and system Analysis

 3.2 Fourier Series Representation of Periodic Signals

Example: Determine the FS coefficients for the signal x(t) depicted in Figure
below.

Figure:Time-domain signal
Solution: The period of x(t) is T = 2, so ω0 = 2π/2 = π. On the interval 0
< t < 2, one period of x(t) is expressed as x(t) = e-2t, so from FS coefficient
eq.

9/23/2022 BY Demis 20 Signal and system Analysis

 3.2 Fourier Series Representation of Periodic Signals

We evaluate the integral to obtain

, since e-jk2π = 1

9/23/2022 BY Demis 21 Signal and system Analysis

 3.2 Fourier Series Representation of Periodic Signals

Check plot of the magnitude spectrum |X[k]| and the phase spectrum
arg{X[k]}. (Page 217 of Haykin)
Magnitude of X[k] is known as the magnitude spectrum of x(t),
while the phase of X[k] is known as the phase spectrum of x(t).
Example: determine the FS representation of the signal

using the method of inspection.

Solution: The fundamental period of x(t) is T = 4. Hence, ω0 = 2π/4 =
π/2, and thus

9/23/2022 BY Demis 22 Signal and system Analysis

 3.2 Fourier Series Representation of Periodic Signals

Using Euler’s formula to expand the cosine yields

Thus comparing these with FS eq.

The magnitude and phase spectra are depicted in Figure below.

9/23/2022 BY Demis 23 Signal and system Analysis

 3.2 Fourier Series Representation of Periodic Signals

Example(Inverse FS): Find the time-domain signal x(t) corresponding to

the FS coefficients

Assume that the fundamental period is T = 2.

Solution: Substituting the values given for X[k] and ω0 = 2π/T = π into
previous eq. yields

The result of summing both infinite geometric series is

9/23/2022 BY Demis 24 Signal and system Analysis

 3.2 Fourier Series Representation of Periodic Signals

Example: Periodic square wave shown below.

Figure: Periodic square wave

We would like to determine its Fourier or

spectral coefficients {𝑋[𝑘]}

𝑘=−
.

9/23/2022 BY Demis 25 Signal and system Analysis

 3.2 Fourier Series Representation of Periodic Signals

In this case , we have

9/23/2022 BY Demis 26 Signal and system Analysis

 3.2 Fourier Series Representation of Periodic Signals

As expected, the amplitudes of the FS coefficients have even symmetry; in this

case, the phases are 0.
Example:T1 = T/4, then , the average valu e is
9/23/2022 BY Demis 27 Signal and system Analysis
 3.2 Fourier Series Representation of Periodic Signals

and

9/23/2022 BY Demis 28 Signal and system Analysis

 3.2 Fourier Series Representation of Periodic Signals

Discrete Time Periodic Signals: DTFS

The DTFS representation of a periodic signal x[n] with fundamental period N
and fundamental frequency 0 = 2/N is given by

where the (possibly complex) Fourier coefficients {X[k]}, also known as

spectral coefficients, are given by

We say that x[n] and X[Jb] are a DTFS pair and denote this relationship as

9/23/2022 BY Demis 29 Signal and system Analysis

 3.2 Fourier Series Representation of Periodic Signals

The notation 𝑘=<𝑁> is used to indicate that the summation is over N

consecutive integers, starting with any value of k.
The above two equations are referred to as the synthesis and analysis equations
respectively.
There are similarities as we ll as differences from the CT case.
Remarks:
The synthesis equation in the DT case involves a finite sum, in sharp contrast
to the CT case which involves an infinite sum. Hence, unlike the CT case,
there are no convergence issues in the DT case.
Unlike the CT case, X[k] = X[k+N].
9/23/2022 BY Demis 30 Signal and system Analysis
 3.2 Fourier Series Representation of Periodic Signals

Example: Find the frequency-domain representation (DTFS coefficients) of

the signal depicted in Fig. below.

