Chapter 03

Spectral Analysis
Fourier analysis
 Fourier analysis is a method of defining periodic waveforms in terms of trigonometric functions.
 The method gets its name from a French mathematician and physicist named Jean Baptiste
Joseph, Baron de Fourier, who lived during the 18th and 19th centuries.
 Fourier analysis is used in electronics, acoustics, and communications.

 Many waveforms consist of energy at a fundamental frequency and also at harmonic frequencies.
The relative proportions of energy in the fundamental and the harmonics determines the shape of
the wave.
 The wave function (usually amplitude, frequency, or phase versus time) can be expressed as of a
sum of sine and cosine functions called a Fourier Series, uniquely defined by constants known as
Fourier coefficients.
 If these coefficients are represented by a , a 1 , a 2 , a 3 , ..., a n , ... and b 1 , b 2 , b 3 , ..., b n , ..., then
the Fourier series F (x), where x is an independent variable (usually time), has the following form:
F ( x ) = a /2 + a 1 cos x + b 1 sin x + a 2 cos 2 x + b 2 sin 2 x + ... + a n cos nx + b n sin nx + ...
 In Fourier analysis, the objective is to calculate
coefficients a , a 1 , a 2 , a 3 , ..., a n and b 1 , b 2 , b 3 , ..., b n up to the largest possible value of n . The
greater the value of n, the more accurate is the Fourier-series representation of the waveform.
Fourier series and Fourier transform
• To know the amplitude and phase of each frequency component
contained in the signal, Fourier analysis is performed.
• If signal is power signal, then fourier series is performed.
• If signal is energy signal, then fourier transform is performed.
• Fourier transform is the process of converting time domain signal to
frequency domain.
Example: Sum of Sinusoids
Consider the CT signal given by
x(t )  A1 cos(t )  A2 cos(4t   / 3)  A3 cos(8t   / 2),
t 
The signal has only three frequency components at 1,4, and
8 rad/sec, amplitudes A1 , A2 , A3 and phases0,  / 3,  / 2
The shape of the signal x(t) depends on the relative
magnitudes of the frequency components, specified in
terms of the amplitudes A1 , A2 , A3
 A1 0.5

 A2 1
 A 0
 3
Fourier Series
 As a matter of fact, sine waves and cosine waves are the basic
building functions for any periodic signal.
 Fourier series is a tool used to analyze any periodic signal.
 We obtain the following information about the signal:
 How many frequency component are present in the signal?
 Their amplitudes
 Their relative phase difference between these frequency components.
Types:
1. Trigonometric
2. Exponential
3. Polar
Drichlet’s condition for Fourier Series
To be described by the Fourier Series the waveform f(t) must satisfy
the following mathematical properties:

1. f(t) is a single-value function except at possibly a finite number

of points.
t0 T
2. The integral t0
f (t ) dt   for any t0.

3. f(t) has a finite number of discontinuities within the period T.

4. f(t) has a finite number of maxima and minima within the
period T.

In practice, f(t) = v(t) or i(t) so the above 4 conditions are always satisfied.
Trigonometric Form
 It is simply a linear combination of sines and cosines at
multiples of its fundamental frequency, f0=1/T.
 
xt  a0   an cos2f 0 nt   bn sin 2f 0 nt 
n 1 n 1

 a0 counts for any dc offset in x(t).

 a0, an and bn are called the trigonometric Fourier Series
Coefficients.
 The nth harmonic frequency is nf0.
Trigonometric Form
To find a0 , an , bn :

1
a0  xt dt
T0 T0
2
an  xt cos2nf 0t dt
T0 T0

2
bn  xt sin 2nf 0t dt
T0 T0
 Example
f(t)

1 f t  a0   an cos2nt  bn sin 2nt 
e-t/2 n 1

1  2 t
 
-p 0 p a0   e dt   e  1 0.504
2 2
 0  
• Fundamental period t
2    2 
T0 = p an   e 2
cos2nt  dt 0.504  2 
• Fundamental frequency  0
 1  16n 
t
f0 = 1/T0 = 1/p Hz 2    8n 
w0 = 2p/T0 = 2 rad/s
bn   e sin 2nt  dt 0.504 
2
2 
 0
 1  16n 
an and bn decrease in amplitude as n  .


