FOURIER TRANSFORM
Learning Outcome:
In this topic, you will learn about:
o the Fourier transform by calculating simple Fourier transforms from
the definition, state how the Fourier transform of a function (signal)
depends on whether that function is even or odd or neither.
o some of the properties of the Fourier transform.
o Finally, you will learn about some special Fourier transform pairs.
Prerequisites:
Before starting this topic you should be familiar with basic Fourier
series, particularly in the complex form.
What is Fourier Transform?
Fourier transform is a basic tool of applied mathematics in which, it is a
mathematical model that decomposes a function or signal into its
constituent frequencies. It is the generalized form of the complex Fourier
series. The Fourier transform is a powerful mathematical tool used to
transform a function or signal from its original time or space domain into
the frequency domain. It helps to transform the signals between two
different domains like transforming the frequency domain to the time
domain. It is a powerful tool used in many fields, such as signal processing,
physics, and engineering, to analyze the frequency content of signals or
functions that vary over time or space. It is particularly used for solving
differential equations (especially partial differential equations) and also in
conjunction with integral equations. Understanding its properties and
applications is essential for engineers. A transform is a mapping between
two domains.
The convergence criteria of the Fourier transform (namely, that the
function be absolutely integrable on the real line) are quite severe due to
the lack of the exponential decay term as seen in the Laplace transform,
and it means that functions like polynomials, exponentials, and
trigonometric functions all do not have Fourier transforms in the usual
sense. However, we can make use of the Dirac delta function to assign
these functions Fourier transforms in a way that makes sense.
Unlike Fourier series, which are mainly useful for periodic functions, the
Fourier transform permits alternative representations of mostly non-
periodic functions. We shall study its various properties.
We will explore the Fourier transform in detail along with the formula,
forward and inverse Fourier transforms, and its properties. There are really
three Fourier transforms, namely:
1. the Fourier Sine transforms
2. Fourier Cosine transforms and
3. A complex form which is usually referred to as the Fourier transform.
Types of Fourier Transforms:
There are several types of Fourier transforms, including
1. the continuous Fourier transform (CFT),
2. the discrete Fourier transform (DFT),
3. the fast Fourier transform (FFT), and
4. the wavelet transforms (WT).
Continuous Fourier Transform (CFT)
For a continuous-time function f(t), the Fourier transform F(ω) is defined
∞
as: 𝑭(𝝎) = ∫−∞ 𝒇(𝒕)𝒆−𝒊𝝎𝒕 𝒅𝒕
where:
• F(ω) is the Fourier transform of f(t)
• ω is the Angular Frequency
• i is the Imaginary Number (𝒊𝟐 = −𝟏)
• t is Time
Fourier transforms helps to expand the non-periodic functions and
convert them into easy sinusoid functions.
There are two types of Fourier transform i.e., forward Fourier transform
and inverse Fourier transform.
Fourier Transform Formula
The formula for the Fourier transforms of a function f(x) is given by:
∞
𝒇(𝒙) = ∫ 𝑭(𝒌)𝑒 𝟐𝝅𝒊𝑲𝒙 𝒅𝒌
−∞
∞
F(k) = ∫−∞ 𝑓(𝑥)𝑒 −𝟐𝝅𝒊𝑲𝒙 𝒅𝑥
Forward Fourier Transform
The forward Fourier transform is a mathematical technique used to
transform a time-domain signal into its frequency-domain representation.
Forward Fourier Transform is represented by F(U). The forward Fourier
∞
transform is defined as: F(U) = ∫−∞ 𝒇(𝒙)𝑒 −𝟐𝝅𝒊𝑈𝒙 𝒅𝒙
If 𝑓(𝑥) is defined on , and 𝑓(𝑥) is square integrable
is finite) then the Fourier transform of 𝑓(𝑥) is given by 𝐹(𝑢) below. 𝐹(𝑢) is
then also square integrable and 𝑓(𝑥) can be recovered form 𝐹(𝑢) by using
the formula for the inverse Fourier transform.
Note that 𝑓 is a function of 𝑥, while 𝐹 is a function of 𝑢. In practical
applications 𝑥 may be measured for example in 𝑚, then 𝑢 is measured in
𝑚−1 , or 𝑥 may be measured for example in 𝑠, then 𝑢 is measured in 𝑠 −1 , i.e.,
Hz. The variable 𝑥 is referred to as ‘physical space’ while the variable
𝑢 is referred to a ‘frequency space’. Both 𝑓(𝑥) and 𝐹(𝑢) represent the same
function: 𝑓(𝑥) is a physical space representation and 𝐹(𝑢) is a frequency
space representation.