Solution: The signal has period N = 5, so 0 = 2/5. Also, the signal has odd
symmetry, so we sum over n = -2 to n = 2:

Using the values of x[n], we get

9/23/2022 BY Demis 31 Signal and system Analysis

 3.2 Fourier Series Representation of Periodic Signals

For one period the DTFS coefficients X[k], k = -2 to k = 2,

9/23/2022 BY Demis 32 Signal and system Analysis

 3.2 Fourier Series Representation of Periodic Signals

Figure:- Magnitude and phase of the DTFS coefficients

9/23/2022 BY Demis 33 Signal and system Analysis
 3.2 Fourier Series Representation of Periodic Signals

Example: Determine the DTFS coefficients of x[n] given by:

Solution: The period of x[n] is N = 6. We expand the cosine by using Euler’s

formula:

Again form DTFS

9/23/2022 BY Demis 34 Signal and system Analysis

 3.2 Fourier Series Representation of Periodic Signals

Equating terms in Eq. (3.13) with those in Eq. (3.14) having equal
frequencies, k/3, gives

Determine the time-domain signal x[n] from the DTFS coefficients depicted
in Fig. below.
Solution: The DTFS coefficients have period 9, so 0 = 2/9. It is
convenient to evaluate over the interval k = - 4 to k = 4 to obtain

9/23/2022 BY Demis 35 Signal and system Analysis

 3.2 Fourier Series Representation of Periodic Signals

Figure: Magnitude and phase of DTFS coefficients for Example

9/23/2022 BY Demis 36 Signal and system Analysis
 3.3 Fourier Transforms and Inverse Transforms

i. Continuous-Time Nonperiodic Signals:The Fourier Transform

The Fourier transform (FT) is used to represent a continuous-time
nonperiodic signal as a superposition of complex sinusoids.
Thus, the FT representation of a continuous-time signal involves an integral
over the entire frequency interval; that is,

where

9/23/2022 BY Demis 37 Signal and system Analysis

 3.3 Fourier Transforms and Inverse Transforms

is the FT of the signal x(t).

We say that x(t) and X(jω) are an FT pair and write

The transform X(jω) describes the signal x(t) as a function of frequency ω and
is termed the frequency-domain representation of x(t).
Equation (3.35) is termed the inverse FT, since it maps the frequency-domain
representation X(jω) back into the time domain.
The FT is used to analyze the characteristics of continuous-time systems and
the interaction between continuous-time signals and systems.
The FT is also used to analyze interactions between discrete- and continuous-
time signals, such as occur in sampling.
9/23/2022 BY Demis 38 Signal and system Analysis
 3.3 Fourier Transforms and Inverse Transforms

The integrals in Eqs. (3.35) and (3.36) may not converge for all functions x(t)
and X(jω).
Convergence is guaranteed t all values of t except those corresponding to
discontinuities if x(t) satisfies the Dirichlet conditions for nonperiodic signals:
x(t) is absolutely integrable:

x(t) has a finite number of maxima, minima, and discontinuities in any finite
interval.
The size of each discontinuity is finite.

9/23/2022 BY Demis 39 Signal and system Analysis

 3.3 Fourier Transforms and Inverse Transforms

Example: Find the FT of x(t) = e-atu(t), shown in Fig. below.

Solution: The FT does not converge for a < 0, since x(t) is not absolutely
integrable; that is,
For a > 0, we have

the magnitude and phase of X(jω) are respectively given by

9/23/2022 BY Demis 40 Signal and system Analysis

 3.3 Fourier Transforms and Inverse Transforms

And

Example: Consider the rectangular pulse depicted in Fig. below and defined
as

Find the FT of x(t).

Solution: The rectangular pulse x(t) is absolutely integrable, provided that T0
< .We thus have

9/23/2022 BY Demis 41 Signal and system Analysis

 3.3 Fourier Transforms and Inverse Transforms

For ω = 0, the integral simplifies to 2T0 . L’Hopital’s rule straightforwardly

shows that
Thus, we usually write

X(jω) is depicted in Fig. below.

9/23/2022 BY Demis 42 Signal and system Analysis

 3.3 Fourier Transforms and Inverse Transforms

Exercise: Example 3.26(p-246), Example 3.27 and Example 3.28 (p-247)

and Problem 3.15 (p-248)

9/23/2022 BY Demis 43 Signal and system Analysis

 3.3 Fourier Transforms and Inverse Transforms

ii. Discrete-time Fourier transform (DTFT) representation of DT

aperiodic signals
The DTFT is used to represent a discrete-time nonperiodic signal as a
superposition of complex sinusoids.
The DTFT would involve a continuum of frequencies on the interval - < 
< , where  has units of radians.
The DTFT representation of a time-domain signal involves an integral over
frequency, namely,

9/23/2022 BY Demis 44 Signal and system Analysis

 3.3 Fourier Transforms and Inverse Transforms

Where

is the DTFT of the signal x[n].