 2 
f t  0.504 1   2
cos2nt  4n sin 2nt 
 n 1 1  16n 
Polar Form
• Using single sinusoid,

xt   C0   Cn cos2nf 0t   n 
 n 1
       
dc component nth harmonic

Where, C0 a0 and Cn , and  n are related to the trigonometric

coefficients an and bn as:

2 2
Cn  an  bn and

 bn 
1
 n  tan  
 an 
Compact Trigonometric

f(t) f t  C0   Cn cos2nt   n 
1 n 1
e-t/2 a0 0.504
 2 
-p 0 p an 0.504  2 
 1  16n 
• Fundamental period
 8n 
T0 = p bn 0.504  2 
 1  16n 
• Fundamental frequency C0 ao 0.504
f0 = 1/T0 = 1/p Hz
 2 
w0 = 2p/T0 = 2 rad/s
2

Cn  a  b 0.504 2

n n 2 
 1  16n 
  bn 
1
 n tan    tan  1 4n
 an 

2
f t  0.504  0.504
2

cos 2nt  tan  1 4n 
n 1 1  16n
Exponential Fourier Series
COMPLEX EXPONENTIALS

e jn0t cos n0t  j sin n0t

 jn0t
e cos n0t  j sin n0t

1 jn0t
cos n0t  e
2
 e  jn0t 
1 jn0t
sin n0t 
2j
e e  jn0t j jn0t

 e
2
  e  jn0t 
Exponential Fourier Series
A periodic signal x(t) is expressed in the exponential Fourier Series form as:

x(t )   Cn e jn0t

n 

Where, Cn is the coefficient of exponential Fourier series and can be calculated as:

t 0 T t 0 T
1 1
 
 jn0t
Cn  x(t )e dt C0  A0  x(t )dt
T T
t0 t0
Example:
Obtain the exponential Fourier series for the waveform shown
in Figure. Also draw the frequency spectrum.

Figure
Solution:
The periodic waveform shown in Figure 3 with a period T = 2π
can be expressed as:

 A, 0 t 
x(t ) 
 A,  t 2
Let t 0 0, t 0  T 2
2 2
And fundamental frequency: 0   1
T 2
 Solution:

Exponential Fourier Series:

T   2 
C0 
1
T 
x(t )dt 
1 
2 


Adt   Adt 
A
 2
t 0  t  0


2
 
0 0  
T
1

 jn0t
Cn  x(t )e dt
T
0
  2 
1   2  A   e  jnt  e  jnt 
 jnt 

A(t )e      
jnt
 dt  A(t )e dt 
2   2    jn    jn  
0    0   
Cont.…

A
jn2
 e

 jn
 0 
e 0   e  j 2n
e 
 jn 2 



A
jn2
 n

( 1)  1  1  ( 1)  j
2A
n
n

 2A
 j , for _ odd _ n
C n   n
 0, for _ even _ n
 
2 A jnt
x(t ) C0  C e
n 
n
jn0t
 
n 
 j
n
e , for _ odd _ n
n 0
2A 2A 2A
C0 0, C1 C  1  , C3 C  3  , C5 C  5 
 3 5
Cont.…
2A 2A 2A
C0 0, C1 C  1  , C3 C  3  , C5 C  5 
 3 5

2A 2A
 Cn 
2A 2A
2A 3 3
2A
5
5

-5 -4 -3 -2 -1 0 1 2 3 4 5
Fourier Transform
 We have seen that periodic signals can be represented with the
Fourier series
 Can aperiodic signals be analyzed in terms of frequency
components?
 Yes, and the Fourier transform provides the tool for this analysis
 The major difference w.r.t. the line spectra of periodic signals is
that the spectra of aperiodic signals are defined for all real values
of the frequency variable  not just for a discrete set of values
The Fourier Transform in the General Case

• Given a signal x(t), its Fourier transform X ( ) is defined as:


 j t
X ( )  x(t )e dt ,   

• A signal x(t) is said to have a Fourier transform in the ordinary sense if the
above integral converges
• Given a signal x(t) with Fourier transform X ( ) , x(t) can be recomputed
from X ( ) by applying the inverse Fourier transform given by:

1 j t
x(t ) 
2 

X ( )e d , t  

x(t )  X ( )
Drichlet condition for fourier transform

• The function x(t) should be single valued in any finite time interval T.
• The signal x(t) is absolutely integral i. e.
• The number of maxima and minima in any finite length of real time is
finite.
• The discontinuities in any finite length of real time is finite.
Properties of the Fourier Transform

x(t )  X ( ) y (t )  Y ( )
• Linearity:
 x(t )   y (t )   X ( )   Y ( )
• Left or Right Shift in Time:
 j t0
x(t  t0 )  X ( )e
• Time Scaling:
1  
x(at )  X  
a a
Properties of the Fourier Transform

