Fourier Transforms

IS3301
Complex Analysis and Mathematical Transforms
Department of Interdisciplinary Studies
Faculty of Engineering
University of Ruhuna
Content
• Introduction
• Fourier Transform
• Inverse Fourier Transforms

• Real and Imaginary Parts of Fourier Transform

• Fourier Sine and Cosine Transforms

• Properties of Fourier Transforms

• Applications
 Introduction: Fourier Transforms
• Fourier Transform is a mathematical technique that can be used to transform a function of time,
𝑓(𝑡) to a function of frequency, 𝐹(𝜔). i.e. used for frequency analysis of signals.

• The Fourier Transform permits the alternative representation of mostly non-periodic functions.

• (help to transform the signals between two different domains: transforming signals from
frequency domain to time domain or vice versa)

• Fourier transform can be derived from the complex exponential form of the Fourier series.

• Fourier Transforms can be used to interpolate functions as well as smooth signals.

• And also used for solving differential equations /partial differential equations.
 Introduction ctd…
• The function 𝐹(𝜔) (or 𝐹(𝑗𝜔) ) is called the Fourier transform of the function 𝑓(𝑡).
• Here, we refer to 𝐹 𝜔 as the frequency domain representation of a function or signal and
𝑓(𝑡) as the time domain representation.

Symbolically we can write Fourier transform (FT) of∞the function 𝑓(𝑡) as

𝐹 𝜔 = ℱ 𝑓(𝑡) = න 𝑓(𝑡)𝑒 −𝑖𝜔𝑡 𝑑𝑡

−∞
• 𝑓(𝑡) is often called the inverse Fourier transform (IFT) of 𝐹(𝜔) and we denote this by
writing

∞
−1
1
𝑓 𝑡 =ℱ 𝐹 𝜔 = න 𝐹 𝜔 𝑒 𝑖𝜔𝑡 𝑑𝜔
2𝜋
−∞
The above two formulas are known as Fourier Transform Pair. They are only valid for
absolutely Integrable signals.
 Dirichlet’s Conditions for Existence of Fourier Transform
• Fourier Transform can be applied to any function 𝑓 𝑡 if it satisfies the following conditions:

1. The function 𝑓(𝑡) (or signal) should have finite number of maxima and minima over any finite
interval.

2. f(t) (or signal) should have finite number of discontinuities over any finite interval.

∞
3. f(t) ((or signal) should be absolutely integrable i.e. ‫׬‬−∞ f(t) dt is convergent.
∞
. ‫׬‬−∞ f(t) dt < ∞

[i.e. an absolutely integrable function is a function whose absolute value is integrable, meaning
that the integral of the absolute value over the whole domain is finite.]

Note: These conditions are sufficient but not necessary. (because PS are not absolutely integrable
signals, but FT will exists, i.e. third condition is not compulsory)
Properties of Fourier Transforms
Linearity Property
• If 𝑓(𝑡), 𝑔(𝑡) are functions with transforms 𝐹(𝜔), 𝐺(𝜔) respectively,
then
𝐹{𝑓(𝑡) + 𝑔(𝑡)} = 𝐹(𝜔) + 𝐺(𝜔)

• If 𝑘 is any constant,
𝐹{𝛼𝑓(𝑡)} = 𝛼𝐹(𝜔)
𝐹{𝛽𝑔(𝑡)} = 𝛽𝐺(𝜔)
𝐹{𝛼𝑓 𝑡 + 𝛽𝑔(𝑡)}} = 𝛼𝐹(𝜔) + 𝛽𝐺(𝜔)
Properties of Fourier Transforms…
Conjugation
• If ℱ 𝑓(𝑡) = 𝐹(𝜔), then
ℱ 𝑓 ∗ (𝑡) = 𝐹 ∗ (−𝜔)
Properties of Fourier Transforms…
Area under 𝑓(𝑡)
• If ℱ 𝑓(𝑡) = 𝐹(𝜔), then
𝐴𝑟𝑒𝑎 𝑢𝑛𝑑𝑒𝑟𝑓 𝑡 = 𝐹(𝜔)ȁ𝜔=0

