You are on page 1of 99

Signals and Systems Analysis

Course Code: (ECEG-3091)


By : Abrham Mesfin(MSc.)
email: abrom321@gmail.com

Course title: Signals and Systems Analysis


ECTS (Credit): 6hr (4hr)
Pre-requisite: Applied Math III
Academic Year: 2019/20
CHAPTER 3
Fourier Representations for
Signals

Monday, April 25, 2022 2


Frequency Response representation
analysis of LTI Systems
Content

2.1 Introduction to Frequency Domain Analysis


2.2 Complex Sinusoids and LTI Systems
2.3 Sinusoidal Steady – State Response of LTI System
2.4 Complex Sinusoids and LTI Systems
2.4 Fourier Transform Representation of Non-periodic Signals
2.5 Fourier Series Representation of Periodic Signals
Monday, April 25, 2022 3
Introduction to Frequency response
representation of LTI system
❑ In this chapter we will discusses on frequency response analysis and
representation of DT and CT LTIS system.

❑A frequency response of a system is defined as the steady-state response of


the system to a sinusoidal input signal.

❑The sinusoidal is a unique input signal, and the resulting input signal for a
linear system, as well as signals throughout the system, is sinusoidal in the
study state; it differs from the in the amplitude and phase angle.
4
Monday, April 25, 2022
Cont.
❑we begin from consider representing a signal as a weighted
superposition of complex sinusoids.

❑ If such a signal is applied to a linear system, then the system


output is a weighted superposition of the system response to each
complex sinusoid.

• By representing signals in terms of sinusoids, we will obtain an


alternative expression for the input – output behavior of an LTI
system.
Monday, April 25, 2022 5
Cont.
• Representation of signals as superposition's of complex sinusoids
not only leads to a useful expression for the system output but also
provides a very insightful characterization of signals and systems.

Sin(𝝎𝒕) CT- LTIS h(t)

Sin(𝛀𝒏) DT- LTIS h[n]

AMiT-ECE SIGNALS AND SYSTEMS CHAPTER


Monday, April 25, 2022 6
THREE: FOURIER REPRESENTATIONS FOR SIGNALS
Sinusoidal Steady – State Response

• Sinusoidal input signals are often used to characterize the response


of a system.

• Here we examine the relationship between the impulse response and


steady state response of an LTI system to a complex sinusoidal
input.

• This relationship is easily established using convolution and complex


sinusoid input signal.
Monday, April 25, 2022 7
Cont.
• Consider the output of a discrete – time system with impulse response 𝒉[𝒏]
and unit amplitude complex sinusoidal input 𝒙 𝒏 = 𝒆𝒋𝜴𝒏 , given by:

𝒚 𝒏 = ෍ 𝒉 𝒌 𝒙[𝒏 − 𝒌]
𝒌=−∞

𝒙 𝒏 = 𝒆𝒋𝜴𝒏

AMiT-ECE SIGNALS AND SYSTEMS CHAPTER


Monday, April 25, 2022 8
THREE: FOURIER REPRESENTATIONS FOR SIGNALS
Cont.
∞ ∞

𝒚 𝒏 = ෍ 𝒉[𝒌]𝒆𝒋𝜴(𝒏−𝒌) = ෍ 𝒉[𝒌] 𝒆𝒋𝜴𝒏 𝒆−𝒋𝜴𝒌


𝒌=−∞ 𝒌=−∞
= 𝑯(𝒆𝒋𝜴 )𝒆 𝒋𝜴𝒏

• Where we have defined: ∞

𝑯 𝒆𝒋𝜴 = ෍ 𝒉[𝒌] 𝒆−𝒋𝜴𝒌


𝒌=−∞

• Hence the output of the system is a complex sinusoid of the same frequency as
the input multiplied by a complex number 𝑯(𝒆𝒋𝜴 ).

AMiT-ECE SIGNALS AND SYSTEMS CHAPTER


Monday, April 25, 2022 9
THREE: FOURIER REPRESENTATIONS FOR SIGNALS
Cont.
• The relationship is depicted in figure 1

• The quantity 𝑯(𝒆𝒋𝜴 ) is not a function of time, 𝒏, but is only a function of


frequency, 𝜴, and is termed the “frequency response of the discrete – time
system”.

• Similar results are obtained for “continuous – time systems”.


10
Monday, April 25, 2022
Cont.
• Let the impulse response of a system be 𝒉(𝒕) and the input be 𝒙 𝒕 = 𝒆.𝒋𝝎𝒕 .

• The convolution integral gives the output as:


∞ ∞

𝒚 𝒕 = න 𝒉 𝝉 𝒙(𝒕 − 𝝉) 𝒅𝝉 = න 𝒉(𝝉)𝒆𝒋𝝎(𝒕−𝝉) 𝒅𝝉
−∞ −∞

= 𝒆𝒋𝝎𝒕 න 𝒉(𝝉) 𝒆−𝒋𝝎𝝉 𝒅𝝉 = 𝑯(𝒋𝝎)𝒆𝒋𝝎𝒕


−∞

• Where we define:

𝑯 𝒋𝝎 = න 𝒉(𝝉) 𝒆−𝒋𝝎𝝉 𝒅𝝉
−∞
Monday, April 25, 2022 11
Cont.
• The output of the system is a complex sinusoid of the same frequency as the
input multiplied by the complex constant 𝑯(𝒋𝝎).

❑𝑯(𝒋𝝎) is a function of only frequency, 𝝎, and not on time,


𝒕. It is termed the frequency response of the continuous –
time system.
• An instinctive interpretation of the sinusoidal steady – state response is
obtained by writing the complex number 𝑯(𝒋𝝎) in polar form.

Monday, April 25, 2022 12


Cont.
• Recall that, if 𝒄 = 𝒂 + 𝒋𝒃 is a complex number, then we may write 𝒄 in polar
form as: 𝒄 = 𝒄 𝒆𝐚𝐫𝐠{𝒄} ,
𝒃
Where 𝒄 = 𝒂𝟐 + 𝒃𝟐 and 𝐚𝐫𝐠 𝒄 = 𝒂𝒓𝒄𝒕𝒂𝒏
𝒂
• Hence we have:

𝑯 𝒋𝝎 = 𝑯(𝒋𝝎) 𝒆𝒋𝒂𝒓𝒈{𝑯 𝒋𝝎 }

AMiT-ECE SIGNALS AND SYSTEMS CHAPTER


Monday, April 25, 2022 13
THREE: FOURIER REPRESENTATIONS FOR SIGNALS
Cont.
• Here:
❑ 𝑯(𝒋𝝎) is termed the magnitude response and
❑𝒂𝒓𝒈{𝑯 𝒋𝝎 } is termed the phase response of the system

• Substituting this polar form in output equation, the output equation will be
expressed as:

𝒚 𝒕 = 𝑯(𝒋𝝎) 𝒆𝒋(𝝎𝒕+𝒂𝒓𝒈 𝑯 𝒋𝝎 )

❑The system modify the amplitude of the input by


𝑯(𝒋𝝎) and the phase by 𝒂𝒓𝒈 𝑯 𝒋𝝎
Monday, April 25, 2022 14
Cont.
• The sinusoidal steady – state response has a similar interpretation for real.
valued sinusoids. Write:
𝒙 𝒕 = 𝑨𝒄𝒐𝒔(𝝎𝒕 + 𝜱)

𝑨 𝒋(𝝎𝒕+𝜱) 𝑨 −𝒋(𝝎𝒕+𝜱)
= 𝒆 + 𝒆
𝟐 𝟐
And use linearity to obtain the output as:
𝑨 𝒋(𝝎𝒕+𝜱+𝒂𝒓𝒈 𝑯 𝒋𝝎 )
𝑨 −𝒋(𝝎𝒕+𝜱−𝒂𝒓𝒈 𝑯 −𝒋𝝎 )
𝒚 𝒕 = 𝑯(𝒋𝝎) 𝒆 + 𝑯(−𝒋𝝎) 𝒆
𝟐 𝟐
• Assuming that 𝒉(𝒕) is real valued, then 𝑯(𝒋𝝎) possesses conjugate
symmetry, that is: 𝑯∗ 𝒋𝝎 = 𝑯(−𝒋𝝎)
Monday, April 25, 2022 15
Cont.
• This implies that, 𝑯(𝒋𝝎) is an even function of 𝝎 while 𝐚𝐫𝐠{𝑯 𝒋𝝎 } is
odd.

• Exploiting these symmetry conditions and simplifying yields:

𝒚 𝒕 = 𝑯 𝒋𝝎 𝑨𝒄𝒐𝒔(𝝎𝒕 + 𝜱 + 𝐚𝐫𝐠 𝑯 𝒋𝝎 )

• As with a complex sinusoidal input, the system modifies the input sinusoid’s
amplitude by 𝑯 𝒋𝝎 and the phase by 𝒂𝒓𝒈 𝑯 𝒋𝝎 .

• This modification is illustrated in figure 2.

Monday, April 25, 2022 16


Cont.

• Similar results are obtained for discrete – time systems using polar form of 𝑯 𝒆𝒋𝜴 .

• Specifically, if 𝒙 𝒏 = 𝒆𝒋𝜴𝒏 is the input, then:

𝒋𝜴 𝒋(𝜴𝒏+𝐚𝐫𝐠 𝑯 𝒆𝒋𝜴 )
𝒚𝒏 = 𝑯 𝒆 𝒆

Monday, April 25, 2022 17


Cont.
• Furthermore, if 𝒙 𝒏 = 𝑨𝒄𝒐𝒔(𝜴𝒏 + 𝜱) is the input to a discrete – time system with real –
valued impulse response, then:

𝒚 𝒏 = 𝑯 𝒆𝒋𝜴 𝑨𝒄𝒐𝒔(𝜴𝒏 + 𝜱 + 𝒂𝒓𝒈 𝑯 𝒆𝒋𝜴 )

• Once again, the system changes the amplitude of the sinusoidal input by 𝑯 𝒆𝒋𝜴 and its

phase by 𝒂𝒓𝒈 𝑯 𝒆𝒋𝜴 .

• The frequency response characterizes the steady – state response of the system to sinusoidal
inputs as a function of the sinusoid’s frequency.

• We say this is a steady – state response because the input sinusoid is assumed to exist
for all time and thus the system is in an equilibrium or steady – state condition.
Monday, April 25, 2022 18
Example: Response of a CT, LTI System to Sinusoidal Inputs

• Suppose that the frequency response of a CT, LTI system is defined by the
following specs:

| H ( ) |
1.5 1.5, 0    20,
| H ( ) |= 
0,   20,
0 20 

arg H ( )
 arg H ( ) = −60 , 
−60
Cont.
• If the input to the system is
x(t ) = 2cos(10t + 90 ) + 5cos(25t + 120 )
• Then the output is

y (t ) = 2 | H (10) | cos(10t + 90 + arg H (10)) +


+ 5 | H (25) | cos(25t + 120 + arg H (25)) =
= 3cos(10t + 30 )
Complex Sinusoids and LTI Systems
• The complex sinusoid input to an LTI system generates an output equal to the
sinusoidal input multiplied by the system frequency response.

• That is, in discrete – time, the input 𝒙 𝒏 = 𝒆𝒋𝜴𝒏 results in the output:

𝒚 𝒏 = 𝑯 𝒆𝒋𝜴 𝒆𝒋𝜴𝒏

Where the frequency response 𝑯 𝒆𝒋𝜴 is defined in terms of the impulse


response 𝒉[𝒏] as:

𝑯 𝒆𝒋𝜴 = ෍ 𝒉[𝒌]𝒆−𝒋𝜴𝒌 DTFS-periodic signal


𝒌=−∞
Monday, April 25, 2022 21
Cont.
• In continuous – time, the input 𝒙 𝒕 = 𝒆𝒋𝝎𝒕 results in the output:
.
𝒚 𝒕 = 𝑯(𝒋𝝎)𝒆𝒋𝝎𝒕
Where the frequency response 𝑯(𝒋𝝎) is defined in terms of the impulse response
𝒉(𝒕) as: ∞

𝑯 𝒋𝝎 = න 𝒉(𝝉)𝒆−𝒋𝝎𝝉 𝒅𝝉
−∞

• We say that the complex sinusoid 𝝍 𝒕 = 𝒆𝒋𝝎𝒕 is an eigen function of the system 𝑯
associated with the eigenvalue 𝝀 = 𝑯(𝒋𝝎) because it satisfies an eigen value problem
described by:
𝑯𝝍 𝒕 = 𝝀𝝍 𝒕
Monday, April 25, 2022 22
This eigen relation is illustrated in figure 3.

Monday, April 25, 2022 23


Cont.
• Signals that are eigen functions of systems play an important role in systems
theory.

• By representing arbitrary signals as weighted superpositions of eigen


functions, we transform the operation of convolution to one of multiplication.

• To see this, consider expressing the input to an LTI system as a weighted sum of 𝑴
complex sinusoids.

𝐱 𝐭 = ෍ 𝐚𝐤 𝐞𝐣𝛚𝐤 𝐭
𝐤=𝟏
Monday, April 25, 2022 24
Cont.
• If 𝒆𝒋𝝎𝒌 𝒕 is an eigen function of the system with eigen value of 𝑯(𝒋𝝎𝒌 ), then
each term in the input, 𝒂𝒌 𝒆𝒋𝝎𝒌 𝒕 , produces an output term, 𝒂𝒌 𝑯(𝒋𝝎𝒌 )𝒆𝒋𝝎𝒌 𝒕 .