We say that X(en) and x[n] are a DTFT pair and write

X(en) describes the signal x[n] as a function of a sinusoidal frequency  and

is termed the frequency-domain representation of x[n].
Equation (3.31) is usually termed the inverse DTFT, since it maps the
frequency-domain representation back into the time domain.
9/23/2022 BY Demis 45 Signal and system Analysis
 3.3 Fourier Transforms and Inverse Transforms

The infinite sum in Eq. (3.32) converges if x[n] has finite duration and is finite
valued.
If x[n] is of infinite duration, then the sum converges only for certain classes
of signals, i.e:

Other notation:

9/23/2022 BY Demis 46 Signal and system Analysis

 3.3 Fourier Transforms and Inverse Transforms

Example: Find the DTFT of the sequence x[n] = anu[n].

Solution: we have

This sum diverges for |α| > 1. For |α| < 1, we have the convergent
geometric series

If a is real valued, we may expand the denominator of Eq. (3.33) using Euler’s
formula to obtain
9/23/2022 BY Demis 47 Signal and system Analysis
 3.3 Fourier Transforms and Inverse Transforms

From this form, we see that the magnitude and phase spectra are given by

And

As with the other Fourier representations, the magnitude spectrum of a signal

is the magnitude of X(ejΩ) plotted as a function of Ω. The phase spectrum is
the phase of X(ejΩ) plotted as a function of Ω.
9/23/2022 BY Demis 48 Signal and system Analysis
 3.3 Fourier Transforms and Inverse Transforms

Example: DTFT of a Rectangular Pulse Let

as depicted in Fig. below. Find the DTFT of x[n].

Solution:We substitute for x[n] to obtain

Now we perform the change of variable m = n + M, obtaining

9/23/2022 BY Demis 49 Signal and system Analysis

 3.3 Fourier Transforms and Inverse Transforms

The expression for X(ejΩ) when Ω ≠ 0, ±2π, ±4π,. . . , may be simplified by

symmetrizing the powers of the exponential in the numerator and
denominator as follows:

9/23/2022 BY Demis 50 Signal and system Analysis

 3.3 Fourier Transforms and Inverse Transforms

We may now write X(ejΩ) as a ratio of sine functions by dividing the

numerator and denominator by 2j to obtain

L’Hopital’s rule gives

hence, rather than write X(ejΩ) as two forms dependent on the value of ft, we
simply write

with the understanding that X(ejΩ) for Ω ≠ 0, ±2π, ±4π,. . . , is obtained as a

limit.
9/23/2022 BY Demis 51 Signal and system Analysis
 3.3 Fourier Transforms and Inverse Transforms
Example: Inverse DTFT of a Rectangular Spectrum Find the inverse DTFT
of

which is depicted in Fig. below.

Solution: First, note that X(ejΩ) is specified only for -π < Ω < π. This is all
that is needed, however, since X(ejΩ) is always 2π-periodic and the inverse
DTFT depends solely on the values in the interval -π < Ω < π.
Thus we can have:
9/23/2022 BY Demis 52 Signal and system Analysis
 3.3 Fourier Transforms and Inverse Transforms

For n = 0, the integrand is unity and we have x[0] = W/π. Using L’Hopital’s
rule, we easily show that

and thus we usually write

with the understanding that the value at n = 0 is obtained

as the limit, we may also write
9/23/2022 BY Demis 53 Signal and system Analysis
 3.3 Fourier Transforms and Inverse Transforms

A graph depicting x[n] versus time n is given in Fig. below.

Figure: Inverse DTFT in the time domain.

Exercise: Find the inverse DTFT of X(ejΩ) = δ(Ω), -π < Ω < π.