Area under 𝐹(𝜔)

• If ℱ 𝑓(𝑡) = 𝐹(𝜔), then
𝐴𝑟𝑒𝑎 𝑢𝑛𝑑𝑒𝑟𝐹(𝜔) = 2𝜋𝑓(𝑡)ȁ𝑡=0
Properties of Fourier Transforms…
Time Reversal
• If ℱ 𝑓(𝑡) = 𝐹(𝜔), then ℱ 𝑓(−𝑡) = 𝐹(−𝜔),

Time Scaling
1 𝜔
• If ℱ 𝑓(𝑡) = 𝐹(𝜔), then ℱ 𝑓(𝑎𝑡) = 𝐹( ), 𝑎 is any real
𝑎 𝑎
constant, 𝑎 ≠ 0
Properties of Fourier Transforms…
Time Shifting
Left Shifting
• If ℱ 𝑓(𝑡) = 𝐹(𝜔), then ℱ 𝑓(𝑡 + 𝑡0 ) = 𝐹(𝜔)𝑒 +𝑖𝜔𝑡0

Right Shifting
• If ℱ 𝑓(𝑡) = 𝐹(𝜔), then ℱ 𝑓(𝑡 − 𝑡0 ) = 𝐹(𝜔)𝑒 −𝑖𝜔𝑡0
Properties of Fourier Transforms…
Frequency Shifting
• If ℱ 𝑓(𝑡) = 𝐹(𝜔), then ℱ 𝑒 ±𝑖𝜔0𝑡 𝑓(𝑡) = 𝐹(𝜔 ∓ 𝜔0 ),

Convolution in Time
The convolution of two functions 𝑓 𝑡 and 𝑔(𝑡)
∞
over the interval (−∞ , ∞) is defined as
1
𝑓 𝑡 ∗𝑔 𝑡 = න 𝑓 𝑢 𝑔 𝑡 − 𝑢 𝑑𝑢
2𝜋
−∞
If ℱ 𝑓1 (𝑡) = 𝐹1 𝜔 and ℱ 𝑓2 (𝑡) = 𝐹2 𝜔 then
ℱ 𝑓1 𝑡 ∗ 𝑓2 (𝑡) = 𝐹1 𝜔 𝐹2 𝜔
Properties of Fourier Transforms…
Multiplication in Time (Convolution in Frequency) or Modulation Property
If ℱ 𝑓1 (𝑡) = 𝐹1 𝜔 and ℱ 𝑓2 (𝑡) = 𝐹2 𝜔 then
1
ℱ 𝑓1 𝑡 𝑓2 (𝑡) = 𝐹1 𝜔 ∗ 𝐹2 𝜔
2𝜋

Differentiation in Time
𝑑𝑓(𝑡)
If ℱ 𝑓(𝑡) = 𝐹(𝜔), then ℱ = (𝑖𝜔)𝐹(𝜔)
𝑑𝑡
𝑑 𝑘 𝑓(𝑡)
ℱ = (𝑖𝜔)𝑘 𝐹(𝜔)
𝑑𝑡 𝑘
Properties of Fourier Transforms…
Integration in Time
𝑡 𝐹(𝜔)
If ℱ 𝑓(𝑡) = 𝐹(𝜔), then ℱ ‫׬‬−∞ 𝑓 𝑡 𝑑𝜏 = + 𝜋𝐹(0)𝛿(𝜔)
𝑖𝜔

