EC 19-203-0403 Signals and Systems

-Dr. P. Mythili

B.Tech Electronics and Communication

School of Engineering,
Cochin University of science and Technology
FOURIER ANALYSIS
 What is Fourier analysis ??

 It is the process of breaking a function or signal or waveform into an alternate

representation, characterized by the sum of sines and cosines or complex
exponentials of different frequencies, phases and amplitudes.

 Fourier transformation is the mathematical procedure which transforms directly

from the time domain ‘t’ to the frequency domain ‘’
 What is the need for Fourier analysis?
By applying Fourier analysis to a signal , we can extract the parameters of any signal
like the
 fundamental frequency,
 the harmonics,
 the energy or power etc.
This can be further helpful in synthesizing, exploring and analyzing the signals.
In the case of systems
 We can find how the system responds to various frequencies (Frequency response)
which can be utilized to design systems based on the frequency response.
 Designing systems based on time domain equations may be difficult while it will
be easier to design based on frequency response. All LTI systems are also called as
filters
 Ideal Differentiator / Integrator
Time domain / Frequency domain
𝐻(𝜔) 𝑌 ( 𝑠 ) =𝑠 𝑋 (𝑠)

Differentiation |𝐻 ( 𝜔 )|=𝜔
𝑑𝑥 (𝑡 )
𝑦 (𝑡)=
𝑑𝑡

𝑦 ( 𝑡 ) =∫ 𝑥(𝑡)𝑑𝑡 𝐻(𝜔) 1
Integrator 𝑌 ( 𝑠) = 𝑋 (𝑠)
𝑠
Integration 1
|𝐻 ( 𝜔 )|=
𝜔

- 0  
 Different variations of Fourier analysis

 Fourier Series (F.S) – continuous time periodic signals

 Fourier Transform (F.T) – continuous time aperiodic signals
 Discrete Fourier Series (DFS) – discrete time periodic signals
 Discrete Time Fourier Transform (D.T.F.T) – discrete time aperiodic
signals
 Discrete Fourier Transform (DFT) – discrete time periodic and aperiodic
signals
Forward and Reverse transforms

Fourier series (F.S) :

For periodic continuous time signals, the magnitudes of all the discrete frequency
components present in the signal are given by the F.S coefficients ak

Dirichlet’s conditions (convergence criteria):

1. The signal must have finite number of maxima and minima within a single period
2. The signal must have finite number of discontinuities within a period
3. The signal must be Absolutely integrable i.e.
Fourier Transform (F.T)

For aperiodic continuous time (C.T) signals, the magnitudes of all the frequency
components present in the signal are given by X(

Dirichlet’s conditions (convergence criteria):

1. The signal must have finite number of maxima and minima
2. The signal must have finite number of discontinuities
3. The signal must be Absolutely integrable i.e.
Discrete Fourier series (D.F.S) :

For periodic discrete time (D.T) signals, the magnitudes of all the discrete
frequency components present in the signal are given by the F.S coefficients ak

Dirichlet’s conditions (convergence criteria):

1. The signal must be absolutely summable i.e.
Discrete time Fourier Transform (D.T.F.T)

For aperiodic discrete time (D.T) signals, the magnitudes of all the
frequency components present in the signal are given by X(

Dirichlet’s conditions (convergence criteria):

The signal must be absolutely summable i.e.
Spectrum
 It is the plot of the frequencies present in a signal versus the
magnitude/strength of the corresponding frequency components.
 vs k Magnitude Spectrum
 vs 
 vs k
Phase Spectrum
 vs 
 vs k - PSD
 Vs  - ESD
Example 1:
Determine the FS and power density spectrum of the rectangular pulse train

A 𝑥(𝑡)

𝜏 𝜏 t
− 2
2
𝑇𝑝
Solution :
Since the signal is CT periodic signal, F.S is used to find the spectrum
At k=0, this function is
indeterminate. To overcome this, we
can make it to the form where
 Spectrum
vs k

If
|𝑎𝑘|
Infinite frequencies /
0.25 𝐴 sinusoids of various
• Every ‘k’ corresponds to a amplitudes make a pulse
train
particular frequency ‘’