• Hence we express the output of the system as:

𝐲 𝐭 = ෍ 𝐚𝐤 𝐇(𝐣𝛚𝐤 )𝐞𝐣𝛚𝐤 𝐭
𝐤=𝟏

Monday, April 25, 2022 25


Cont.
• The output is a weighted sum of 𝑴 complex sinusoid, with the weights,

𝒂𝒌 , modified by the system frequency response, 𝑯(𝒋𝝎𝒌 ).

• The operation of convolution, 𝐡 𝐭 ∗ 𝐱(𝐭) , becomes multiplication,

𝒂𝒌 𝑯(𝒋𝝎𝒌 )𝐞𝐣𝛚𝐤 𝐭 , because 𝒙(𝒕) is expressed as a sum of eigen functions.

❑The analogous relationship holds in the discrete – time

case.

Monday, April 25, 2022 26


Cont.
• This property is a powerful motivation for representing signals as weighted
superposition of complex sinusoids.

• In addition, the weights provide an alternative interpretation of


signals.

• Rather than describing the signal behavior as a function of time, the


weights describe the signal as a function of frequency.

AMiT-ECE SIGNALS AND SYSTEMS CHAPTER


Monday, April 25, 2022 27
THREE: FOURIER REPRESENTATIONS FOR SIGNALS
Fourier Representation for Four Signal Classes
• There are four distinct Fourier representations, each applicable to a
different class of signals.

• These four classes are defined by the periodicity properties of a signal


and whether it is continuous or discrete. See in Table 1.

• Periodic signals have Fourier Series representations.


❑The Fourier Series (FS) applied to CT periodic signals and
❑The Discrete – Time Fourier Series (DTFS) applied to DT periodic
signals.
28
Cont.

• Non – periodic signals have Fourier Transform representations.


❑ If the signal is CT, the representation is termed Fourier
Transform (FT) and
❑ If it is DT, then it is Discrete Time Fourier Transform
(DTFT)

Monday, April 25, 2022 29


Cont.

Monday, April 25, 2022 30


Discrete – Time Non – Periodic Signals:
The Discrete – Time Fourier Transform (DTFT)
Derivation:
• We develop the DTFT from the DTFS by describing a non – periodic signal as
the limit of a periodic signal whose period, N, approaches infinity.

• For this approach to be meaningful, we assume that the non – periodic signal
is represented by a single period of the periodic signal that is centered on the
origin, and that the limit as N approaches infinity is taken in a symmetric
manner.
Monday, April 25, 2022 31
Cont.
• Let 𝒙
ෝ[𝒏] be a periodic signal with period 𝑵 = 𝟐𝑴 + 𝟏.

• Define the finite – duration non – periodic signal 𝒙[𝒏] as one period of 𝒙
ෝ[𝒏], as
shown by:
𝒙ෝ𝒏, −𝑴 ≤ 𝒏 ≤ 𝑴
𝒙𝒏 =ቊ
𝟎, 𝒏 >𝑴
• This relationship is illustrated in figure 11.

• Note that as M increases, the periodic replicates of 𝐱[𝐧] that are present in 𝐱ො [𝐧]
move further and further away from the origin.

• Eventually, as 𝑴 → ∞, these replicates are removed to infinity.


Monday, April 25, 2022 32
Cont.

Monday, April 25, 2022 33


Cont.
• Thus we may write:
ෝ[𝒏]
𝒙 𝒏 = 𝐥𝐢𝐦 𝒙
𝑴→∞

• Begin with the DTFS representation for the periodic signal 𝒙


ෝ[𝒏].

• We have the DTFS pair:

𝑴
DTFS-periodic signal
ෝ𝒏 = ෍
𝒙 𝑿[𝒌]𝒆𝒋𝒌𝜴𝒐𝒏
𝒌=−𝑴
𝑀
1 −𝑗𝑘Ω𝑜 𝑛 DTFS-non periodic
𝑋𝑘 = ෍ 𝑥[𝑛]𝑒

2𝑀 + 1 signal
𝑛=−𝑀

Monday, April 25, 2022 34


Cont.
• Since 𝒙
ෝ 𝒏 = 𝒙[𝒏] for −𝑴 ≤ 𝒏 ≤ 𝑴, we may rewrite the last equation as:
𝑀 ∞
1 −𝑗𝑘Ω𝑜 𝑛
𝟏
𝑋𝑘 = ෍ 𝑥[𝑛]𝑒 = ෍ 𝒙[𝒏]𝒆−𝒋𝒌𝜴𝒐𝒏
2𝑀 + 1 𝟐𝑴 + 𝟏
𝑛=−𝑀 𝒏=−∞

Where the second line follows from the fact that 𝒙 𝒏 = 𝟎 for 𝒏 > 𝑴.

• We now define a continuous function of frequency, 𝑿 𝒆𝒋𝜴 , whose samples at 𝒌𝜴𝒐


are equal to the DTFS coefficients normalized by 𝟐𝑴 + 𝟏. That is:

𝑿 𝒆𝒋𝜴 = ෍ 𝒙[𝒏]𝒆−𝒋𝜴𝒏
𝒏=−∞
Monday, April 25, 2022 35
Cont.
So that:
𝑿 𝒆𝒋𝒌𝜴𝒐
𝑿𝒌 =
𝟐𝑴 + 𝟏
• Substitute this definition for 𝑿[𝒌] in to the equation 𝒙
ෝ 𝒏 to obtain:

𝑴
𝟏
ෝ𝒏 =
𝒙 ෍ 𝑿 𝒆𝒋𝒌𝜴𝒐 𝒆𝒋𝒌𝜴𝒐 𝒏
𝟐𝑴 + 𝟏
𝒌=−𝑴

𝟐𝝅
• Using the relationship 𝜴𝒐 = = 𝟐𝝅/(𝟐𝑴 + 𝟏), we write:
𝑵

𝑴
𝟏
ෝ𝒏 =
𝒙 ෍ 𝑿 𝒆𝒋𝒌𝜴𝒐 𝒆𝒋𝒌𝜴𝒐𝒏 Ω𝑜
𝟐𝝅
Monday, April 25, 2022
𝒌=−𝑴 36
Cont.
• At this point we invoke the fact that 𝒙[𝒏] is the limiting value of 𝒙
ෝ 𝒏 as 𝑴 → ∞.

• However, let us first consider the effect of 𝑀 → ∞ on the fundamental


frequency, 𝜴𝒐 .

• As M increases, 𝜴𝒐 decreases and the spacing between harmonics in the DTFS


decreases.

• This decreases in harmonic spacing is illustrated in figure 12 by depicting


𝑿 𝒆𝒋𝒌𝜴𝒐 for increasing values of M.

Note that: 𝑿 𝒆𝒋𝜴 is 𝟐𝝅 periodic in 𝜴.