Figure: Unit impulse in the frequency domain,

9/23/2022 BY Demis 54 Signal and system Analysis
 3.3 Fourier Transforms and Inverse Transforms

9/23/2022 BY Demis 55 Signal and system Analysis

 3.3 Fourier Transforms and Inverse Transforms

Properties of Fourier Representations

1. Linearity and Symmetry Properties
All four Fourier representations satisfy the linearity property:

In the cases of the FS and DTFS, the signals being summed are assumed to
have the same fundamental period.
Example: Suppose z(t) is the periodic signal depicted in Fig. (a) below. Use
the linearity property to determine the FS coefficients Z[ k].
9/23/2022 BY Demis 56 Signal and system Analysis
 3.3 Fourier Transforms and Inverse Transforms

Solution:Write z(t) as a sum of signals; that is,

where x(t) and y(t) are depicted in Figs. (b) and (c) below, respectively.Then:

The linearity property implies that

9/23/2022 BY Demis 57 Signal and system Analysis
 3.3 Fourier Transforms and Inverse Transforms

9/23/2022 BY Demis 58 Signal and system Analysis

 3.3 Fourier Transforms and Inverse Transforms

Symmetry Properties: Real and Imaginary Signals

for both real and imaginary valued signals, this property can be summarized as
follows.

9/23/2022 BY Demis 59 Signal and system Analysis

 3.3 Fourier Transforms and Inverse Transforms

Symmetry Properties: Even and Odd Signals

Suppose that x(t) is real valued and has even symmetry.
If x(t) is real and even, then X(j) is real and the following hold true.

Similarly, we may show that if x(t) is real and odd, then X(j) is imaginary
and
Identical symmetry relationships hold for all four Fourier representations.
Real and even time-domain signals have real and even frequency-domain
representations, and real and odd time-domain signals have imaginary and odd
frequency domain representations.
9/23/2022 BY Demis 60 Signal and system Analysis
 3.3 Fourier Transforms and Inverse Transforms

2. Convolution Properly
The convolution property is a consequence of complex sinusoids being
eigenfunctions of LTI systems.
With the convolution property, we may analyze the input-output behavior of a
linear system in the frequency domain by multiplying transforms instead of
convolving time signals.
a. Convolution of Nonperiodic Signals
The convolution of two nonperiodic continuous-time signals x(t) and h(t)
corresponds to multiplication of their Fourier transforms, H(j) and X(j),
in the frequency domain.
9/23/2022 BY Demis 61 Signal and system Analysis
 3.3 Fourier Transforms and Inverse Transforms

Now we express x(t — t) in terms of its FT:

Substituting this expression into the convolution integral yields

We recognize the inner integral over  as the FT of h(), or H(j). Hence,

y(t) may be rewritten as
9/23/2022 BY Demis 62 Signal and system Analysis
 3.3 Fourier Transforms and Inverse Transforms

and we identify H(j)X(j) as the FT of y(t), i.e:

Example: Let x(t) = (l/(t)) sin(t) be the input to a system with impulse
response h(t) = (l/(t)) sin(2t). Find the output y(t) = x(t) * h(t).
Solution: This problem is extremely difficult to solve in the time domain.
However, it is simple to solve in the frequency domain if we use the
convolution property.
9/23/2022 BY Demis 63 Signal and system Analysis
 3.3 Fourier Transforms and Inverse Transforms

Thus:

And

and we conclude that y(t) = (1/(t)) sin(t).

9/23/2022 BY Demis 64 Signal and system Analysis

 3.3 Fourier Transforms and Inverse Transforms

A similar property holds for convolution of discrete-time nonperiodic signals:

If

b. Filtering
The multiplication that occurs in the frequency-domain representation gives
rise to the notion of filtering.
The term “filtering” implies that some frequency components of the input are
eliminated while others are passed by the system unchanged.
We may describe systems in terms of the type of filtering that they perform
on the input signal.
9/23/2022 BY Demis 65 Signal and system Analysis
 3.3 Fourier Transforms and Inverse Transforms

A low-pass filter attenuates high-frequency components of the input and

passes the lower frequency components.
A highpass filter attenuates low frequencies and passes the high frequencies.
A band-pass filter passes signals within a certain frequency band and
attenuates signals outside that band.