1
𝜔 𝑜=2 𝜋 𝐹 𝑜 , 𝐹 𝑜=
𝑇𝑝
−8 −4 - 0 2 4 8 12 𝑘
Example 2:
Determine the spectrum of the signal

𝑥(𝑡)
A

𝜏 𝜏
− 2 t
2

Solution :
This signal a CT aperiodic hence FT
can be applied.
 |𝑋 ( 𝜔)|
𝐴𝜏
Infinite frequencies /
sinusoids of various
amplitudes make a
square pulse

4𝜋 2𝜋 0 2𝜋 4𝜋 6𝜋 𝜔
− −
𝜏 𝜏 𝜏 𝜏 𝜏
Properties of F.S
Sl. No Time domain Frequency Domain

2. Linearity
3. Time shifting
4. Frequency Shifting

5. Convolution
6. Time reversal
7. Parsevals Theorem
Properties of F.T
Sl. No Time domain Frequency Domain

1. Linearity
2. Time shifting
3. Frequency Shifting

4. Convolution
5. Time reversal
6. Parsevals Theorem
Example 1:
Find the spectrum of
Solution:
Check for convergence criteria.

|𝑋 ( 𝜔)|
1

0 𝜔
Inverse Transforms
 Inverse Fourier series or transforms can be obtained using look up tables and
duality property.
 The Duality Property tells us that if has a Fourier Transform , then if we
have a function of time that has the functional form of the transform it will
have a Fourier Transform  that has the functional form of the original time
function (but is a function of frequency).
Time domain Frequency domain
|𝑋 ( 𝜔 )|
𝑥 ( 𝑡 ) =𝛿(𝑡 )
1
1

0 t 0 𝜔

𝑥(𝑡) 𝑋 ( 𝜔 ) =2 𝜋 𝛿(𝜔)
1
1

0 t 0 t
 𝐺(𝜔)

2𝜋 2𝜋
− 𝜔
𝜏 𝜏 𝜔

𝐺(𝜔)

4𝜋 2𝜋 2𝜋 4𝜋
− − − −
𝜏 𝜏 𝜏 𝜏
 Example :2
Find the spectrum of
Solution:
This signal is absolutely integrable.

Example:3
Find the spectrum of
Solution:

The signal is not absolutely integrable. This doesn’t mean there are no
frequencies
An alternate way to find the spectrum will be to make some approximations
and make it integrable.

As 𝑋(𝜔)

0 
Example 4:
Find the spectrum of
Solution:
Convergence criteria not satisfied as

𝑋(𝜔)

0 
Example 5:
Find the FT of the signal
Solution:
Use the frequency shifting property

Here

𝑋 ( 𝜔 ) =2 𝜋 𝛿(𝜔 − 𝜔1 )
1

0 𝜔1 t
Example : 6
Find the spectrum of
Solution:

Whenever we get expressions in the form of sines and cosines, we can

express it in terms of complex exponentials and find the spectrum using FT

Though we found out the FT, we ended with impulses, Therefore the Fourier
series coefficients can be extracted as
 |𝑋 ( 𝜔 )| |𝑎 𝑘|
½ ½

− 𝜔1 0 𝜔1  −1 0 1 k

The spectrum obtained using Fourier Transforms /

series have a mirror image. i.e., Negative frequencies
appear due to the mathematics involved in them
 Class work
 Find the spectrum of the signal

|𝑎 𝑘|
|𝑋 ( 𝜔 )|

−500 −100 0 1 00 500  −5 −1 0 1 5 

Problem :
Find the spectrum of a periodic pulse train or
𝑥(𝑡 )

… …

−4𝑇 −3𝑇 −2𝑇 −𝑇 0 𝑇 2𝑇 3𝑇 4 𝑇 t

Solution :

T is the period of the pulse.

  Since it is periodic, we can find the F.S coefficients

|𝑎 𝑘|
Each k corresponds to a
frequency
… …

−4 −3 −2 −1 0 1 2 3 4 k
𝑋 (𝜔)

… …

−
6 𝜋 4 𝜋 2𝜋
− −−
0 2𝜋 4 𝜋 2𝜋
𝜔
𝑇 𝑇 𝑇 𝑇 𝑇 𝑇
 The spectrum of a pulse train is another pulse train
System Function/ System response

 , where is the output response and is the input response and is the
system response.