Differentiation in Frequency
𝑑𝐹(𝜔)
If ℱ 𝑓(𝑡) = 𝐹(𝜔), then ℱ 𝑡𝑓(𝑡) = 𝑖
𝑑𝜔

𝑛
𝑛 𝑛
𝑑 𝐹(𝜔)
ℱ 𝑡 𝑓(𝑡) = 𝑖
𝑑𝜔 𝑛
Properties of Fourier Transforms…
Modulation
If ℱ 𝑓(𝑡) = 𝐹(𝜔), then

1
ℱ 𝑓(𝑡) cos 𝜔0 𝑡 = 𝐹 𝜔 + 𝜔0 + 𝐹 𝜔 − 𝜔 0
2

𝑖
ℱ 𝑓(𝑡) sin 𝜔0 𝑡 = 𝐹 𝜔 + 𝜔0 − 𝐹 𝜔 − 𝜔0
2
Real and Imaginary Parts of Fourier Transform
• If the Fourier transform of the function 𝑓(𝑡), 𝐹 ∞𝜔 ;

𝐹 𝜔 = ℱ 𝑓(𝑡) = න 𝑓(𝑡)𝑒 −𝑖𝜔𝑡 𝑑𝑡

−∞
• By replacing 𝑒 −𝑖𝜔𝑡 = cos 𝜔𝑡 − sin 𝜔𝑡,
∞
then ∞

𝐹 𝜔 = ℱ 𝑓(𝑡) = න 𝑓(𝑡) cos 𝜔𝑡 𝑑𝑡 − 𝑖 න sin 𝜔𝑡 𝑓(𝑡) 𝑑𝑡

−∞ −∞
Here, both integrals are real, assuming that 𝑓(𝑡) is real.

Hence the real and imaginary parts of the Fourier Transforms are:
∞ ∞
𝑅𝑒 𝐹 𝜔 = ‫׬‬−∞ 𝑓(𝑡) cos 𝜔𝑡 𝑑𝑡 𝐼𝑚 𝐹 𝜔 = − ‫׬‬−∞ sin 𝜔𝑡 𝑓(𝑡) 𝑑𝑡
Fourier Sine and Cosine Transforms
• There are really three Fourier transforms,

• A complex form that is usually referred to as the Fourier

transform.

• Fourier Sine transforms

• Fourier Cosine transforms

Fourier Sine Transforms

Fourier Sine Transform:

∞
2
𝐹 𝜔 = න 𝑓 𝑡 sin 𝜔𝑡 𝑑𝑡
𝜋
0

Inverse Fourier Sine Transform:

∞
2
𝑓 𝑡 = න sin 𝜔𝑡 𝐹 𝜔 𝑑𝜔
𝜋
0
Fourier Cosine Transforms
Fourier Cosine Transform:
∞
2
𝐹 𝜔 = න 𝑓 𝑡 cos 𝜔𝑡 𝑑𝑡
𝜋
0

Inverse Fourier Cosine Transform:

∞
2
𝑓 𝑡 = න cos 𝜔𝑡 𝐹 𝜔 𝑑𝜔
𝜋
0
Examples
• Find the Fourier Transform of 𝑥 𝑡 = 𝛿(𝑡)
Examples
• Show that 𝑥 𝑡 = 𝐴0 (DC value) is not an absolutely integrable.
Examples
• Find the Fourier Transform of 𝑥 𝑡 = 𝐴0 (DC value).
Example
• Find the Fourier Transform of Unit Step Function 𝑈(𝑡) (u(t) is a power
signal which is not absolutely integrable)
Examples
• Find the Fourier Transform of 𝑥 𝑡 = 𝑒 −𝑎𝑡 𝑢 𝑡 , 𝑎 > 0
Examples
• Find the Fourier Transform of 𝑥 𝑡 = 𝑒 −𝑎 𝑡 , 𝑎 > 0
Some Useful Results in Computation of the Fourier
Transforms
∞ −𝑎𝑡 𝑎 𝑒 𝑖𝑎𝑡 −𝑒 −𝑖𝑎𝑡
• ‫׬‬0 𝑒 cos 𝜔𝑡 𝑑𝑡 = • sin 𝑎𝑡 =
𝑎2 +𝜔2 2𝑖