Figure: ideal low-pass, high-pass, and band-pass filters

9/23/2022 BY Demis 66 Signal and system Analysis

 3.3 Fourier Transforms and Inverse Transforms

The convolution property implies that the frequency response of a system

may be expressed as the ratio of the FT or DTFT of the output to that of the
input.
Specifically, we may write, for a continuous-time system.

and, for a discrete-time system,

Hence, if the input spectrum is nonzero at all frequencies, the frequency

response of a system may be determined from knowledge of the input and
output spectra.
9/23/2022 BY Demis 67 Signal and system Analysis
 3.3 Fourier Transforms and Inverse Transforms
Example: The output of an LT1 system in response to an input x(t) = e-2tu(t)
is y(t) = e-tu(t). Find the frequency response and the impulse response of this
system.
Solution:We take the FT of x(t) and y(t), obtaining

And

Now we use the definition

9/23/2022 BY Demis 68 Signal and system Analysis

 3.3 Fourier Transforms and Inverse Transforms
This equation may be rewritten as

We take the inverse FT of each term to obtain the impulse response of the
system:
we can recover the input of the system from the output as

And

9/23/2022 BY Demis 69 Signal and system Analysis

 3.3 Fourier Transforms and Inverse Transforms

where Hinv(j) = l / H(j) and Hinv(ej) = l/H(ej) are the frequency

responses of the respective inverse systems. An inverse system is also known
as an equalizer, and the process of recovering the input from the output is
known as equalization.
c. Convolution of Periodic Signals
We define the periodic convolution of two continuous-time signals x(t) and
z(t), each having period T, as

where the symbol  denotes that integration is performed over a single

period of the signals involved.
9/23/2022 BY Demis 70 Signal and system Analysis
 3.3 Fourier Transforms and Inverse Transforms
Substituting the FS representation of z{t) into the convolution integral leads
to the property

Again, we see that convolution in time transforms to multiplication of the

frequency domain representations.
The discrete-time convolution of two N-periodic sequences x[n] and z[n] is
defined as

Substitution of the DTFS representation for z[n] results in the property

9/23/2022 BY Demis 71 Signal and system Analysis

 3.3 Fourier Transforms and Inverse Transforms

The convolution properties of all four Fourier representations are

summarized in table below.

3. Dffierentiation and Integration Properties

As differentiation and integration are operations that apply to continuous
functions, we may consider the effect with respect to time for a continuous-
time signal or with respect to frequency in the FT and DTFT
9/23/2022 BY Demis 72 Signal and system Analysis
 3.3 Fourier Transforms and Inverse Transforms

a. Differentiation in Time
Consider the effect of differentiating a nonperiodic signal x(t). First, recall
that x(t) and its FT, X(j), are related by

Differentiating both sides of this equation with respect to t yields

from which it follows that

That is, differentiating a signal in the time domain corresponds to multiplying

its FT by j in the frequency domain.
9/23/2022 BY Demis 73 Signal and system Analysis
 3.3 Fourier Transforms and Inverse Transforms

Example:The differentiation property implies that

Verify this result by differentiating and taking the FT of the result.

Solution: Using the product rule for differentiation, we have

Taking the FT of each term and using linearity, we may write

9/23/2022 BY Demis 74 Signal and system Analysis

 3.3 Fourier Transforms and Inverse Transforms

Example:The differentiation property implies that

Similarly, if x(t) is a periodic signal, then we have the FS representation

Differentiating both sides of this equation gives

and thus we conclude that

9/23/2022 BY Demis 75 Signal and system Analysis

 3.3 Fourier Transforms and Inverse Transforms

b. Differentiation in Frequency
FT is:

we differentiate both sides of this equation with respect to  and obtain

from which it follows that

The operation of differentiation does not apply to discrete-valued quantities,

and thus a frequency-differentiation property for the FS or DTFS does not
exist.
9/23/2022 BY Demis 76 Signal and system Analysis
 3.3 Fourier Transforms and Inverse Transforms

