ELECTRICAL & ELECTRONICS ENGINEERING,

FOURAH BAY COLLEGE, UNIVERSITY OF

SIERRA LEONE

SIGNAL ANALYSIS (EENG-411)

Lecture Presentation By:

Dr. ABDULAI SANKOH

 FOURIER TRANSFORM AND ITS SPECTRA
❑ Although, the CTFS is a powerful concept, its suffer from one
disadvantage, i.e it is only applicable to periodic signals. Most
signals of practical important are not periodic.
❑ The analysis of non-periodic signals requires the use of the
Fourier Transform or Fourier Integral.
❑ In the derivation of the Fourier Transform, the Fourier series
expression can be used by making a number of inferences.
Module Content
❑ Assume that s(t) is initially periodic. Hence, from Eq. (15) and
(16) we can write:
𝒗 𝒕 = σ∞
𝒏=−∞ 𝑽𝒏 𝒆
𝒋𝒏𝝎𝟎 𝒕
(16)
𝑻
𝟏 𝟐 (15)
where 𝑽𝒏 = ‫𝒗 ׬‬ 𝒕 𝒆−𝒋𝒏𝝎𝟎 𝒕 𝒅𝒕
𝑻 −𝑻
𝟐

❑ In the limit, for a single pulse we note the following:

𝜔 = 2𝜋Τ𝑇 and as 𝑇 tends to ∞ , 𝛚 = 𝑑𝜔 , a small incremental
1 𝑑𝜔
quantity ∴ 𝑇 = 2𝜋
 FOURIER TRANSFORM AND ITS SPECTRA
❑ Also, in the limit, the σ(summation sign) tends to the
‫( ׬‬integral sign). Eq. (15a) can be written as:
𝒅𝝎 ∞
𝑽𝒏 = න 𝒗 𝒕 𝒆−𝒋𝝎𝒕 𝒅𝒕
𝟐𝝅 −∞
❑ where nɷo= ɷ giving a value for Vn for the continuum of
frequencies, i.e. a continuous spectrum. Hence, Eq. (16)
becomes: ∞ Module Content
𝑑𝜔 ∞
𝑣 𝑡 =න න 𝑣 𝑡 𝑒 −𝑗𝜔𝑡 𝑑𝑡 𝑒 𝑗𝜔𝑡
−∞ 2𝜋 −∞
❑ The term in brackets can be written as:
∞
𝑉 𝜔 = ‫׬‬−∞ 𝑣 𝑡 𝑒 −𝑗𝜔𝑡 𝑑𝑡 (17)
or
∞
𝑉 𝒇 = ‫׬‬−∞ 𝑣 𝑡 𝑒 −𝑗𝟐𝝅𝒇𝑡 𝑑𝑡 (18)
❑ Eq. (17) & (18) are called the FOURIER TRANSFORM or
FOURIER INTEGRAL of v(t)
 FOURIER TRANSFORM AND ITS SPECTRA
1 ∞
∴𝑣 𝑡 = න 𝑉 𝜔 𝑒 𝑗𝜔𝑡 𝑑𝜔 (19)
2𝜋 −∞
❑ Now ꞷ = 2πf, Given dꞷ = 2πdf, Hence
∞
𝑣 𝑡 = න 𝑉 𝑓 𝑒 𝑗2𝜋𝑓𝒕 𝑑𝑓 (20)
−∞
❑ Eq. (19) & (20) are therefore referred to as the INVERSE
TRANSFORM of V(f). Module Content
❑ The combination of Eqs. (17) and (18) as well as Eq. (19) and (20)
respectively are referred to as a Fourier Transform pair.
Notational representation
❑ The Fourier transform pair can be represented by the curly ′ℑ′ as
follows: ℑ 𝑣(𝑡) = 𝑉(𝑓)

ℑ−1 𝑉(𝑓) = 𝑣(𝑡)

❑ The double-ended arrow notation is often used, relating the time
function to its transform function known as a transform pair, we
can write
 FOURIER TRANSFORM AND ITS SPECTRA
v(t)  V(f)
V(f)  v(t)

❑ In words, we can say from the above expressions, that the

Fourier Transform of v(t) is given by V(f) and that the Fourier
Inverse Transform of V(f) is v(t).
In summary we note that:
1 ∞
1. 𝑣 𝑡 = ‫𝑉 ׬‬ 𝜔 𝑒 𝑗𝜔𝑡 𝑑𝜔 ………. Synthesis equation
2𝜋 −∞ Module Content
∞
𝑉 𝜔 = ‫׬‬−∞ 𝑣 𝑡 𝑒 −𝑗𝜔𝑡 𝑑𝑡 ………. Analysis equation
2. Whereas the FS is a discrete function of frequency in that it is
only specified at particular frequencies and each frequency has
an amplitude, the FT is a continuous function of frequency and
𝑿 𝝎 represent the distribution of amplitude with frequency.
3. For the periodic time functions only discrete frequencies which
are harmonically related exist while for non-periodic time
functions all the frequencies exists to give a continuous
distribution from 0 to .
 FOURIER TRANSFORM AND ITS SPECTRA
❑ The synthesis equation enables us to synthesize or reconstruct
the original waveform v(t), from its Fourier transform
characteristics V(ω).
❑ The Analysis equation provides us with the tool to determine the
spectral behaviour of the waveform and thereby estimate
bandwidth and other requirements for transmission.