Note also, the there are many choices of representations for example,
the equation:
And the inverse is
Note that these equations use a 𝜉 (the Greek letter Xi) to imply frequency
instead of 𝜔 (Omega) which generally refers to angular frequency (𝜔 =
2𝜋𝜉). The Fourier transform of a time dependent signal produces a
frequency dependent function.
A lot of engineers use omega because it is used in transfer functions, but
here we are just looking at frequency. If we use the angular frequency
instead of frequency, then we would have to apply a factor of 2𝜋 to either
the transform or the inverse. The general rule is that the unit of the
Fourier transform variable is the inverse of the original function’s
variable.
Inverse Fourier Transform
The inverse Fourier transform is the process of converting a frequency-
domain representation of a signal back into its time-domain form.
This is the reverse process of the forward Fourier transform. Inverse
Fourier Transform is defined as:
∞
−1
𝒇(𝒙) = 𝐹𝐾 [𝑭(𝒌)] (𝒙) = ∫ 𝑭(𝒌)𝑒 𝟐𝝅𝒊𝑲𝒙 𝒅𝒌
−∞
Properties of Fourier Transform
Various properties of Fourier transform are:
• If 𝑎(𝑡) has a Fourier transform 𝐴(𝑓), then Fourier transform of 𝐴(𝑡)
is 𝑎(−𝑓). It is called the duality property.
• Fourier transform is a linear transform. It is called linear transform.
• Modulation property is the property in which the function is
modulated by other function.
• A shift in the time domain corresponds to a phase shift in the
frequency domain in Fourier Transform
• Multiplying a time-domain signal by a complex exponential
corresponds to a shift in the frequency domain in Fourier
Transform
• In Fourier Transform taking the complex conjugate of the time-
domain signal corresponds to taking the complex conjugate of the
frequency-domain signal and reversing the frequency.
• The Fourier transform and its inverse are linear operators, and
therefore they both obey superposition and proportionality.
•
•
•
Fourier Transform Table
The table below shows the Fourier transform of various functions.
Functions f(x) F(k) = Fx[f(x)]
1 1 𝛿(𝑘)
1
Sine Function 𝑠𝑖𝑛(2𝜋𝑘0𝑥) ( ) × 𝑖 × [𝛿 (𝑘 + 𝑘0) − 𝛿 (𝑘 − 𝑘0)]
2
Cosine Function 𝑐𝑜𝑠(2𝜋𝑘0𝑥) (1/2) × [𝛿(𝑘 + 𝑘0) + 𝛿(𝑘 − 𝑘0)]
Inverse Function −𝑃𝑉(1/𝜋𝑥) 𝑖[1 − 2𝐻(−𝑘)]
Exponential Function 𝑒 −2𝜋𝑘0|𝑥| (1/𝜋)[𝑘0 / (𝑘2 + 𝑘20)]
Gaussian Function 𝑒 −𝑎𝑥2𝑒−𝑎𝑥 2
𝜋𝑎𝑒 − 𝜋2𝑘2/𝑎𝑎𝜋𝑒 − 𝜋2𝑘2/𝑎
Applications of Fourier Transform
Some applications of Fourier transform are as follows:
• Fourier transforms are used in signal processing,
telecommunications, audio processing, and image processing.
• Fourier transforms are used to reduce noise, compression, etc.
• It is also used to represent the wave propagation, analysis of
electrical signals and many more.
• The special form of Fourier transforms is used to represent
periodic functions and infinite series in mathematics.
Fourier transform is a common and useful engineering tool for
analyzing signals and vibrations, but sometimes it can produce some
hard to interpret results.
Euler's Formula
A key property of complex numbers is called Euler’s formula, which
states:
This exponential representation is very common with the Fourier
transform. Note that an imaginary number of the format R + jI can be
written as Aejξ where A is the magnitude and ξ is the angle.
Examples of the Fourier Transform
To find the Fourier transform of a function,
first substitute the function into the definition of the Fourier
transform. As with the Laplace transform, calculating the Fourier
transform of a function can be done directly by using the definition. We
will use the example function:
Which definitely satisfies our convergence criteria.