∞ −𝑎𝑡 𝜔
• ‫׬‬0 𝑒 sin 𝜔𝑡 𝑑𝑡 = • cos 𝑎𝑡 =
𝑒 𝑖𝑎𝑡 +𝑒 −𝑖𝑎𝑡
𝑎2 +𝜔2
2

∞ sin 𝜔𝑡 𝜋
• ‫׬‬0 𝑑𝑡 = ; 𝜔>0 •
∞ −𝑎2 𝑡 2 𝜋
𝑡 2 ‫׬‬0 𝑒 𝑑𝑡 =
2𝑎

∞ sin 𝑡 𝜋
• When 𝜔 = 1, ‫׬‬0 𝑡 𝑑𝑡 = • When a = 1,
2 ∞ −𝑡 2 𝜋
‫׬‬0 𝑒 𝑑𝑡 =
2
 Example
2
Find the Fourier Transform of 𝑒 −𝑡 .
2
Hence find the Fourier transform of 𝑒 −𝑎𝑡 .
∞
• Answer: 1 𝜔 2 𝜔2
2 − 𝑡+𝑖 −
𝐹{𝑒 −𝑡 } =
∞
𝐹 𝜔 = න 𝑒 2 4 𝑑𝑡 𝜔2 ∞
=
1 2
න 𝑒 −𝑡 𝑒 −𝑖𝜔𝑡 𝑑𝑡
2𝜋
−∞
𝜔2 ∞
2𝑒 − 4 −𝑧 2
2𝜋 = න𝑒 𝑑𝑡
−∞
∞ 𝑒− 4 𝜔 2
− 𝑡+𝑖 2 2𝜋
1 2 +𝑖𝜔𝑡) = න𝑒 𝑑𝑡 0 2
= න 𝑒 −(𝑡 𝑑𝑡 2𝜋 𝜔
2𝜋 −∞ −4
−∞
∞ 𝜔 2𝑒 𝜋
1 −(𝑡 2 +𝑖𝜔𝑡)
By putting 𝑧 = 𝑡 + 𝑖
2
=
= න𝑒 𝑑𝑡 𝜔2 ∞ 2𝜋 2 2
2𝜋
∞ −∞ 𝑒− 4 −𝑧 2
𝜔
−4
1 𝜔2
−(𝑡 2 +𝑖𝜔𝑡− 4 + 4 )
𝜔2 = න𝑒 𝑑𝑡 𝑒
= න 𝑒 𝑑𝑡 2𝜋 =
2𝜋 −∞
−∞
As 𝑒 −𝑧 2
is even function;
2
Now
2
−𝑎𝑡 2 − 𝑎𝑡 𝜔2
𝐹{𝑒 }=𝐹 𝑒 −
−𝑎𝑡 2 1 𝑒 4𝑎
1 𝜔 𝐹{𝑒 }=
= 𝐹 𝑎 2
𝑎 𝑎 1 𝜔2
= 𝑒 − 4𝑎
𝜔 2
2𝑎
𝑎
− 4
1 𝑒
=
𝑎 2
 Fourier Transform to Solve Differential Equations
• Solve the following differential equation
𝑑2 𝑥 𝑑𝑥
• + 2𝑘 + 𝜆2 𝑥 = 𝑓(𝑡)
𝑑𝑡 2 𝑑𝑡
Let’s take 𝐹 𝑥(𝑡) = 𝑋(𝜔) and 𝐹 𝑓(𝑡) = 𝐹(𝜔)