37
Monday, April 25, 2022
Cont.

Monday, April 25, 2022 38


Cont.
• This follows from the equation:

𝑿 𝒆𝒋𝜴 = ෍ 𝒙[𝒏]𝒆−𝒋𝜴𝒏
𝒏=−∞
and the 𝟐𝝅 periodicity of 𝒆−𝒋𝜴𝒏 .
𝟐𝝅
• Combining the equations: 𝜴𝒐 = = 𝟐𝝅/(𝟐𝑴 + 𝟏),
𝑵

𝑥 𝑛 = lim 𝑥[𝑛]
ො 𝑎𝑛𝑑 ෝ𝒏
𝒙
𝑀→∞
𝑴
𝟏
We have: = ෍ 𝑿 𝒆𝒋𝒌𝜴𝒐 𝒆𝒋𝒌𝜴𝒐𝒏 𝜴𝒐
𝟐𝝅
𝒌=−𝑴

𝑴
𝟏
𝒙 𝒏 = 𝒍𝒊𝒎 ෍ 𝑿 𝒆𝒋𝒌𝜴𝒐 𝒆𝒋𝒌𝜴𝒐 𝒏 𝜴𝒐
𝑴→∞ 𝟐𝝅
𝒌=−𝑴
Monday, April 25, 2022 39
Cont …
• In equation given above we are summing values of a function
𝑿 𝒆𝒋𝜴 𝒆𝒋𝜴𝒏

evaluated at 𝒌𝜴𝒐 multiplied by the width between samples, 𝜴𝒐 .

• Taking the limit and identifying 𝜴 = 𝒌𝜴𝒐 so that 𝒅𝜴 = 𝜴𝒐 , the sum in equation passes
to the integer:
𝑴
𝟏
𝒙 𝒏 = 𝒍𝒊𝒎 ෍ 𝑿 𝒆𝒋𝒌𝜴𝒐 𝒆𝒋𝒌𝜴𝒐𝒏 𝜴𝒐
𝑴→∞ 𝟐𝝅
𝒌=−𝑴

𝝅
𝟏
𝒙𝒏 = න 𝑿 𝒆𝒋𝜴 𝒆𝒋𝜴𝒏 𝒅𝜴
𝟐𝝅
−𝝅

Monday, April 25, 2022 40


Cont.
• The limits on the integral are obtained by noting that:
𝐥𝐢𝐦 𝑴 𝜴𝒐 = 𝝅
𝑴→∞

𝟐𝝅
Where, 𝜴𝒐 = and N=2M+1
𝑵

• We have thus expressed 𝒙[𝒏] as a weighted superposition of discrete – time


sinusoids.

• In this case the superposition is an integral and the weighting on each sinusoid
is:
𝟏
𝑿 𝒆𝒋𝜴 𝒅𝜴
𝟐𝝅
Monday, April 25, 2022 41
The DTFT Representation
• The DTFT representation is expressed as:
𝝅
𝟏
𝒙𝒏 = න 𝑿 𝒆𝒋𝜴 𝒆𝒋𝜴𝒏 𝒅𝜴
𝟐𝝅
−𝝅
Where:

𝑿 𝒆𝒋𝜴 = ෍ 𝒙[𝒏]𝒆−𝒋𝜴𝒏
𝒏=−∞

• We say that 𝑿 𝒆𝒋𝜴 and 𝒙 𝒏 are a DTFT pair and write:


𝑫𝑻𝑭𝑻
𝒙[𝒏] 𝑿 𝒆𝒋𝜴

Monday, April 25, 2022 42


Cont.
.
• The transform 𝑿 𝒆𝒋𝜴 describes the signal 𝒙[𝒏] as a function of sinusoidal
frequency 𝜴 and is termed the frequency – domain representation of 𝒙[𝒏].

• We say that the DTFT of 𝒙[𝒏] is:


𝑿 𝒆𝒋𝜴 = ෍ 𝒙[𝒏]𝒆−𝒋𝜴𝒏
𝒏=−∞

❑ Since it converts the time domain signal into frequency – domain


representation.

43
Cont.
• The inverse DTFT is:
𝝅
𝟏
𝒙𝒏 = න 𝑿 𝒆𝒋𝜴 𝒆𝒋𝜴𝒏 𝒅𝜴
𝟐𝝅
−𝝅

Since it converts the frequency – domain representation back in to the time –


domain.

Monday, April 25, 2022 44


Example 5:
Exponential Sequence
• Find the DTFT of the sequence
𝒙 𝒏 = 𝜶𝒏 𝒖[𝒏]
Solution:
• Using the general equation we have:
∞ ∞

𝑿 𝒆𝒋𝜴 = ෍ 𝜶𝒏 𝒖[𝒏]𝒆−𝒋𝜴𝒏 = ෍ 𝜶𝒏 𝒆−𝒋𝜴𝒏


𝒏=−∞ 𝒏=𝟎
• This sum diverges for 𝜶 ≥ 𝟏.

Monday, April 25, 2022 45


Cont.
• For α < 1 we have the convergent geometric series:

𝒏 𝟏
𝑿 𝒆𝒋𝜴 =෍ 𝜶𝒆 −𝒋𝜴 =
𝟏 − 𝜶𝒆−𝒋𝜴
𝒏=𝟎
• If 𝜶 is real valued, we may write the magnitude and phase of 𝑋 𝑒 𝑗Ω as:
𝒋𝜴
𝟏 𝟏
𝑿 𝒆 = 𝟏
= 𝟏
𝟐 𝟐 𝟐 𝟐 𝜶𝟐 + 𝟏 − 𝟐𝜶𝒄𝒐𝒔𝜴 𝟐
𝟏 − 𝜶𝒄𝒐𝒔𝜴 + 𝜶 𝒔𝒊𝒏 𝜴
𝒋𝜴
𝜶𝒔𝒊𝒏𝜴
𝒂𝒓𝒈 𝑿 𝒆 = −𝒂𝒓𝒄𝒕𝒂𝒏
𝟏 − 𝜶𝒄𝒐𝒔𝜴

𝟏 𝟏
𝑿 𝒋𝝎 = , 𝑿 𝒋𝝎 = 𝟏/𝟐
𝒂+𝒋𝝎 𝒂𝟐 +𝝎𝟐

Monday, April 25, 2022 46


Cont.
• The magnitude and phase of 𝑿 𝒆𝒋𝜴 are depicted graphically in figure 13 for
𝜶 = 𝟎. 𝟓. note that both are 𝟐𝝅 periodic.

• As in the other Fourier representation:

➢The magnitude spectrum of a signal is the magnitude of 𝑿 𝒆𝒋𝜴 depicted


as a function of 𝜴.

➢The phase spectrum is the phase of 𝑿 𝒆𝒋𝜴 depicted as a function of 𝜴.

Monday, April 25, 2022 47


Cont.