• Time Reversal: x( t )  X (  )

n
n nd
• Multiplication by a Power of t: t x (t )  ( j ) n
X ( )
d
j 0 t
• Multiplication by a Complex Exponential: x(t )e  X (   0 )

• Multiplication by a Sinusoid (Modulation):

j
x(t )sin( 0t )   X (   0 )  X (   0 ) 
2
1
x(t ) cos( 0t )   X (   0 )  X (   0 ) 
2
Properties of the Fourier Transform
n
d n
• Differentiation in the Time Domain:
n
x(t )  ( j ) X ( )
dt

• Convolution in the Time Domain: x(t )  y (t )  X ( )Y ( )

• Multiplication in the Time Domain: x(t ) y (t )  X ( )  Y ( )

Example: Impulse Function (t)

 e j0

F ()  F (t)   (t)e dt  e0 1
 jt t  jt



(t)  1
1  2()

F (t)
1


1 if t  0
(t)  Delta function has unity area.
0 if t 
0
Example: Fourier Transform of Single Rectangular Pulse

1 for  t
f (t)  rectt   II(t  2 2
/)  
0 for all t  2

f (t)  rectt   II(t 

/ 2
/ )
F ()   f (t)e jt dt   e jt
Pulse of 1
width  dt  jt / 2  j/ 2
ej/ 2 e
 e 

 
/ 2  j
  j / 2
 2  2 time t
 j2sin /2 sin /2
2
 
0  j ( /2
Remember  = 2f
 sin /2
F ()  
sincf
 /2
1
sinc function
F()


 time t
 2 2
2
0
Note the pulse is
time centered
  2 3 
 3  2
    
 0
Fourier Transform of Complex Exponentials


F 1  ( f  fc ) ( f  fC )e j2f t

df

Evaluate for


f  fc
F 1  ( f  fc )   j2fct 
e df  e

 f
f j2fct
c
 j2f ct
 ( f  fc )  and
e 

F  ( f  cf ) 
1
C
( f  f
)e j2f t df

Evaluate for
f   fc
F 1  ( f  cf ) 
f  fc
e j2fct df  e
 j2fct
 ( f  fc )  j2f ct

e
Power and Energy of signal
Energy signal
• Energy is finite and non zero.
• Non periodic signal are energy signal.
• These signals are time limited.
• The power of energy signal is zero.
• Example: A single rectangular pulse.

Power signal
• The normalized average power is finite and non zero.
• Periodic (practical) signal are power signals.
• These signals can exist infinite time.
• Energy of power signal is infinite.
• Example: A train of rectangular pulse.
Parseval’s Theorem for Energy signals:
• Parseval's theorem refers to that information is not lost in Fourier transform.
• It states that the total energy of the signal x(t) is equal to the sum of energies of the
individual spectral components in the frequency domain.
• We can find the energy of signal without knowing its time domain according to
Parseval’s theorem.
• Suppose that x(t) and X() are the fields in the time and frequency domain respectively,
where X() is obtained by the Fourier transform of x(t). According to Parseval's theorem,
the following equation holds:
Proof:

• Let there be a signal x(t) with its Fourier transform

X(𝛚) . The energy E of signal x(t) is expressed as:

Inverse Fourier transform x(t) can be defined as:

x(t)
Replacing one x(t)
Interchanging the order of integration, we get

We know that,
X( so, X(

Therefore,

For a real signal x(t), the Fourier transform X() and X(-) are complex conjugates.
Hence
=
Energy Spectral Density:
• The squared amplitude spectrum is called as the Energy spectral density or energy
density spectrum. Defined as:

ESD = ψ(f) =

• Let us apply x(t) to an ideal low pass filter

• The response or output of a system is expressed as:
Y()=X()H()
Where, Y() and X() are Fourier transform of y(t) and x(t)
• Using Parseval’s theorem
• Here H()=0 for all the frequencies except for the narrow band - m to m for which it is unity.
Therefore,
=
• Again we may assume that FT X() is constant with frequency for a narrow band - m to m .
• The energy of the signal over transmission band ΔB = 2Bm will be
= (2Bm)
Substituting 2Bm = Δ
(Δ
=
Where, represents energy per unit bandwidth and is known as ESD or
energy density spectrum, denoted by Ψ().

Ψ() =
Parseval’s Theorem for Power Signals:
• This theorem relates average power of a periodic signal to its fourier series
coefficients
• It states the total average power of a periodic signal x(t) is equal to the sum of
average powers of individual fourier coefficients (Cn).
• The average power of a signal x(t) is defined as:

• If the signal can be represented as sum of fourier series components and power is
defined as:

Where Cn is the amplitude of nth harmonic component of fourier series.