A frequency-differentiation property does exist for the DTFT. By definition,

Differentiation of both sides of this expression with respect to frequency leads

to the property

c. Integration
We may integrate with respect to time in both the FT and FS and with respect
to frequency in the FT and DTFT.
We limit our consideration here to integrating nonperiodic signals with
respect to time.
9/23/2022 BY Demis 77 Signal and system Analysis
 3.3 Fourier Transforms and Inverse Transforms

c. Integration
We may integrate with respect to time in both the FT and FS and with respect
to frequency in the FT and DTFT.
We limit our consideration here to integrating nonperiodic signals with
respect to time.
And for such signal the integration property is:

9/23/2022 BY Demis 78 Signal and system Analysis

 3.3 Fourier Transforms and Inverse Transforms

4.Time- and Frequency-Shift Properties

a.Time-Shift Property
Let z(t) = x(t - t0) be a time-shifted version of x(t). The goal is to relate the
FT of z(t) to the FT of x(t).We have

Next, we effect the change of variable r = t - t0, obtaining

9/23/2022 BY Demis 79 Signal and system Analysis

 3.3 Fourier Transforms and Inverse Transforms

The result of time-shifting the signal x(t) by t0 is to multiply the FT X(jω) by

𝑒 −𝑗𝜔𝑡0 .
Note that |Z(jω)| = |X(jω)| and arg{Z(jω)} = arg{X(jω)}-ωto .
A similar property holds for the other three Fourier representations, as
indicated in Table 3.7.

9/23/2022 BY Demis 80 Signal and system Analysis

 3.3 Fourier Transforms and Inverse Transforms

Example: Use the FT of the rectangular pulse x(t) depicted in Fig. (a) to
determine the FT of the time-shifted rectangular pulse z(t) depicted in Fig. (b)
below.

Figure: Application of the time-shift property

Solution: First, we note that z(t) = x(t – T1), so the time-shift property
implies that Z(jω) = 𝑒 −𝑗𝜔𝑇1 X(jω).

9/23/2022 BY Demis 81 Signal and system Analysis

 3.3 Fourier Transforms and Inverse Transforms

We obtain:

Thus, we have

b. Frequency-Shift Property
In the previous subsection, we considered the effect of a time shift on the
frequency-domain representation.
In the current subsection, we consider the effect of a frequency shift on the
FT time-domain signal.
Our aim is to express the inverse FT of Z(jω) = X(j(ω-γ)) in terms of x(t).
9/23/2022 BY Demis 82 Signal and system Analysis
 3.3 Fourier Transforms and Inverse Transforms

Let

By the definition of the inverse FT, we have

We effect the substitution of variables η = ω - γ, obtaining

A frequency shift corresponds to multiplication in the time domain by a

complex sinusoid whose frequency is equal to the shift.
9/23/2022 BY Demis 83 Signal and system Analysis
 3.3 Fourier Transforms and Inverse Transforms

Since all the Fourier representations are based on complex sinusoids, they all
share this property, as summarized in Table 3.8.

The other observation is that the frequency-shift property is the “dual” of the
time-shift property.
Example: Use the frequency-shift property to determine the FT of the
complex sinusoidal pulse
9/23/2022 BY Demis 84 Signal and system Analysis
 3.3 Fourier Transforms and Inverse Transforms

The other observation is that the frequency-shift property is the “dual” of the
time-shift property.

Solution: We may express z(t) as the product of a complex sinusoid ej10t and a
rectangular pulse

Thus:

and employing the frequency-shift property

9/23/2022 BY Demis 85 Signal and system Analysis
 3.3 Fourier Transforms and Inverse Transforms

we obtain

5. Finding Inverse Fourier Transforms by Using Partial-Fraction

Expansions
Suppose X(jω) is expressed as a ratio of polynomials in jω.

Then we may determine the inverse FT of such ratios by using a partial-

fraction expansion
We assume that M < N. If M > N, then we may use long division to express
X(jω) in the form
9/23/2022 BY Demis 86 Signal and system Analysis
 3.3 Fourier Transforms and Inverse Transforms

Let the roots of the denominator polynomial A(jω) be dk , k = 1, 2, . . . , N.