AMiT-ECE SIGNALS AND SYSTEMS CHAPTER


Monday, April 25, 2022 48
THREE: FOURIER REPRESENTATIONS FOR SIGNALS
Exercise:

Find the DTFT of

𝒙 𝒏 = 𝟐 𝟑 𝒏 𝒖[−𝒏].
Example 6: Rectangular pulse
• Let 𝟏, 𝒏 ≤𝑴
𝒙𝒏 =ቊ
𝟎, 𝒏 >𝑴

as depicted if figure 14(a). Find the DTFT of 𝒙[𝒏].


AMiT-ECE SIGNALS AND SYSTEMS CHAPTER
Monday, April 25, 2022 49
THREE: FOURIER REPRESENTATIONS FOR SIGNALS
Solution:
• Substitute for 𝒙[𝒏] for DTFT equation to obtain:
∞ 𝑴

𝑿 𝒆𝒋𝜴 = ෍ 𝒙[𝒏]𝒆−𝒋𝜴𝒏 = ෍ 𝟏𝒆−𝒋𝜴𝒏


𝒏=−∞ 𝒏=−𝑴

Monday, April 25, 2022 50


Cont.
• Now perform the change of variable, 𝒎 = 𝒏 + 𝑴, obtaining:

𝟐𝑴 𝟐𝑴

𝑿 𝒆𝒋𝜴 = ෍ 𝒆−𝒋𝜴(𝒎−𝑴) = 𝒆𝒋𝜴𝑴 ෍ 𝒆−𝒋𝜴𝒎


𝒎=𝟎 𝒎=𝟎

𝟏 − 𝒆 −𝒋𝜴(𝟐𝑴+𝟏)
𝒋𝜴𝑴
𝑿 𝒆𝒋𝜴 =൞ 𝒆 𝜴 ≠ 𝟎, ±𝟐𝝅, ±𝟒𝝅, ⋯
𝟏 − 𝒆−𝒋𝜴
𝟐𝑴 + 𝟏 𝜴 = 𝟎, ±𝟐𝝅, ±𝟒𝝅, ⋯

• The expression for 𝑿 𝒆𝒋𝜴 , when 𝜴 ≠ 𝟎, ±𝟐𝝅, ±𝟒𝝅, … , may be simplified by


symmetrizing the power of the exponential in the numerator and
denominator, as shown by:
Monday, April 25, 2022 51
Cont.
𝑒 −𝑗Ω(2𝑀+1)/2 𝑒 𝑗Ω(2𝑀+1)/2 − 𝑒 −𝑗Ω(2𝑀+1)/2
𝑋 𝑒 𝑗Ω = 𝑒 𝑗Ω𝑀
𝑒 −𝑗Ω/2 𝑒 −𝑗Ω/2 − 𝑒 −𝑗Ω/2
𝜴
𝐬𝐢𝐧 (𝟐𝑴 + 𝟏)
= 𝟐
𝜴
𝐬𝐢𝐧
𝟐
• Note that L’Hopital’s rule gives:
Ω
sin (2𝑀 + 1)
lim 2 = 2𝑀 + 1
Ω→0,±2π,±4π,…, Ω
sin
2

AMiT-ECE SIGNALS AND SYSTEMS CHAPTER


Monday, April 25, 2022 52
THREE: FOURIER REPRESENTATIONS FOR SIGNALS
Cont.
• Hence, rather than writing 𝑿 𝒆𝒋𝜴 as two forms dependent on the value of Ω, we simply write:
.
𝜴
𝒔𝒊𝒏 (𝟐𝑴 + 𝟏)
𝑿 𝒆𝒋𝜴 = 𝟐
𝜴
𝒔𝒊𝒏
𝟐
with the understanding that 𝑿 𝒆𝒋𝜴 , for 𝜴 = 𝟎, ±𝟐𝝅, ±𝟒𝝅, … , is obtained as a limit,

• In this example 𝑿 𝒆𝒋𝜴 is purely real.

• A graph of 𝑿 𝒆𝒋𝜴 as a function of 𝜴 is given in figure 14(b)

• We see that as M increases, the time extent of 𝑥[𝑛] increases while the energy in 𝑿 𝒆𝒋𝜴
becomes more concentrated near 𝜴 = 𝟎.
Monday, April 25, 2022 53
Example 7: Discrete – Time sinc Function
• Find the inverse DTFT of
𝒋𝜴 𝟏, 𝜴 ≤𝑾
𝑿 𝒆 =ቊ
𝟎, 𝑾< 𝜴 <𝝅
as depicted in figure 15(a).

AMiT-ECE SIGNALS AND SYSTEMS CHAPTER


Monday, April 25, 2022 54
THREE: FOURIER REPRESENTATIONS FOR SIGNALS
Cont.
Solution:

• First note that 𝑿 𝒆𝒋𝜴 is specified only for −𝝅 < 𝜴 < 𝝅.

• This is all that is needed, since 𝑿 𝒆𝒋𝜴 is always 𝟐𝝅 periodic and the inverse DTFT
depends only on the values in the interval −𝝅 < 𝜴 < 𝝅.

• Substituting for 𝑋 𝑒 𝑗Ω in equation for inverse DTFT gives:

𝑾
𝟏 𝟏 𝑾
𝒙𝒏 = න 𝒆𝒋𝜴𝒏 𝒅𝜴 = 𝒆𝒋𝜴𝒏 ቮ 𝒏≠𝟎
𝟐𝝅 𝟐𝝅𝒏𝒋
−𝑾 −𝑾
AMiT-ECE SIGNALS AND SYSTEMS CHAPTER
Monday, April 25, 2022 55
THREE: FOURIER REPRESENTATIONS FOR SIGNALS
Cont.

𝟏
𝒙𝒏 = 𝒔𝒊𝒏 𝑾𝒏 𝒏≠𝟎
𝝅𝒏
• For 𝒏 = 𝟎, the integrand is unity and we have 𝒙 𝟎 = 𝐖/𝝅.

• It is easy to show using L’Hopital’s rule that:

𝟏 𝑾
𝐥𝐢𝐦 𝒔𝒊𝒏 𝑾𝒏 =
𝒏→𝟎 𝝅𝒏 𝝅
and thus we usually write:

Monday, April 25, 2022 56


Cont.
𝟏
𝒙𝒏 = 𝒔𝒊𝒏 𝑾𝒏
𝝅𝒏
as the inverse DTFT of 𝑿 𝒆𝒋𝜴 with the understanding that the value at 𝒏 = 𝟎 is obtained
as the limit.