❑ Eg.-1 Find the Fourier transform

Module Content
of a rectangular pulse as shown
in Fig. 1:
Fig. 1

Solution:
𝐴; 𝑡 < 𝜏ൗ2
𝑣 𝑡 =
❑ The function of the waveform is given as: 0; 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
 FOURIER TRANSFORM AND ITS SPECTRA
❑ Let V() be the Fourier transform of v(t) then from Eq. (17) we
can write 𝝉
𝝉
Τ𝟐
∞
−𝒋𝝎𝒕 Τ𝟐 𝒆−𝒋𝝎𝒕
𝑽 𝝎 =‫׬‬ 𝒗 𝒕 𝒆 𝒅𝒕 = ‫𝑨 𝝉 ׬‬. 𝒆−𝒋𝝎𝒕 𝒅𝒕 = 𝑨
−∞ − Τ𝟐 −𝒋𝝎
−𝝉Τ𝟐
𝑨 −𝒋𝝎𝝉Τ𝟐 𝒋𝝎𝝉Τ𝟐 𝑨 𝒋𝝎𝝉Τ𝟐 −𝒋𝝎𝝉Τ𝟐 𝝉𝑺𝒊𝒏 𝝎𝝉Τ𝟐
= −𝒋𝝎 𝒆 −𝒆 = 𝒋𝝎 𝒆 −𝒆 =𝑨 𝝎𝝉ൗ
𝟐

𝝎𝝉
𝑽 𝝎 = 𝑨𝝉𝑺𝒊𝒏𝒄
𝟐
Module Content
❑ The spectrum has the sinc(x) shape and is continuous for all
frequency values as shown in the figure below
❑ The zero crossings are similarly derived. Here 𝝎𝝉Τ𝟐 = ±𝒏𝝅.
Given 𝝎𝒛𝒄 = ±𝟐𝒏𝝅/𝝉, where 𝒏 is an integer
2𝜋 4𝜋 6𝜋
𝝎𝒛𝒄 =± , ± , ± 𝜏
𝜏 𝜏

❑ Similarly we can express the zero crossings in terms of frequency

which results in 1 2 3
𝒇𝒛𝒄 = ± ,
𝜏
±
𝜏
, ±𝜏
 FOURIER TRANSFORM AND ITS SPECTRA

Fig. 2 The spectrum of v(t)

Module Content
1. The spectrum is a sinc(𝒙) curve
2. The zero crossings occur at values of 𝒙 = ±𝒏𝝅 , similar to that
for periodic signals. However in this case the spectrum is
continuous.
3. Amplitude values decrease with increasing values of 𝒙.
4. The peak value occurs at 𝒙 = 𝟎, with an amplitude value of A.

❑ It can be seen that the phenomenon of reciprocal spreading is

also observed for non-periodic functions.
 FOURIER TRANSFORM AND ITS SPECTRA
❑ In this case the first zero crossing occurs at 𝝎𝒛𝒄 = ±𝟐𝝅/𝝉 i.e
𝒇𝒛𝒄 = ±𝟏/𝝉 , where  is the pulse width.

❑ It can be seen that as  increases the zero crossings get closer

to the origin and the filter characteristics become more
selective, i.e. less bandwidth is required to transmit a larger
component of the signal as contained in the main lobe of the
spectrum. Module Content

❑ Conversely as the pulse width reduces the zero crossings are

farther removed from the origin. Thus a wider bandwidth will
be required to transmit such narrow width pulses. In other
words the narrower the pulse the greater the bandwidth
required to transmit it and vice versa. This phenomenon is
referred to as ‘reciprocal spreading’ similar to the case for
periodic signals.
 ENERGY SPECTRUM
❑ For a continuous signal the normalized power is given by:
𝑻𝒐
𝟏
𝑷𝑻𝒐𝒕𝒂𝒍 = ‫׬‬ 𝟐
𝑻 𝒗𝟐 (𝒕) 𝒅𝒕 i.e Parseval’s theorem
𝑻 − 𝒐
𝟐

❑ For a discontinuous signal such as a single pulse, the average

power tends to zero, since T tends to ∞
❑ We therefore consider the energy content associated with the
pulse in a unit resistance thereby
Module referring to the signal as an
Content
energy signal.
❑ Energy W is defined as:
∞
𝑬=‫׬‬ 𝒇(𝒕)𝟐 𝒅𝒕
−∞

❑ It can be recall that f(t) , i.e. the Synthesis Equation (19) is given
by: 𝟏 ∞ 𝒋𝝎𝒕
𝒇(𝒕) = ‫׬‬
𝟐𝝅 −∞
𝑮 𝝎 𝒆 𝒅𝝎

∞ 𝟏 ∞
❑ Therefore: 𝑬 = ‫׬‬−∞ 𝒇(𝒕) 𝟐𝝅 ‫׬‬−∞ 𝑮 𝝎 𝒆𝒋𝝎𝒕 𝒅𝝎 𝒅𝒕
 ENERGY SPECTRUM
𝟏 ∞ ∞
𝑬= න 𝒇 𝒕 . න 𝑮 𝝎 𝒆𝒋𝝎𝒕 𝒅𝝎 𝒅𝒕
𝟐𝝅 −∞ −∞

❑ Re-arranging we obtain :
𝟏 ∞ ∞
𝑬= ‫׬‬−∞ 𝑮 𝝎 𝒅𝝎 ‫׬‬ 𝒇 𝒕 𝒆𝒋𝝎𝒕 𝒅𝒕
𝟐𝝅 −∞

∞
Since 𝑮 𝝎 = ‫׬‬−∞ 𝒇 𝒕 𝒆−𝒋𝝎𝒕 𝒅𝒕

𝟏 ∞ Module Content
𝑬= 𝟐𝝅 ‫׬‬−∞ 𝑮 𝝎 𝑮 −𝝎 𝒅𝝎

∞ 𝟐
𝑬=
𝟏
‫׬‬−∞ 𝑮 𝝎 𝒅𝝎 (21)
𝟐𝝅

❑ In other words, 𝑮 𝝎 𝟐 Τ𝟐𝝅 defines how the energy in the signals

is distributed with frequency. Thus, 𝑮 𝝎 𝟐 Τ𝟐𝝅 is known as the
energy spectral density (ESD)
❑ Alternatively noting that  = 2f, giving d =2df we obtain from
Eq. (21) ∞ 𝟐 (22)
𝑬= 𝑮 𝒇 ‫׬‬−∞
𝒅𝒇
 ENERGY SPECTRUM