Proof:
Power Spectral Density Function:

•In applying frequency-domain techniques to the analysis of random signals

the natural approach is to Fourier transform the signals.

•Unfortunately the Fourier transform of a stochastic process does not,

strictly speaking, exist because it has infinite signal energy.

•But the Fourier transform of a truncated version of a stochastic

process does exist.
• The expression for power spectral density may be derived by
assuming the power signal as a limiting case of an energy signal.
• Let us consider a power signal x(t) which is extended to infinity.
• The signal x τ(t) may be expressed as
x τ(t) = x(t) < τ/2

0 elsewhere
Now the terminated signal is of finite duration τ, therefore it is an
energy signal.
• Let the energy of this signal is denoted by E τ
Eτ=
Where if fourier transfer of
• It may be observed that x(t) over the interval (- τ/2, τ/2) will be same
as over the interval (-∞, ∞)

So,

=
And,
P=
In the limit may be approach a finite value and is represented by S()
such that S()= known as PSDF.
Auto correlation function and Psdf:
• The correlation function of two signal x1(t) and x2(t) is defined as,
τ) =
Where, τ is delayed parameter.
• It is measure of similarity between the signal x1(t) and time delay
version of another signal x2(t). This correlation is called cross-
correlation.
• Special case of cross-correlation is auto correlation in which both
signal are same. i.e. x1(t) = x2(t) = x(t)
τ) =
Properties of Autocorrelation function:

1. The autocorrelation function and psdf form a Fourier

transform pair.
2. If τ) and -τ) are the complex conjugates of each other then
they must be equal.
3. The autocorrelation function at origin is equal to the average
power of a signal.
4. The autocorrelation function has maximum value at the
origin.
Psdf of white noise
• Noise signals are random signals so, any two samples of noise signals are
uncorrelated.
• The special case of noise signal is white noise whose psdf is independent of
frequency ( i.e, constant over entire frequency range).
• The signal contains equal power within a fixed bandwidth at any center frequency.
 What is SPECTRUM
ANALYZER
 The problems associated with non-real-time analysis in the frequency
domain can be eliminated by using a spectrum analyzer. A spectrum
analyzer is a real-time analyzer, which means that it simultaneously
displays the amplitude of all the signals in the frequency range of the
analyzer.

 Spectrum analyzers, like wave analyzers, provide information about the

voltage or energy of a signal as a function of frequency. Spectrum
analyzers provide a graphical display on a CRT.
 SPECTRUM
ANALYZER
• A spectrum analyzer measures the
magnitude of an input signal
versus frequency within the full
frequency range of the instrument.

• The input signal that a spectrum

analyzer measures is electrical,
however, spectral compositions of
other signals, such as acoustic
pressure waves and optical light
waves, can be considered through
the use of an appropriate transducer.
 SPECTRUM ANALYZER
• By analyzing the spectra of
electrical signals, dominant
frequency, power, distortion,
harmonics, bandwidth, and other
spectral components of a signal
can be observed that are not easily
detectable in time domain
waveforms.

• The display of a spectrum analyzer

has frequency on the horizontal
axis and the amplitude displayed
on the vertical axis.
Analog spectrum analyzer
• A spectrum analyzer is an instrument is used to visualize and measure PSDF of
power type signal.
• Measurement of PSDF is equivalent to passing the signal x(t) through a narrow
band filter whose centered frequency can be sweep from near DC to infinity and
observing or measuring output of the filter at each and every tuned frequency.
• As we know if x(t) is passed through a NB(narrow band) filter tuned at wo the
average power of the output signal y(t) would be:
The above (i) can be obtained by passing x(t) through the NB filter centered at
f0, squaring the o/p of the filter and integrating it.

Now if we want to measure PSDF at frequency range we need weep the

centered frequency of NF form zero to ∞ .
The accuracy of PSDF measurement increases with decrease in the bandwidth
of narrow band filter.
In practice it is very difficult to weep the centre frequency of narrow band
filter as they are constructed using quartz crystals.
Therefore another method in which narrow band filter is tuned to a precise
frequency and the x(t) is frequency shifted in such a way that the frequency
range of frequency shifted x(t) sequentially fall within the pass band of the
filter, is employed to measure PSDF.
Figure : Analog spectrum analyzer
Assignment
• PSDF of harmonic signal
• Prove that in a LTI system when a power signal x(t) is applied the PSD
of its output is equal to the PSD of its input multiplied by the squared
amplitude response of the system.
Thank
You