Secondly, Evaluate the integral using any means possible. This integral
resist the techniques of elementary calculus, but we can make use of
residue theory instead. To use residues, we create a
contour 𝜸 consisting of a concatenation of the real line and a
semicircular arc in the lower half plane that circles clockwise. The goal
is to show that the real integral equals the contour integral by showing
that the arc integral vanishes.
We may factor the denominator to show that the function has simple
poles at 𝒕±= ±𝒊. Since only 𝒕 − is being enclosed, we can use the residue
theorem to calculate the value of the contour integral.
SOLVED PROBLEMS:
Example 2 Find the Fourier transforms of:
(a)
(b)
(c)
by making use of the standard
integral
(d)
In each case confirm
that the Fourier transform vanishes as
Solution
(a)
As sin ωa is bounded, it follows directly that
(b)
As the numerator of G(ω) is bounded, it follows that
This
example shows that although f (x) may be real, its Fourier transform
can be
complex.
(c)
The integrand of the second integral is odd, so the value of the integral
is zero. Using the standard resultlt
in the remaining integral on the right, we find that
In this case the factor
ensures that
(d)
As the numerator of the Fourier transform is bounded, the
denominator causes the transform to vanish as
This example shows that a complex function can also have a Fourier
transform and, in general, that the transform will be complex.
Example 3: What is the Fourier transform of the function 𝑥(𝑡) = 2?
Solution:
Plugging this equation into the Fourier transform, we get:
Integrals around infinity start to behave oddly, so this example will not
be mathematically rigorous, but intuitively we can see that the integral
of 𝑠𝑖𝑛(𝜉𝑡) and 𝑐𝑜𝑠(𝜉𝑡) as 𝑡 goes from negative infinity to positive infinity
should be 0, unless 𝜉 = 0. At that point the equation simplified
dramatically to:
We can write the equation for 𝑋(𝜉) using the Dirac delta function, 𝛿(𝑥),
which is defined as:
Example 4: Find the Fourier transform of 𝑓(𝑥), where:
Solution:
Example 5: Find the Fourier transform of 𝑓(𝑥), where 𝜔 is a constant.
Solution:
Example 6: Find the Fourier transform of the function using Euler
formula approach 𝒆^{−|𝒙|}.
Solution:
THE SIN AND COS TRANSFORMS
It is important to note that square integrable functions have the
property that lim 𝑓(𝑥) = 0. This is also true for any derivative of a
𝑥→+∞
square integrable function.
The Fourier transform is applied to functions defined on 𝑥 ∈ (−∞, ∞), and
therefore it can only be used in the solution od pdes where the domain
is 𝑥 ∈ (−∞, ∞). Often the domain is rather 𝑥 ∈ [0, ∞) and we need to have
a transform with appropriate derivative properties that acts on this
domain.
Both the Sin Transform and the Cos transform can perform this role.
With both the Sin transform and the Cos transform, there is, however,
an extra piece of information that is needed at the point 𝑥 = 0 in order
to apply their respective derivative properties. But, in a well-posed
boundary condition on the domain 𝑥 ∈ [0, ∞) , there is also an extra
condition needed at the point 𝑥 = 0, and this supplies the information
needed by the Sin or Cos transform.
The Cos Transform
We now introduce a new transform called the Cos-transform,
abbreviated with CosTF. It is actually not really new, it is just a variant of
the FT. We shall use the notation
to denote the CosTF of the function f(x), where the frequency variable
associated with 𝑥 is 𝜉. We shall also often employ the shorter notation
FC(𝜉 ) to denote the CosTF of 𝑓(𝑥).
Consider an even real function 𝑓(𝑥) defined on 𝑥 ∈ [0, ∞). Its Fourier
transform is given by
Let us define the relationship between the CosTF and the FT as
In other words
Although the original 𝑓(𝑥) was defined over the entire real line, the CosTF
can be viewed as a transform for functions only defined on 𝑥 ∈ [0, ∞).
Note that 𝐹𝐶 (𝜉) is an even function of 𝜉 and at this stage it may be defined
for all 𝑥 ∈ (−∞, ∞), even though 𝑥 lies only in [0, ∞).
In order to transform back, note that
Therefore, Inverse 𝐶𝑜𝑠 transform, abbreviated by I𝐶𝑜𝑠TF, is given by
We now have a transform that operates on the region [0, ∞), and this
can be used to solve boundary value problems on the semi-infinite
domain 𝑥 ∈ [0, ∞), 𝑡 ∈ [0, ∞).