Fig. 3

Module Content

❑ Note the factor of 2 for the double-sided to single sided spectral

conversion
 ENERGY SPECTRUM
❑ This result is known as Rayleigh’s theorem and the quantity
G(f)|2 is called the energy spectral density and is given as
energy per unit frequency or Joules/Hz.
❑ For energy signals we associate energy with a continuous
band of frequencies. Hence the energy contained in the
frequency band (1, 2) is merely the area under the () curve
between 1 and 2.
Module Content
❑ Eg.-2 Given the current pulse, 𝒊 𝒕 = 𝒕𝒆−𝒃𝒕 𝒖 𝒕 ;
(a) Find the total 𝟏𝛀 associated with this waveform
(b) What fraction of this energy is present in the frequency
band from −𝒃 𝐭𝐨 𝒃 𝒓𝒂𝒅/𝒔
Solution
❑ The total 𝟏𝛀 energy associated with either a current or voltage
waveform can be found by use of Parseval's theorem
∞
𝟏
𝑬=
𝟐𝝅
න 𝑮𝑻 (𝝎) 𝟐
𝒅𝝎
−∞
 ENERGY SPECTRUM
❑ Where 𝑮𝑻 (𝒋𝝎) is the Fourier transform of the current or voltage
waveform. Hence, the Fourier transform of the current is
∞ ∞

𝑰(𝒋𝝎) = න 𝒕𝒆−𝒃𝒕 𝒆−𝒋𝝎𝒕 𝒅𝒕 = න 𝒕𝒆− 𝒋𝝎+𝒃 𝒕 𝒅𝒕

−∞ −∞
∞
𝒆− 𝒋𝝎+𝒃 𝒕 𝒆− 𝒋𝝎+𝒃 𝒕 𝟏
𝑰 𝒋𝝎 = 𝒕. − 𝟏. =
−(𝒋𝝎 + 𝒃) 𝒋𝝎 + 𝒃 𝟐 𝟎 𝒋𝝎 + 𝒃 𝟐
Module Content
𝟐
𝟐
𝟏 𝟏 𝟏
𝑰 𝒋𝝎 = 𝟐 = =
𝒋𝝎 + 𝒃 𝒃𝟐 − 𝝎𝟐 𝟐 + 𝟒𝝎𝟐 𝒃𝟐 𝒃𝟐 + 𝝎𝟐 𝟐

❑ The total energy associated with the current is

∞
𝟏 𝟐
𝑬𝑻 = න 𝑰 (𝒋𝝎) 𝒅𝝎
𝟐𝝅
−∞
𝟏 ∞ 𝟐
𝟏 ∞ 𝟏
𝑬𝑻 = න 𝑰 (𝒋𝝎) 𝒅𝝎 = න 𝟐 𝒅𝝎
𝝅 𝟎 𝝅 −∞ 𝒃𝟐 + 𝝎𝟐
 ENERGY SPECTRUM
❑ If we take the trigonometric substitution 𝝎 = 𝒃 𝒕𝒂𝒏 𝜽, then
𝟐 𝟐 𝟐
𝒃𝟐 + 𝝎𝟐 = 𝒃𝟐 + 𝒃𝟐 𝒕𝒂𝒏𝟐 𝜽 = 𝒃𝟐 𝒔𝒆𝒄𝟐 𝜽

Also 𝒅𝝎 = 𝒃𝒔𝒆𝒄𝟐 𝜽 𝒅𝜽

𝟏 ∞ 𝒃 𝒔𝒆𝒄𝟐 𝜽 𝟏 ∞ 𝟏 𝟏 ∞
𝑬𝑻 = න 𝟒 𝟒 𝒅𝜽 = න 𝟑 𝟐 𝒅𝜽 = 𝟑 න 𝒄𝒐𝒔𝟐 𝜽 𝒅𝜽
𝝅 −∞ 𝒃 𝒔𝒆𝒄 𝜽 𝝅 −∞ 𝒃 𝒔𝒆𝒄 𝜽 𝝅𝒃 −∞

𝟏 𝟏 𝟏
𝑬𝑻 = 𝜽 + 𝒔𝒊𝒏 𝟐𝜽 Module Content
𝝅𝒃𝟑 𝟐 𝟒
𝝎
❑ S𝐢𝐧𝐜𝐞 𝝎 = 𝒃 𝒕𝒂𝒏 𝜽, then, 𝜽 = tan−𝟏 𝒃

❑ Then, ∞
𝟏 𝟏 −𝟏
𝝎 𝟏 −𝟏
𝝎
𝑬𝑻 = tan + 𝒔𝒊𝒏 𝟐 tan
𝝅𝒃𝟑 𝟐 𝒃 𝟒 𝒃 𝟎

❑ S𝐢𝐧𝐜𝐞 tan−𝟏 ∞ = 𝝅Τ𝟐 then

𝟏 𝟏 𝝅 𝟏
𝑬𝑻 = . =
𝝅𝒃𝟑 𝟐 𝟐 𝟒𝒃𝟑
 ENERGY SPECTRUM
❑ To find the energy present in the frequency band −𝒃 < 𝒇 <
𝒃 𝒓𝒂𝒅/𝒔 we use Parseval's theorem and integrate.
𝒃
𝟏 𝟐 𝟏 𝒃 𝟏 𝒃
𝟏
𝑬𝒃 = න 𝑰 (𝒋𝝎) 𝒅𝝎 = න 𝑰 (𝒋𝝎) 𝟐
𝒅𝝎 = න 𝟐 𝒅𝝎
𝟐𝝅 𝝅 𝟎 𝝅 𝟎 𝒃𝟐 + 𝝎𝟐
−𝒃
𝒃
𝟏 𝟏 −𝟏
𝝎 𝟏 −𝟏
𝝎
𝑬𝒃 = tan + 𝒔𝒊𝒏 𝟐 tan
𝝅𝒃𝟑 𝟐 𝒃 𝟒 𝒃 𝟎

𝟏 𝟏 𝟏 𝟏 𝟏
𝑬𝒃 = tan −𝟏
𝟏 + 𝒔𝒊𝒏 𝟐 Module
tan −𝟏
𝟏 −Content
tan −𝟏
𝟎 + 𝒔𝒊𝒏 𝟐 tan −𝟏
𝟎
𝝅𝒃𝟑 𝟐 𝟒 𝟐 𝟒