A graph depicting 𝒙[𝒏] is given in figure 15(b)

• We may also write:

𝑾 𝑾𝒏
𝒙 𝒏 = 𝒔𝒊𝒏𝒄
𝝅 𝝅
Using the sinc function notation defined earlier.
AMiT-ECE SIGNALS AND SYSTEMS CHAPTER
Monday, April 25, 2022 57
THREE: FOURIER REPRESENTATIONS FOR SIGNALS
Example 8: The impulse.
• Find the DTFT of 𝒙 𝒏 = 𝜹[𝒏]. .
Solution:
• for 𝑥 𝑛 = δ[𝑛], we have:

𝑿 𝒆𝒋𝜴 = ෍ 𝜹[𝒏]𝒆−𝒋𝜴𝒏 = 𝟏
𝒏=−∞
• Hence,
𝐷𝑇𝐹𝑇
δ[𝑛] 1
• This DTFT pair is depicted in Figure 16.

AMiT-ECE SIGNALS AND SYSTEMS CHAPTER


Monday, April 25, 2022 58
THREE: FOURIER REPRESENTATIONS FOR SIGNALS
Cont …
Cont …

Example 9: Find the inverse DTFT of:


𝑿 𝒆𝒋𝜴 = 𝜹 𝜴 −π<Ω≤π
Solution: By definition,
𝝅 𝝅
𝟏 𝒋𝜴 𝒋𝜴𝒏
𝟏
𝒙𝒏 = න𝑿 𝒆 𝒆 𝒅𝜴 = 𝒙 𝒏 = න 𝜹 𝜴 𝒆𝒋𝜴𝒏 𝒅𝜴
𝟐𝝅 𝟐𝝅
−𝝅 −𝝅
AMiT-ECE SIGNALS AND SYSTEMS CHAPTER
Monday, April 25, 2022 59
THREE: FOURIER REPRESENTATIONS FOR SIGNALS
Cont …
• Using the shifting property of the impulse function to obtain: .
𝟏
𝒙𝒏 =
𝟐𝝅
and thus write:
𝟏 𝑫𝑻𝑭𝑻
𝜹 𝜴 −𝝅<𝜴≤𝝅
𝟐𝝅
• In this example we have again defined only one period of 𝑋 𝑒 𝑗Ω .
• We can alternatively define 𝑿 𝒆𝒋𝜴 overall 𝜴 by writing it is an infinite sum of
delta functions shifted by integer multiples of 𝟐𝝅

𝑿 𝒆𝒋𝜴 = ෍ 𝜹 𝜴 − 𝒌𝟐𝝅
𝒌=−∞
Monday, April 25, 2022 60
Cont.
• Both definitions are common.
• This DTFT pair is depicted in figure 17.

AMiT-ECE SIGNALS AND SYSTEMS CHAPTER


Monday, April 25, 2022 61
THREE: FOURIER REPRESENTATIONS FOR SIGNALS
Cont.
Exercise:
• Find the inverse DTFT of:
𝑿 𝒆𝒋𝜴 = 𝟐𝒄𝒐𝒔 𝟐𝜴
• Find the DTFT of:
𝟐𝒏 , 𝟎≤𝒏≤𝟗
𝒙𝒏 =ቊ
𝟎 𝒐𝒕𝒉𝒆𝒓𝒘𝒊𝒔𝒆

AMiT-ECE SIGNALS AND SYSTEMS CHAPTER


Monday, April 25, 2022 62
THREE: FOURIER REPRESENTATIONS FOR SIGNALS
Continuous – Time Non – Periodic Signals:
The Fourier Transform (FT)
• The Fourier Transform is used to represent a continuous – time non –
periodic signal as a superposition of complex sinusoids.

• We shall simply present the FT in this section.

• An interpretation of the FT as the limiting form of the FS is developed as


problem 3.34.

• recall that the continuous non – periodic nature of a time signal implies that
the superposition of complex sinusoids involves a continuum of
frequencies ranging from −∞ 𝒕𝒐 ∞.
63
Monday, April 25, 2022
Cont.

• Thus the FT representation for a time signal involves an integral over


frequency, as shown by: ∞
𝟏
𝒙 𝒕 = න 𝑿(𝒋𝝎) 𝒆𝒋𝝎𝒕 𝒅𝝎
𝟐𝝅
−∞

• Where: ∞

𝑿 𝒋𝝎 = න 𝒙(𝒕)𝒆−𝒋𝝎𝒕 𝒅𝒕
−∞
In the first equation we have expressed 𝒙(𝒕) as a weighted superposition of
sinusoids.

Monday, April 25, 2022 64


Cont.
• We say that 𝒙(𝒕) and 𝑿(𝒋𝝎) are a FT pair and write:
𝑭𝑻
𝒙(𝒕) 𝑿 𝒋𝝎

• The transform 𝑿 𝒋𝝎 describes the signal 𝒙(𝒕) as a function of sinusoidal


frequency 𝝎 and is termed the frequency – domain representation for
𝒙(𝒕).

• The second equation is termed the FT of 𝒙(𝒕) since it converts the time –
domain signal in to its frequency – domain representation.

AMiT-ECE SIGNALS AND SYSTEMS CHAPTER


Monday, April 25, 2022 65
THREE: FOURIER REPRESENTATIONS FOR SIGNALS
Convergence of Fourier Transform:
Just as in the case of periodic signals, the sufficient conditions for the
convergence of 𝑿(𝝎) are the following (again referred to as the Dirichlet
conditions):

1. x(t) is absolutely integrable over any period, that is,

2. x(t) has a finite number of maxima and minima within any


finite interval of t
3. x(t) has a finite number of discontinuities within any finite
interval of t, and each of these discontinuities is finite.
Monday, April 25, 2022 66
Cont.
Example 10: Real Exponential.

• Find the FT of:


𝒙 𝒕 = 𝒆−𝒂𝒕 𝒖(𝒕)

Solution:

• The FT does not converge for 𝒂 ≤ 𝟎 since 𝑥(𝑡) is not absolutely integrable,
as shown by:

න 𝒆−𝒂𝒕 𝒅𝒕 = ∞ 𝒂≤𝟎
Monday, April 25, 2022 −∞ 67
Cont.
• For 𝒂 > 𝟎

∞ ∞ ∞

𝑿 𝒋𝝎 = න 𝒙(𝒕)𝒆−𝒋𝝎𝒕 𝒅𝒕 = න 𝒆−𝒂𝒕 𝒖(𝒕)𝒆−𝒋𝝎𝒕 𝒅𝒕 = න 𝒆−(𝒂+𝒋𝝎)𝒕 𝒅𝒕


−∞ −∞ 𝟎

𝟏 ∞ 𝟏
=− 𝒆− 𝒂+𝒋𝝎 𝒕
ቮ =
𝒂 + 𝒋𝝎 𝟎 𝒂 + 𝒋𝝎
• Converting to polar form, the magnitude and phase of 𝑿 𝒋𝝎 are given by:

𝟏 𝝎
𝑿 𝒋𝝎 = 𝟐 𝒂𝒏𝒅 𝒂𝒓𝒈 𝑿 𝒋𝝎 = −𝒂𝒓𝒄𝒕𝒂𝒏
𝒂 + 𝝎𝟐 𝟏/𝟐 𝒂

Monday, April 25, 2022 68


Cont.