These roots are found by replacing jω with a generic variable v and
determining the roots of the polynomial

We may then write

Assuming that all the roots dk , k = 1, 2, . . . , N, are distinct, we may write

9/23/2022 BY Demis 87 Signal and system Analysis

 3.3 Fourier Transforms and Inverse Transforms

where the coefficients Ck = 1, 2, . . . , N, are determined either by solving a

system of linear equations or by the method of residues.
Then we can have the FT pair

Assuming that the real part of each dk, k = 1 , 2, ..., N, is negative, we use
linearity to write

9/23/2022 BY Demis 88 Signal and system Analysis

 3.3 Fourier Transforms and Inverse Transforms

Exercise: Use partial-fraction expansions to determine the time-domain

signals corresponding to the following FT’S:

Answers:

9/23/2022 BY Demis 89 Signal and system Analysis

 3.3 Fourier Transforms and Inverse Transforms

Suppose again X(e jΩ) is given by a ratio of polynomials in ejΩ ; that is,

We factor the denominator polynomial as

In this case we replace ejΩ with the generic variable v and find the dk from the
roots of the polynomial

Assuming that M < N and all the dk are distinct, we may express X(e jΩ) as

9/23/2022 BY Demis 90 Signal and system Analysis

 3.3 Fourier Transforms and Inverse Transforms

Since

the linearity property implies that

Example: Find the inverse DTFT of

Solution:The roots of the polynomial

Solving this gives d1= -1/2 and d2 = 1/3.

9/23/2022 BY Demis 91 Signal and system Analysis
 3.3 Fourier Transforms and Inverse Transforms

We seek coefficients C1 and C2 such that

Using the method of residues

Hence:

9/23/2022 BY Demis 92 Signal and system Analysis

 3.3 Fourier Transforms and Inverse Transforms

6. Multiplication Property
If x(t) and z(t) are nonperiodic signals, then we wish to express the FT of the
product y(t) = x(t)z(t) in terms of the FT of x(t) and z(t).
We represent x(t) and z(t) in terms of their respective FT’s as

The product term, y(t), may thus be written in the form

9/23/2022 BY Demis 93 Signal and system Analysis

 3.3 Fourier Transforms and Inverse Transforms

Now we effect the change of variable η = ω — v to obtain

Hence, we identify this convolution, scaled by 1/(2π), asY(jω); that is,

Where

Similarly, if x[n] and z[n] are discrete-time nonperiodic signals, then the DTFT
of the product y[n] = x[n]z[n] is given by the convolution of the their DTFT’s
and multiplication by 1/(2π); that is,

9/23/2022 BY Demis 94 Signal and system Analysis

 3.3 Fourier Transforms and Inverse Transforms

Where

The multiplication property for periodic signals is analogous to that for

nonperiodic signals.And given by:

Where

The multiplication property for discrete-time periodic signals is

9/23/2022 BY Demis 95 Signal and system Analysis

 3.3 Fourier Transforms and Inverse Transforms

Where

is the periodic convolution of DTFS coefficients. Again, all three time-domain

signals have a common fundamental period N.
The multiplication properties are summarized for all four Fourier
representations in Table 3.9.

9/23/2022 BY Demis 96 Signal and system Analysis

 3.3 Fourier Transforms and Inverse Transforms

7. Scaling Properties
The effect of scaling the time variable on the frequency-domain
representation of a signal is given as:

Hence, scaling the signal in time introduces the inverse scaling in the
frequency-domain representation and an amplitude change.
Example: Let the rectangular pulse

Find the FT of the scaled rectangular pulse

9/23/2022 BY Demis 97 Signal and system Analysis

 3.3 Fourier Transforms and Inverse Transforms

Solution:

Note that y(t) = x(t/2). Hence, application of the scaling property with a =
1/2 gives:

The FS coefficients of x(t) and x(at) are identical; the scaling operation simply
changes the harmonic spacing from ω0 to aω0.

9/23/2022 BY Demis 98 Signal and system Analysis

 3.3 Fourier Transforms and Inverse Transforms

The scaling operation has a slightly different character in discrete time than in
continuous time.
First of all, z[n] = x[pn] is defined only for integer values of p. Second, if |p|
> 1, then the scaling operation discards information, since it retains only
every pth value of x[n].
This loss of information prevents us from expressing the DTFT or DTFS of
z[n] in terms of the DTFT or DTFS of x[n] in a manner similar to the way we
did for the continuous-time results derived earlier.