𝟏 𝝅 𝟏
𝑬𝒃 = + − 𝟎 − 𝟎
𝝅𝒃𝟑 𝟖 𝟒

𝝅+𝟐 𝟏
𝑬𝒃 = . 𝟑
𝟖 𝝅𝒃
𝝅+𝟐 𝟏
𝑬𝒃 .
𝟖 𝝅𝒃𝟑 𝝅 + 𝟐
❑ The fraction of energy present is: 𝑬𝑻
=
𝟏
=
𝟐𝝅
= 𝟎. 𝟖𝟏𝟖
𝟒𝒃𝟑
 NETWORK RESPONSE
❑ Consider the network
Linear Filter
𝑣𝑖 (𝑡) 𝑭(𝒋𝝎)
𝑣0 (𝑡)

Fig.4 A Linear Network

❑ F() is the transfer function of the system = 𝐹 𝜔 ∠𝜑(𝜔)
𝒊𝒇 𝒗𝒊 𝒕 ⟷ 𝑮𝒊 (𝝎)
𝒗𝒐 𝒕 ⟷ 𝑮𝒐 (𝝎)
Module Content
Then for a linear system
𝑮𝒐 𝝎
𝑭 𝝎 =
𝑮𝒊 𝝎

i.e 𝑮𝒐 𝝎 = 𝑭 𝝎 𝑮𝒊 𝝎 (23)
∞
Where 𝑮𝒊 𝝎 = ‫׬‬ 𝒗𝒊 𝒕 𝒆−𝒋𝝎𝒕 𝒅𝒕
−∞

❑ The output of the network vo(t) is obtained from the Fourier

inverse of 𝑮𝒐 𝝎 i.e.
 NETWORK RESPONSE
❑ Consider the network
1 ∞
𝒗𝟎 𝑡 = න 𝑮𝟎 𝜔 𝑒 𝑗𝜔𝑡 𝑑𝜔 (24)
2𝜋 −∞

❑ Using Eq. (23) for Go(), in Eq. (24), we obtain

∞
𝟏
𝒗𝟎 𝒕 = න 𝑭(𝒘)𝑮𝒊 𝝎 𝒆𝒋𝝎𝒕 𝒅𝝎 (25)
𝟐𝝅 −∞
Module Content
❑ The output of Eq. (25) is expressed in terms of the Transfer
Function of the network and the input amplitude spectrum.
 PROPERTIES OF THE FOURIER TRANSFORM
The Shift Theorem
𝒊𝒇 𝒇 𝒕 ⟷ 𝑭 (𝝎)
(24)
𝒇 𝒕 − 𝝉𝟎 ⟷ 𝒇 (𝒋𝝎)𝒆−𝒋𝝎𝝉𝟎
❑ Proof
Let 𝒇𝟏 𝒕 ⟷ 𝑮𝟏 (𝝎) and
𝒇𝟐 𝒕 ⟷ 𝑮𝟐 (𝝎)
❑ Hence we can define Module Content
∞
𝑮𝟐 𝝎 = න 𝒇(𝒕 − 𝝉𝟎 ) 𝒆−𝒋𝝎𝝉𝒐 𝒅𝒕
−∞
∞
−𝒋𝝎𝒕′ (*)
=න 𝒇(𝒕′ ) 𝒆 . 𝒆−𝒋𝝎𝝉𝒐 𝒅𝒕′
−∞
❑ Where
𝒕′ = 𝒕 − 𝝉𝟎 ⟹ 𝒕 = 𝒕′ + 𝝉𝟎 and 𝒅𝒕 = 𝒅𝒕′ 𝐬𝐢𝐧𝐜𝐞 𝝉𝟎 𝐢𝐬 𝐚 𝐜𝐨𝐧𝐬𝐚𝐧𝐭
∞ ′ −𝒋𝝎𝒕′ ∞ ′ −𝒋𝝎𝒕′
But ‫׬‬−∞ 𝒇(𝒕 )𝒆 = 𝑮𝟏 𝝎 i.e 𝑮𝟏 𝝎 = ‫׬‬−∞ 𝒇(𝒕 )𝒆
 PROPERTIES OF THE FOURIER TRANSFORM
❑ Since t’ is dummy variable.
❑ Hence from Eq. (*) from above, we can write
∞
−𝒋𝝎𝝉𝒐 ′ −𝒋𝝎𝒕′
𝑮𝟐 𝝎 = 𝒆 න 𝒇(𝒕 ) 𝒆 𝒅𝒕′
−∞

Hence 𝑮𝟐 𝝎 = ℒ 𝒇(𝒕 − 𝝉𝟎 ) = 𝑮𝟏 𝝎 𝒆−𝒋𝝎𝝉𝒐 (25)

❑ In general if Module Content

𝒇 𝒕 ⟷ 𝑭 (𝒋𝝎)
❑ Then
𝒇 𝒕 − 𝝉𝟎 ⟷ 𝒇 (𝒋𝝎)𝒆−𝒋𝝎𝝉𝒐
❑ The function f(t-o) is the delayed version of f(t) by a factor of o.
❑ It can be observe that the amplitude spectrum of this shifted
function is not affected by the shift.
❑ However the phase spectrum is modified by an amount (-o),
i.e a shift of o in the time domain implies a phase shift of (-o).
 PROPERTIES OF THE FOURIER TRANSFORM
Eg-3: Consider the double gated function as shown
V(t)

T T

t
D D
Solution
❑ The double gated function of the
Module figure above is the sum of two
Content
time shifted versions of the standard pulse as shown.
f(t)
A Rectangular Pulse
𝜏 𝑓(𝑡) =
𝐴;𝑡 < 𝜏Τ2
given A=1
0; 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
𝝉 𝝉 t
−
𝟐 𝟐
❑ From the solution of Eg.-1
𝝎𝝉
𝕵 𝒇(𝒕) = 𝑨𝝉𝑺𝒊𝒏𝒄
𝟐
𝝎𝑫 𝝎𝑫
❑ Hence, 𝕵 𝒗(𝒕) = 𝒆𝒋𝝎𝑫 . 𝑻𝝉𝑺𝒊𝒏𝒄 +𝒆−𝒋𝝎𝑫 . 𝑻𝝉𝑺𝒊𝒏𝒄
𝟐 𝟐
 PROPERTIES OF THE FOURIER TRANSFORM
❑ Applying the shift theorem to the components of v(t). Hence, with
further manipulation we obtain the following:
𝝎𝝉
𝒗 𝒕 = 𝟐𝝉. 𝒄𝒐𝒔 𝑾𝑫 . 𝑺𝒊𝒏𝒄
𝟐
2. Linearity Theorem
𝒊𝒇 𝒇𝟏 𝒕 ⟷ 𝑭𝟏 (𝒋𝝎) and
𝒇𝟐 𝒕 ⟷ 𝑭𝟐 (𝒋𝝎)

Then 𝒂𝒇𝟏 𝒕 + 𝒃𝒇𝟐 𝒕 ⟷ 𝒂𝑭𝟏 Module

𝒋𝝎 + 𝒃𝑭Content
𝟐 (𝒋𝝎)

❑ where a and b are constants.