AMiT-ECE SIGNALS AND SYSTEMS CHAPTER


Monday, April 25, 2022 69
THREE: FOURIER REPRESENTATIONS FOR SIGNALS
Cont.
Exercise:
• Find the FT of 𝒙(𝒕) assuming 𝒂 > 𝟎
𝒙 𝒕 = 𝒆𝒂𝒕 𝒖(−𝒕)

Example 11: Rectangular Pulse.

• Consider the rectangular pulse depicted in figure 19(a) and defined as:

𝟏 −𝑻≤𝒕≤𝑻
𝒙 𝒕 =ቊ
𝟎 𝒕 >𝑻

Find the FT of 𝒙(𝒕)


Monday, April 25, 2022 70
Cont.

Solution:

• The rectangular pulse is absolutely integrable provided 𝑇 < ∞.


➢For ω ≠ 0 we have:

∞ 𝑻
𝟏 −𝒋𝝎𝒕 𝑻 𝟐
𝑿 𝒋𝝎 = න 𝒙(𝒕)𝒆−𝒋𝝎𝒕 𝒅𝒕 = න 𝒆−𝒋𝝎𝒕 𝒅𝒕 =− 𝒆 ቤ = 𝐬𝐢𝐧(𝝎𝑻)
𝒋𝝎 −𝑻 𝝎
−∞ −𝑻 71
Cont.
• For ω = 0, the integral simplifies to 2𝑇.

• It is straightforward to show using L’Hopital’s rule that:

𝟐
𝐥𝐢𝐦 𝒔𝒊𝒏(𝝎𝑻) = 𝟐𝑻
𝝎→𝟎 𝝎

• Thus; we write for all 𝝎

𝟐
𝑿 𝒋𝝎 = 𝒔𝒊𝒏(𝝎𝑻)
𝝎
With the understanding that the value at 𝝎 = 𝟎 is obtained by evaluating a limit.
In this case 𝑿 𝒋𝝎 is real.

• It is depicted in figure 19(b).


Monday, April 25, 2022 72
Cont.
• The magnitude spectrum is:
𝐬𝐢𝐧 𝝎𝑻
𝑿 𝒋𝝎 =𝟐
𝝎
• And the phase spectrum is:
𝒔𝒊𝒏 𝝎𝑻
𝟎 >𝟎
𝒂𝒓𝒈 𝑿 𝒋𝝎 𝝎
𝒔𝒊𝒏 𝝎𝑻
𝝅 <𝟎
𝝎

Monday, April 25, 2022 73


Cont.
• We may write 𝑋(𝑗ω) using the sinc function notation as:

𝝎𝑻
𝑿 𝒋𝝎 = 𝟐𝑻𝒔𝒊𝒏𝒄
𝝅

• This example illustrate a very important property of the Fourier Transform.

• Consider the effect of changing T.


❑As T increases, 𝒙(𝒕) becomes less concentrated about the time origin, while
𝑿(𝒋𝝎) becomes more concentrated about the frequency origin.
❑As T decreases, 𝒙(𝒕) becomes more concentrated about the time origin,
while 𝑿(𝒋𝝎) becomes less concentrated about the frequency origin.
74
Monday, April 25, 2022
Cont.
• In a certain sense, the “width” of 𝒙(𝒕) is inversely related to the “width” of
𝑿(𝒋𝝎).

• As a general principle, we shall see that signals concentrated in one domain


are spread out in the other domain.

Example 12: The sinc Function.

• Find the inverse FT of the rectangular spectrum depicted in Figure 20(a) and
given by:

𝟏 −𝑾≤𝝎≤𝑾
𝑿 𝒋𝝎 = ቊ
Monday, April 25, 2022 𝟎 𝝎 >𝑾 75
Cont.

AMiT-ECE SIGNALS AND SYSTEMS CHAPTER


Monday, April 25, 2022 76
THREE: FOURIER REPRESENTATIONS FOR SIGNALS
Cont.
Solution:
• Using equation for the inverse FT gives for 𝑡 ≠ 0
∞ 𝑊
1 𝑗ω𝑡
1 𝑗ω𝑡
1 𝑗ω𝑡 𝑊
𝑥 𝑡 = න 𝑋(𝑗ω) 𝑒 𝑑ω = න𝑒 𝑑ω = 𝑒 ቤ
2π 2π 2𝑗π𝑡 −𝑊
−∞ −𝑊
𝟏 𝑾 𝑾𝒕
= 𝐬𝐢𝐧 𝑾𝒕 = 𝒔𝒊𝒏𝒄
𝝅𝒕 𝝅 𝝅
• When 𝒕 = 𝟎, the integral simplifies to 𝑾/𝝅. Since:
𝟏 𝑾
𝐥𝐢𝐦 𝒔𝒊𝒏 𝑾𝒕 =
𝒕→𝟎 𝝅𝒕 𝝅

AMiT-ECE SIGNALS AND SYSTEMS CHAPTER


Monday, April 25, 2022 77
THREE: FOURIER REPRESENTATIONS FOR SIGNALS
Cont.
• We write for all "𝑡“ 𝟏
𝒙 𝒕 = 𝒔𝒊𝒏 𝑾𝒕
𝝅𝒕
With the understanding that the value at 𝒕 = 𝟎 is obtained at a limit.

• Figure 20(b) depicts 𝒙(𝒕).

• Note again the inverse relationship between the concentration of the signal about the
origin in the time and frequency domains.
❑As W increases, the frequency – domain representation becomes less concentrated
about 𝝎 = 𝟎,
❑ while the time – domain representation becomes more concentrated about 𝒕 = 𝟎.

Monday, April 25, 2022 78


Cont.
• Another interesting observation can be made by considering this and the previous
example.

• In the previous example:


➢A rectangular time – domain pulse is transformed to a sinc function in
frequency

• In this example:
➢A sinc function in time – domain is transformed to a rectangular function in
frequency.

❖This “duality” is a consequence of the similarity between the forward trans form
and inverse transform.
79
Monday, April 25, 2022
Cont.
Example 13: The impulse.
• Find the FT of:
𝒙 𝒕 = 𝜹(𝒕)
Solution:
• This 𝑥(𝑡) does not satisfy the Dirichlet conditions, since the impulse δ(𝑡) is
only defined within an integral.

• We attempt to proceed in spite of this potential problem using the inverse FT to



write:
𝑿 𝒋𝝎 = න 𝜹(𝒕) 𝒆−𝒋𝝎𝒕 𝒅𝒕 = 𝟏
Monday, April 25, 2022
−∞ 80
Cont.
• The solution follows from the shifting property of the impulse function.

• Hence:
𝑭𝑻
𝜹 𝒕 𝟏

• And the impulse contains unity contributions from complex sinusoids of all
frequencies, from ω = −∞ 𝑡𝑜 ω = ∞.