Applying F.T. to the both sides

𝑑2 𝑥 𝑑𝑥 2
𝐹 2
+ 2𝑘𝐹 + 𝜆 𝐹 𝑥 𝑡 = 𝐹{𝑓(𝑡)}
𝑑𝑡 𝑑𝑡
𝑖𝜔 2 𝑋 𝜔 + 2𝑘𝑖𝜔𝑋 𝜔 + 𝜆2 𝑋 𝜔 = 𝐹(𝜔)
𝐹 𝜔 = 𝜆2 − 𝜔2 + 2𝑖𝑘𝜔 𝑋(𝜔)
𝐹 𝜔
𝑋 𝜔 = 2
𝜆 − 𝜔 2 + 2𝑖𝑘𝜔
• Taking inverse F.T.,
∞
1
𝑥 𝑡 = න 𝑒 𝑖𝜔𝑡 𝑋 𝜔 𝑑𝜔
2𝜋
−∞
𝐹 𝜔
By replacing 𝑋 𝜔 by ,
𝜆2 −𝜔2∞
+2𝑖𝑘𝜔
1 𝐹 𝜔
𝑥 𝑡 = න 𝑒 𝑖𝜔𝑡 2 2
𝑑𝜔
2𝜋 𝜆 − 𝜔 + 2𝑖𝑘𝜔
−∞
Example
• Using the F.T. of the signals given
by,
𝐹 𝑥1 𝑡 ∗ 𝑥2 𝑡 = 𝑋1 𝜔 . 𝑋2 𝜔
𝑥1 𝑡 = 𝑡𝑒 −𝑡 𝑢(𝑡) ,
𝑥2 𝑡 = 𝑡𝑒 −2𝑡 𝑢(𝑡). 𝑥1 𝑡 ∗ 𝑥2 𝑡 = 𝐹 −1 𝑋1 𝜔 . 𝑋2 𝜔

Answer: 1 1
𝐹 𝑒 −𝑡 𝑢(𝑡) =
1
and = 𝐹 −1 .
1+𝑖𝜔 2 2+𝑖𝜔 2
1+𝑖𝜔
1
𝐹 𝑒 −2𝑡 𝑢(𝑡). =
2+𝑖𝜔 −2 1 2 1
= 𝐹 −1 + 2
+ + 2
(1 + 𝑖𝜔) 1 + 𝑖𝜔 (2 + 𝑖𝜔) 2 + 𝑖𝜔
Then using Frequency differentiation
property,
𝑥 𝑡 = −2𝑒 −𝑡 𝑢 𝑡 + t𝑒 −𝑡 𝑢 𝑡 + 2𝑒 −2𝑡 𝑢 𝑡 + 2𝑒 −2𝑡 𝑢(𝑡)
1
𝐹 𝑡𝑒 −𝑡 𝑢(𝑡) = and
1+𝑖𝜔 2

1
𝐹 𝑡𝑒 −2𝑡 𝑢(𝑡). =
2+𝑖𝜔 2
In elementary cases, we can use a Table of standard Fourier transforms
together, if necessary, with the appropriate properties of the Fourier
transform.
 Examples
• Find the Fourier Transform of the 7. 2𝑒 −𝑡 𝑢 𝑡 + 3𝑒 −2𝑡 𝑢 𝑡
following functions.
1. 𝑒 −𝑡 𝑢(𝑡)
2. 𝑒 −3𝑡 𝑢(𝑡) 4 ; 𝑡 ≤3
8. 𝑓 𝑡 = ቊ
3. 𝑒 −𝑡/2 𝑢(𝑡) 0 ; 𝑡 >3
1 ; −𝑎 < 𝑡 < 𝑎
4. 𝑝𝑎 𝑡 = ቊ
0 ; 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒 1 ;3 ≤ 𝑡 ≤ 5
9. g 𝑡 = ቊ
0 ; 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
1 ; 𝑡 ≤1
5. 𝑓 𝑡 = ቊ
0 ; 𝑡 >1
10. Find
1/4 ; 𝑡 ≤ 3
6. 𝑓 𝑡 = ቊ
0 ; 𝑡 >3 11. Find
• Find
END