❑ The process of transforms implies integration which is a linear
process.
3. Scaling Property
𝒊𝒇 𝒇 𝒕 ⟷ 𝑭(𝒋𝝎)

❑ Then for a real constant ‘a’.

𝟏 𝒋𝝎
𝒇 𝒂𝒕 ⟷ 𝑭
𝒂 𝒂
 PROPERTIES OF THE FOURIER TRANSFORM
Proof:
❑ Assume a>0 then the transform of 𝒇 𝒂𝒕 is given by definition
as: ∞

𝕵 𝒇(𝒂𝒕) = න 𝒇(𝒂𝒕)𝒆−𝒋𝝎𝒕 𝒅𝒕
−∞
Let 𝐱 = 𝒂𝒕 ⇒ 𝒅𝒙 = 𝒂𝒅𝒕
❑ Hence, we obtain
∞ Module Content
𝒙
−𝒋 𝝎𝒂 𝒅𝒙 𝟏 𝒋𝝎
𝕵 𝒇(𝒂𝒕) = න 𝒇 𝒙 𝒆 = 𝑭
𝒂 𝒂 𝒂
−∞

❑ If a>0, it can be shown that

𝟏 𝒋𝝎
𝕵 𝒇(𝒂𝒕) = 𝑭
𝒂 𝒂
❑ Combining these results we obtain
𝟏 𝒋𝝎
𝒇 𝒂𝒕 ⟷ 𝑭
𝒂 𝒂
 PROPERTIES OF THE FOURIER TRANSFORM
3. Frequency shift theorem (Modulation)
❑ Applying the shift theorem to the components of v(t). Hence,
𝒊𝒇 𝒇 𝒕 ⟷ 𝑭(𝒋𝝎)
Then 𝒇 𝒕 𝒆𝒋𝝎𝟎𝒕 ⟷ 𝑭(𝒋𝝎 − 𝒋𝝎𝟎 )
Proof:
❑ By definition.
∞ ∞
Module Content
𝕵 𝒇(𝒕)𝒆𝒋𝝎𝟎 𝒕 𝒋𝝎 𝒕
= න 𝒇(𝒕)𝒆 𝟎 𝒆 −𝒋𝝎𝒕 𝒅𝒕 = න 𝒇 𝒕 𝒆−𝒋 𝝎−𝝎𝟎 𝒕 𝒅𝒕 = 𝑭(𝒋𝝎 − 𝒋𝝎𝟎 )
−∞ −∞

i.e. 𝒇 𝒕 𝒆𝒋𝝎𝟎 𝒕 ⟷ 𝑭(𝒋𝝎 − 𝒋𝝎𝟎 )

❑ This forms the mathematical basis of the Modulation (frequency
transition) in communications.
Eg-4: Given f(t) to be a modulating signal and 𝒄𝒐𝒔𝝎𝟎 𝒕, a carrier,
Determine the spectrum of the product 𝒇(𝒕)𝒄𝒐𝒔𝝎𝟎 𝒕
𝒆𝒋𝝎𝟎 𝒕 + 𝒆−𝒋𝝎𝟎𝒕
𝕵 𝒇(𝒕)𝒄𝒐𝒔𝝎𝟎 𝒕 = 𝕵 𝒇(𝒕)
𝟐
 PROPERTIES OF THE FOURIER TRANSFORM
𝟏 𝟏
𝕵 𝒇(𝒕)𝒄𝒐𝒔𝝎𝟎 𝒕 = 𝕵 𝒇(𝒕)𝒆𝒋𝝎𝟎 𝒕 + 𝕵 𝒇(𝒕)𝒆−𝒋𝝎𝟎𝒕
𝟐 𝟐
❑ i.e. Invoking the frequency shift theorem
𝟏 𝟏
𝕵 𝒇(𝒕)𝒄𝒐𝒔𝝎𝟎 𝒕 = 𝑭 𝒋𝝎 − 𝒋𝝎𝟎 + 𝑭 𝒋𝝎 + 𝒋𝝎𝟎
𝟐 𝟐

❑ Therefore, it should be noted that multiplying a time function

by 𝒄𝒐𝒔𝝎𝟎 𝒕 shifts the original spectrum of the signal such that
Module
half of its is centred about 𝝎Content
𝟎 and the other half is centred
around −𝝎𝟎 .
4. Time differentiation and Integration
𝒊𝒇 𝒇 𝒕 ⟷ 𝑭(𝒋𝝎)

𝒅𝒇(𝒕) Then
Then (1) 𝒅𝒕
⟷ 𝒋𝝎𝑭(𝒋𝝎) i.e. Time differentiation
∞ 𝟏
(2) ‫׬‬−∞ 𝒇 𝒕′ 𝒅𝒕′ ⟷ 𝒋𝝎 𝑭(𝒋𝝎) i.e. Integration
 PROPERTIES OF THE FOURIER TRANSFORM
𝟏 ∞
𝒇 𝒕 = 𝟐𝝅 ‫׬‬−∞ 𝑭(𝒋𝝎)𝒆𝒋𝝎𝒕 𝒅𝝎
i.e. differentiating inside the integral
𝒅𝒇(𝒕) 𝟏 ∞
△ = 𝟐𝝅 ‫׬‬−∞ 𝑭 𝒋𝝎 . 𝒋𝝎𝒆𝒋𝝎𝒕 𝒅𝝎
𝒅𝒕