Example 14: DC Signal.

• Find the inverse FT of:


𝑿(𝒋𝝎) = 𝟐𝝅𝜹(𝝎)

AMiT-ECE SIGNALS AND SYSTEMS CHAPTER


Monday, April 25, 2022 81
THREE: FOURIER REPRESENTATIONS FOR SIGNALS
Cont.
Solution:
• Here again we may expect convergence irregularities since 𝑋 𝑗ω has an infinite discontinuity
at the origin.
• Using the equation for FT, we find:
∞ ∞
𝟏 𝒋𝝎𝒕
𝟏
𝒙 𝒕 = න 𝑿(𝒋𝝎) 𝒆 𝒅𝝎 = න 𝟐𝝅𝜹(𝝎) 𝒆𝒋𝝎𝒕 𝒅𝝎 = 𝟏
𝟐𝝅 𝟐𝝅
−∞ −∞
• Hence, we identify:
𝑭𝑻
𝟏 𝟐𝝅𝜹 𝝎
As a FT pair.
• This implies that the frequency content of a DC signal is concentrated entirely at 𝝎 = 𝟎,
which is intuitively satisfying.

Monday, April 25, 2022 82


Cont.
Exercise: Find the inverse FT of:

𝟐𝒄𝒐𝒔𝝎 𝝎 ≤𝝅
𝑿 𝒋𝝎 = ቊ
𝟎 𝝎 >𝝅

Exercise: Find the FT of the time signal depicted in figure 21 and given by:

𝒕 𝒕 ≤𝟏
𝒙 𝒕 =ቊ
𝟎 𝒕 >𝟏

Monday, April 25, 2022 83


Cont.

Monday, April 25, 2022 84


Cont.

Monday, April 25, 2022 85


THE FREQUENCY RESPONSE OF CONTINUOUS-TIME LTIS
A. Frequency Response:
we showed that the output y(t) of a continuous-time LTI system equals the
convolution of the input x(t) with the impulse response h(t); that is,

Applying the convolution property we obtain

• where 𝒀(𝝎), 𝑿(𝝎),and H(𝝎), are the Fourier transforms of 𝒚(𝒕), 𝒙(𝒕), and
𝒉(𝒕), respectively.
AMiT-ECE SIGNALS AND SYSTEMS CHAPTER
Monday, April 25, 2022 86
THREE: FOURIER REPRESENTATIONS FOR SIGNALS
Cont.

❑The function H(𝝎) is called the frequency response of the system.


Relationships represented by Let

❑Then /H(𝝎)/ is called the magnitude response of the system, and 𝜽𝑯 (𝝎), the
phase response of the system.

AMiT-ECE SIGNALS AND SYSTEMS CHAPTER


Monday, April 25, 2022 87
THREE: FOURIER REPRESENTATIONS FOR SIGNALS
Cont.
❑Then /𝑯(𝝎)/ lis called the magnitude response of the system, and 𝜽𝒌 (𝝎) the phase response
of the system.

Fig. Relationships between inputs and outputs in an LTI system.


Monday, April 25, 2022 88
B. LTI Systems Characterized by Differential Equations:

A continuous-time LTI systems of practical interest are described by linear


constant-coefficient differential equations of the form

with 𝑀 ≤ 𝑵 . Taking the Fourier transform of both sides of and using the linearity
property and the time-differentiation property we have

Monday, April 25, 2022 89


Cont.
Thus, from

which is a rational function of 𝝎.The result is the same as the Laplace transform counter part
𝑯(𝒔) = 𝒀(𝒔)/𝑿(𝒔) with s =j𝝎.

AMiT-ECE SIGNALS AND SYSTEMS CHAPTER


Monday, April 25, 2022 90
THREE: FOURIER REPRESENTATIONS FOR SIGNALS
Exercise on Fourier Transfer
1. Find the Fourier transform of each of the following signals:

2. Consider the RC filter shown in Fig. Find the frequency response 𝑯(𝒘) of this
filter and discuss the type of filter.

Monday, April 25, 2022 91


The Frequency Response of Discrete-time LTI Systems
A. Frequency Response:
we showed that the output 𝒚[𝒏] of a discrete-time LTI system equals the
convolution of the input 𝒙[𝒏] with the impulse response 𝒉[𝒏]; that is,

Applying the convolution property we obtain

where Y(Ω), X(Ω), and H(Ω) are the Fourier transforms of y[n], x [n] , and
h[n], respectively.
92
Monday, April 25, 2022
Cont.
we have:

Therefore;

Then /H(𝜴)/ is called the magnitude response of the system, and 𝜽𝑯 (𝛀), the
phase response of the system.

Monday, April 25, 2022 93


LTI Systems Characterized by Difference Equations:
As discussed in chapter 2, many discrete-time LTI systems of practical interest
are described by linear constant-coefficient difference equations of the form

with 𝑴 ≤ 𝑵. Taking the Fourier transform of both sides of Eq. and using the
linearity property and the time-shifting), we have

Monday, April 25, 2022 94


Cont.
• The result is the same as the Z-transform counterpart 𝑯(𝛀) = 𝒀(𝛀)/
𝑿(𝛀), with z = 𝒆𝒋𝜴 that is,

Periodic Nature of the Frequency Response:


• we have

• Thus, unlike the frequency response of continuous-time systems, that of all


discrete-time LTI systems is periodic with period 2𝝅.

• Therefore, we need observe the frequency response of a system only over the
frequency range 0 ≤ 𝜴 ≤ 2𝝅 or − 𝝅 ≤ 𝜴 ≤ 𝝅.
Monday, April 25, 2022 95
Properties of Fourier Representations
• The Four Fourier representations discussed in this chapter are summarized in
table below.

• This table provides a convenient reference for both the definition of each
transform and identification of the class of signals to which each applies.

• All the Four Fourier representations are based on the complex sinusoid.

• Consequently, they all share a set of common properties that follow from the
characteristics of complex sinusoids.

• This section examines the properties of Fourier representations.

Monday, April 25, 2022 96


Reading Assignment:
Read on the following properties of Fourier Representations:

• Periodic signals have Fourier Series representations.


➢The Fourier Series (FS) applied to CT periodic signals and
➢The Discrete – Time Fourier Series (DTFS) applied to DT periodic
signals.
• Convolution Properties(𝒚(𝒕) = 𝒙(𝒕) ∗ 𝒉(𝒕))
𝒚(𝒋𝝎) = 𝒙 𝒋𝝎 𝑯(𝒋𝝎)
𝒚(𝒆𝒋𝝎 ) = 𝒙(𝒆𝒋𝝎 )𝑯(𝒆𝒋𝝎 )
Monday, April 25, 2022 97
Cont.

Monday, April 25, 2022 98


END OF
CHAPTER
THREE
Monday, April 25, 2022 99

You might also like