9/23/2022 BY Demis 99 Signal and system Analysis

 3.3 Fourier Transforms and Inverse Transforms

8. Parseval Relationships
The Parseval relationships state that the energy or power in the time-domain
representation of a signal is equal to the energy or power in the frequency-
domain representation.
Hence, energy and power are conserved in the Fourier representation. And
given by:

9/23/2022 BY Demis 100 Signal and system Analysis

 3.3 Fourier Transforms and Inverse Transforms

Hence, the energy in the time-domain representation of the signal is equal to

the energy in the frequency-domain representation, normalized by 2π.
The quantity |X(jω)|2 plotted against ω is termed the energy spectrum of
the signal
Example: Using Parseval’s theorem Calculate the Energy in a Signal

Solution: Using the DTFT Parseval relationship in, we have

where

9/23/2022 BY Demis 101 Signal and system Analysis

 3.3 Fourier Transforms and Inverse Transforms
Since

it follows that

Reading Assignment
9. Time—Bandwidth Product
10 Duality

9/23/2022 BY Demis 102 Signal and system Analysis

 3.4 The Sampling Theorem
Question:Why would we want to sample a continuous-time signal?
Processing of a CT signal using a discrete-time system.
Question: Under what conditions can we recover (or reconstruct) the CT
signal exactly from its samples?
The answer is provided by the sampling theorem.
Impulse train sampling
The impulse train or comb function

shown in the figure below

9/23/2022 BY Demis 103 Signal and system Analysis
 3.4 The Sampling Theorem

has as FT

9/23/2022 BY Demis 104 Signal and system Analysis

 3.4 The Sampling Theorem
The parameter T is the sampling period and its reciprocal 1/T is the sampling
rate.
Consider the system shown below:

Example signals:

9/23/2022 BY Demis 105 Signal and system Analysis

 3.4 The Sampling Theorem
We can write

We are interested in examining the FT of xs(t). To derive it, we proceed as

follows:
From the multiplication property, we have

Using the fact that

9/23/2022 BY Demis 106 Signal and system Analysis
 3.4 The Sampling Theorem

and the distributive property of ∗, we can write

Finally, using the replication property of δ(·), we obtain

9/23/2022 BY Demis 107 Signal and system Analysis

 3.4 The Sampling Theorem
and the distributive property of ∗, we can writeFinally, using the replication
property of δ(·), we obtain
We conclude that the spectrum of xs(t) consists of scaled (by 1/T) replicas of
the original spectrum X(jω) spaced by the sampling rate.
Reconstruction of the CT signal from its samples
A sketch of an example spectrum X(jω) is shown below, where ωM represents
the highest frequency component present in x(t).

9/23/2022 BY Demis 108 Signal and system Analysis

 3.4 The Sampling Theorem
A corresponding sketch for Xs(jω) is shown below. For this sketch, it is
assumed that ωs > 2 ωM.

From the above sketch, we observe the following:

9/23/2022 BY Demis 109 Signal and system Analysis

 3.4 The Sampling Theorem
As long as there is no overlap between the replicas, i.e. ωs > 2 ωM, we can
reconstruct x(t) by passing the samples through an ideal low-pass filter (LPF)
with gain T and cut-off frequency of say ωs/2.
We can now state the sampling theorem: A band-limited signal x(t) with
spectrum X(jω) = 0 for |ω| > ωM is completely
determined by its samples taken at a rate ωs > 2 ωM, where 2ωM is referred
to as the Nyquist rate.
If ωs < 2ωM, there will be overlap between the replicas and in general it will
not be possible to recover x(t) exactly.

9/23/2022 BY Demis 110 Signal and system Analysis

 3.4 The Sampling Theorem
When overlap occurs, we say that there is aliasing: a high frequency
component appears as a lower frequency component in the reconstructed
signal.
Typical sampling rates:
Telephone quality voice: 8 kHz
CD audio: 44.1 kHz

9/23/2022 BY Demis 111 Signal and system Analysis

 Refereces
❖ Oppenheim (MIT Professor., Signals and Systems 2nd edition
❖ Simon Haykin and Barry Van Veen Signals and Systems, Second Edition, John,
Wiley & Sons, 2003.

9/23/2022 BY Demis 112 Signal and system Analysis

 Thank You

BY Demis
9/23/2022 113
Signal and system Analysis