∞
𝒅𝒇(𝒕) 𝒋𝝎 ∞
❑ Hence 𝕵
𝒅𝒕
=න න 𝑭 𝒋𝝎 𝒆𝒋𝝎𝒕 . 𝒅𝝎 𝒆−𝒋𝝎𝒕 𝒅𝒕
−∞ 𝟐𝝅 −∞

❑ Re-arranging we obtain
Module Content
∞
𝒅𝒇(𝒕) −𝒋𝝎𝒕
𝟏 ∞
𝕵 = 𝒋𝝎 න 𝒆 . 𝒅𝒕 න 𝑭(𝒋𝝎)𝒆𝒋𝝎𝒕 𝒅𝝎
𝒅𝒕 −∞ 𝟐𝝅 −∞
∞
𝟏 ∞
= 𝒋𝝎 න 𝒇(𝒕)𝒆−𝒋𝝎𝒕 . 𝒅𝒕 𝐬𝐢𝐧𝐜𝐞 𝒇 𝒕 = න 𝑭(𝒋𝝎)𝒆𝒋𝝎𝒕 𝒅𝝎
−∞ 𝟐𝝅 −∞
𝒅𝒇(𝒕)
𝕵 = 𝒋𝝎𝑭(𝒋𝝎) Then
❑ Hence 𝒅𝒕
∞
❑ Since 𝑭 𝒋𝝎 = න 𝒇 𝒕 𝒆−𝒋𝝎𝒕 𝒅𝒕 i.e.
−∞
𝒅𝒇(𝒕)
𝕵 ⟷ 𝒋𝝎𝑭(𝒋𝝎) (a*)
𝒅𝒕
 PROPERTIES OF THE FOURIER TRANSFORM
❑ Extended to the nth order derivative by repeated differentiation
within the integral yields
𝒅" 𝒇(𝒕) "
"
⟷ (𝒋𝝎) 𝑭(𝒋𝝎)
𝒅𝒕
ii Integration
∞
Let 𝒈 𝒕 = ‫׬‬−∞ 𝒇(𝒕′ ) 𝒅𝒕′
A𝐥𝐬𝐨 𝒈 𝒕 ⟷ 𝑮(𝝎) Module Content
𝒅𝒈(𝒕)
❑ Recall that = 𝒇(𝒕)
𝒅𝒕
❑ Using the differentiation relationship derived in Eq. (a*) and
taking the transforms of both sides, we obtain
Then
𝒋𝝎𝑮(𝒋𝝎) = 𝑭(𝒋𝝎)
or
𝟏
𝑮(𝒋𝝎) = 𝑭(𝒋𝝎)
𝒋𝝎
 PROPERTIES OF THE FOURIER TRANSFORM
5. Symmetry or Duality Theorem
𝒊𝒇
(i) 𝒗 𝒕 ⟷ 𝑽 𝒇 and if there exists a time function z 𝒕 such that
(ii) 𝒛 𝒕 = 𝑽 𝒕 i.e. z 𝒕 is the same as 𝑽 𝒇 albeit in the time domain

❑ Then
𝕵 𝒛(𝒕) = 𝒗(−𝒇), i.e. 𝒗 𝒕 ⟷ 𝑽 −𝒇

Where 𝒗 −𝒇 = 𝒗 𝒕 i.e. 𝒕 = −𝒇
Module
❑ Similarly, we can write the Content in the angular frequency
equivalents
domain, i.e.
If 𝒇 𝒕 ⟷ 𝑭 𝝎 , then F 𝒕 ⟷ 𝟐𝝅𝑭 −𝝎 ,

❑ i.e. replacing 𝒕 by −𝛚 in 𝒇(𝒕) and multiplying by 𝟐𝝅

Proof: Then
For v 𝒕 ⟷ 𝑽 𝒇
❑ We can write by definition
∞
𝒗 𝒕 = 𝕵−𝟏 𝑽(𝒇) = න 𝑽(𝒇)𝒆𝒋𝟐𝝅𝒇𝒕 𝒅𝒇
−∞
 PROPERTIES OF THE FOURIER TRANSFORM
❑ i.e. the above is the synthesis equation
❑ Replacing 𝒇 with the dummy variable 𝛌
∞
𝒗 𝒕 = ‫׬‬−∞ 𝑽(𝝀)𝒆𝒋𝟐𝝅𝝀𝒕 𝒅𝝀
❑ Furthermore, since 𝒕 is also a dummy variable and noting that
𝐳 𝐭 =𝒗 𝒕
❑ By definition, we obtain
∞ Module Content
𝕵 𝒛(𝝀) = ‫׬‬−∞ 𝒛(𝝀)𝒆𝒋𝟐𝝅𝒇𝝀 𝒅𝝀 :i.e. The analysis Equation
∞
= ‫׬‬−∞ 𝑽(𝝀)𝒆𝒋𝟐𝝅𝝀(−𝒇) 𝒅𝝀 :i.e. The inverse Transform of 𝒗 −𝒇

❑ Hence, the expression above defines the synthesis equation for

𝒗 −𝒇 , hence
Then
𝕵 𝒛(𝒕) = 𝒗(−𝒇)

Eg-5: From the figure below, given the sinc pulse 𝒛 𝒕 = 𝑨𝒔𝒊𝒏𝒄𝟐𝝎𝒕
❑ Now let us first of all define the unit rectangular pulse as follows:
 PROPERTIES OF THE FOURIER TRANSFORM

❑ Symbolically this can be written as:

𝒕 𝟏; 𝒕 < 𝝉Τ𝟐 Module Content
𝚷 =ቊ
𝝉 𝟎; 𝒕 > 𝝉Τ𝟐

❑ Hence, a general rectangular pulse can be defined as follows:

𝒕 𝑩; 𝒕 < 𝝉Τ𝟐
𝑩𝚷 =ቊ
𝝉 𝟎; 𝒕 > 𝝉Τ𝟐 Then
❑ The above yields:
𝒕
𝒗 𝒕 = 𝑩𝚷 ⟷ 𝑩𝝉𝑺𝒊𝒏𝒄𝒇𝝉 i.e.
𝝉

𝒗 𝒇 = 𝑩𝝉𝑺𝒊𝒏𝒄𝒇𝝉 (a*)
 PROPERTIES OF THE FOURIER TRANSFORM
❑ Rewriting the equation 𝒛 𝒕 = 𝑨𝒔𝒊𝒏𝒄𝟐𝝎𝒕 in the form of the
equation above, we obtain
𝑨
𝒛 𝒕 = 𝟐𝝎 𝑺𝒊𝒏𝒄 𝟐𝝎 𝒕 (b*)
𝟐𝝎

❑ Hence, we require the Rect function that will yield this Sinc function
Hence, for 𝒛 𝒕 = 𝑽 𝒕 and comparing Eqs. (a*) and (b*), we obtain
𝒛 𝒕 =𝑽 𝒕 Module Content
❑ Hence, for 𝒛 𝒕 = 𝑽 𝒕 and comparing Eqs. (a*) and (b*), we
𝑨
obtain, 𝝉 = 𝟐𝝎 𝒂𝒏𝒅 𝑩 =
𝟐𝝎
𝒇
❑ The duality theorem gives 𝕴 𝒛(𝒕) = 𝒗(−𝒇) = 𝑩𝜫(− )
𝝉
Then
❑ where 𝝉 has been replaced by −𝒇
𝑨 𝒇
= 𝜫 −
𝟐𝝎 𝟐𝝎
 PROPERTIES OF THE FOURIER TRANSFORM
𝑨 𝒇
❑ Noting 𝝉 = 𝟐𝝎 or 𝕴 𝒛(𝒕) = 𝜫 −
𝟐𝝎 𝟐𝝎

❑ Since the rectangular pulse has even symmetry.

Module Content
The band-limited spectrum of z(t)
❑ Example: Find the Fourier transform of δ(t) an impulse of unit
strength.
❑ Solution: Now recap that the unit impulse 𝜹(𝒕) = 𝟎 except at t =
∞ Then
0, and also the integral ‫׬‬−∞ 𝜹(𝒕)𝒅𝒕 =1
∞
❑ 𝒗 𝒇 = ‫׬‬−∞ 𝜹(𝒕)𝒆−𝒋𝟐𝝅𝒇𝒕 𝒅𝒕 = 𝟏

i.e 𝛿(𝑡) 1 ; ∀𝑓
 PROPERTIES OF THE FOURIER TRANSFORM
❑ i.e. the spectral density of the delta function or unit impulse is
1 volt Hz-1, i.e. it has a White spectral density of 1 Volt/Hz for all
frequencies.

Module Content

Fig. 16a The impulse ∂(t) Fig. 16b. The spectrum of the
delta pulse
❑ Consider the time shifted unit impulse δ(t - t0)
Then
❑ From the shift theorem we note

𝒇 𝒕 − 𝒕𝟎 𝑭(𝒋𝝎)𝒆−𝒋𝝎𝒕𝟎
 PROPERTIES OF THE FOURIER TRANSFORM
❑ Now we have shown that:
i.e 𝜹(𝒕) 𝟏 ; ∀𝒇

❑ Hence applying the shift theorem we obtain

𝝏(𝒕 − 𝒕𝟎 ) 𝟏. 𝒆−𝒋𝒘𝒕𝟎

Module Content

Fig. 18a The delayed pulse Fig. 18b. The spectrum of a time-
Then
shifted delta function

❑ Note that the amplitude spectrum maintains its characteristics

similar to the original function. However, the phase component
is affected.
 PROPERTIES OF THE FOURIER TRANSFORM
❑ Now since 𝝏(𝒕) 𝟏 and 1 𝜹(−𝒇), from the duality theorem
❑ 𝒇(𝒕) 𝒇(𝒋𝝎) then 𝒇 𝒕 𝟐𝝅𝒇 −𝝎 , 𝐡𝐞𝐧𝐜𝐞 1 𝜹(−𝒇),

❑ 𝐇𝐞𝐧𝐜𝐞 1 𝟐𝝅𝜹(𝝎) in radians per second or angular frequency.

❑ Also, using the frequency shift theorem on this expression, we
obtain
𝟏. 𝒆𝒋𝒘𝒐𝒕 𝟐𝝅𝜹(𝝎 − 𝝎𝟎 )
Module Content

❑ Example: Given a network whose transfer function is H(f), a unit

impulse δ(t) is applied at its input. Show that the response vo(t)
=h(t) i.e. its impulse response, is the Inverse transform of H(f),
i.e. 𝒉 𝒕 = 𝕴−𝟏 𝑯(𝒇)
Then
❑ Solution: Let vi(t)= ∂(t)
❑ To show that v0(t) = h(t)
❑ The output Spectral response of the linear network, is given by
 PROPERTIES OF THE FOURIER TRANSFORM
❑ 𝑽𝟎 (𝒇) = 𝑯(𝒇)𝑽𝒊 (𝒇)

❑ Where, 𝒗𝒊 (𝒕) 𝑽𝒊 (𝒇)

❑ Now we note that the input has a transform given by:

𝜹(𝒕) 𝟏 since 𝒗𝒊 𝒕 = 𝜹(𝒕)

❑ Hence 𝑽𝟎 (𝒇) = 𝑯(𝒇)𝒙𝟏 = 𝑯(𝒇)

❑ Also, 𝒗𝟎 (𝒕) = 𝒉(𝒕) = 𝕴−𝟏 Module

𝑽𝟎 (𝒇) Content

❑ i.e 𝒉(𝒕) = 𝕴−𝟏 𝑯(𝒇) , i.e. the inverse transform of the

transfer function of a network is its impulse response.
∞
❑ i.e. 𝒉 𝒕 = ‫׬‬−∞ 𝑯 𝒇 𝒆𝒋𝟐𝝅𝒇𝒕 𝒅𝒇 = ℑ−1 𝐻(𝑓)
Then
Special Case:
❑ Now if 𝑯(𝒇) = 𝟏 then the response of the network h(t) to an
impulse 𝝏 𝒕 is itself an impulse.
 PROPERTIES OF THE FOURIER TRANSFORM
❑ This therefore provides us with a definition of the impulse.
❑ From the expression above we note for H(f) = 1

❑ Now we note in general for a network with Transfer function H(f) and
an input of vi(t) = ∂(t), where 𝒗𝒊 (𝒕) ⟷ 𝑽(𝒇)

❑ We can write the impulse response as:

∞ Module Content
∞ 𝒋𝟐𝝅𝒇𝒕
𝒉 𝒕 = න 𝑯 𝒇 𝑽𝒊 𝒇 𝒆𝒋𝟐𝝅𝒇𝒕 𝒅𝒇) = න 𝟏. 𝒆 𝒅𝒇
−∞ −∞
∞ −𝒋𝟐𝝅𝒇𝒕
= න 𝟏. 𝒆 = 𝜹(𝒕
−∞

Then
❑ Since 𝑽𝒊 𝒇 = 𝟏 𝒊. 𝒆𝜹 𝒕 ⟷ 𝟏, This therefore provides us with the definition
for the unit impulse. This is an important expression as it forms the
basis of defining the delta function.
 FOURIER TRANSFORM OF PERIODIC SIGNAL
❑ As mentioned in the introduction we cannot normally talk of the
transform of periodic signals. However due to the peculiar
characteristics of the delta function the Fourier transform of
periodic signals can be expressed mathematically.
❑ Let us see how this is achieved in the solution of the problem given.
❑ Example: Find the Fourier Transform of 𝐯 𝒕 = 𝒄𝒐𝒔𝝎𝟎 𝒕
❑ Module Content
Solution: The exponential Fourier series representation of v(t) is given
𝟏 𝟏 𝟐𝝅
by: 𝒗 𝒕 = 𝒆𝒋𝝎𝒐 𝒕 + 𝒆−𝒋𝝎𝒐 𝒕 ; 𝝎𝒐 =
𝟐 𝟐 𝑻𝒐
𝟏
❑ Thus 𝑽𝟏 = 𝑽−𝟏 = & Vn= 0, n ≠ ±1
𝟐

∞
❑ The F.T of v(t) is therefore givenThen
by: 𝒗 𝒇 = ‫∞׬‬ 𝒄𝒐𝒔 𝝎𝒐 𝒕. 𝒆−𝒋𝟐𝝅𝒇𝒕 𝒅𝒕

∞ ∞
𝟏 𝟏
𝒗 𝒇 = න 𝒆−𝒋𝟐𝝅(𝒇−𝒇𝒐)𝒕 𝒅𝒕 + න 𝒆−𝒋𝟐𝝅(𝒇+𝒇𝒐)𝒕 𝒅𝒕
𝟐 ∞
𝟐 ∞
 FOURIER TRANSFORM OF PERIODIC SIGNAL
∞ ∞
❑ Now since ‫𝒆 ∞׬‬ −𝒋𝟐𝝅𝒇𝒕
𝒅𝒇 = ‫𝒆 ׬‬−𝒋𝟐𝝅𝒇𝒕 𝒅𝒇 = 𝜹(𝒕)
∞

❑ Comparing with the equation above, we obtain

𝟏 𝟏
𝒗 𝒇 = 𝜹(𝒇 − 𝒇𝟎 ) + 𝜹(𝒇 + 𝒇𝒐 )
𝟐 𝟐

❑ i.e. the Fourier transform of a sinusoidal signal or other periodic

signals consists of impulses located at each harmonic frequency of the
signal. Module Content

❑ Illustrative Example: Find the Fourier Transform of the periodic ON-

OFF rectangular pulse train.

Then
 FOURIER TRANSFORM OF PERIODIC SIGNAL
❑ Now from previous work 𝒇 𝒕 = σ∞ 𝒋𝒏𝒘𝒐𝒕
𝒏=−∞ 𝑭𝒏 𝒆

❑ i.e. the exponential representation of the F-series of f(t)

❑ Hence, Fn = 𝑨𝒅
𝑻
𝑺𝒊𝒏𝒄
𝒏𝝅𝒅
𝑻
, from previous work for a train of rectangular
pulses
❑ Taking the F.T on both sides of the equation, we obtain
𝕴 𝒇(𝒕) = Module
𝕴 σ𝒏 𝑭𝒏Content
𝒆𝒋𝒏𝒘𝒐 𝒕

𝑨𝒅 𝒏𝝅𝒅 𝒋𝒏𝒘𝒐 𝒕
= 𝕴 σ𝝅 𝒔𝒊𝒏 𝒄 ( )𝒆
𝑻 𝑻

❑ Now we note that from theory:

𝒆𝒋𝒏𝝎𝒐 𝒕 Then − 𝒏𝝎𝒐 )
𝟐𝝅𝜹(𝝎
𝟐𝝅𝑨𝒅 𝒏𝝅𝒅 𝟐𝝅
❑ Hence, 𝕴 𝒇(𝒕) = 𝑻
σ𝒏 𝒔𝒊𝒏 𝒄 𝑻
𝜹 𝝎 − 𝒏𝝎𝒐 𝒘𝒉𝒆𝒓𝒆 𝝎 𝑻
 FOURIER TRANSFORM OF PERIODIC SIGNAL
❑ Show sketch of spectrum using impulses
❑ The impulse train

Module Content
❑ Now 𝒇 𝒕 = σ∞
𝒌=−∞ 𝜹(𝒕 − 𝒌𝑻), i.e. a periodic waveform

❑ The complex Fourier series is given by

𝟏
𝒇 𝒕 = σ∞ 𝑭
𝒏=−∞ 𝒏 𝒆𝒋𝒏𝒘𝒐 𝒕
𝒘𝒉𝒆𝒓𝒆 𝑭𝒏 =
𝑻

∞ Then
𝟏
𝒇 𝒕 = ෍ 𝒆𝒋𝒏𝒘𝒐 𝒕
𝑻
𝒏=−∞

❑ Taking the Fourier Transform of both sides we obtain:

FOURIER TRANSFORM OF PERIODIC SIGNAL
∞
𝟏 𝟐𝝅
𝒇 𝒋𝝎 = 𝕴
𝑻
෍ 𝒆𝒋𝒏𝒘𝒐 𝒕 =
𝑻
෍ 𝜹(𝝎 − 𝒏𝝎𝒐 )
𝒏=−∞

𝐢. 𝐞 σ∞
𝐤=−∞ 𝛅(𝐭 − 𝐤𝐓) 𝐰𝐨 σ∞
𝐧=−∞ 𝛅(𝛚 − 𝐧𝛚𝐨 );

i.e An impulseModule Content

train of weight 𝛚𝐨

Then
THANK YOU