Transcript Slide 1

Fourier Transforms
1
Background
While the Fourier series/transform is very important for
representing a signal in the frequency domain, it is also
important for calculating a system’s response (convolution).
• A system’s transfer function is the Fourier transform of its
impulse response
• Fourier transform of a signal’s derivative is multiplication in
the frequency domain: jwX(jw)
• Convolution in the time domain is given by multiplication in
the frequency domain (similar idea to log transformations)
Fourier Series in exponential form
Consider the Fourier series of the 2T periodic function:

a
˜f (x)  0  (a cos nx  b sin nx )
n
2 n 1 n
T
T
Due to the Euler formula
e i  cos   isin
It can be rewritten as

f˜ (x) 


inx
c
e
n
n 
(1)
With the decomposition coefficients calculated as:

1 T in T t
cn 
e
f (t)

2T T
(2)
3
Fourier transform
The frequencies are
n
wn 
T
and
w 

T
Therefore (1) and (2) are represented as

 iwx T iwt

1
f˜ (x)   e  e f (t)dtw
2 n   T


(3)
Since, on one hand the function with period T has also the periods kT for any integer k, and
on the other hand any non-periodic function can be considered as a function with infinite
period, we can run the T to infinity, and obtain the Riemann sum with ∆w→∞, converging to

the integral:
T


1
iwx
iwt
f˜ (x) 
e  e f (t)dtd w

2   T


(4)
4
Fourier transform definition
The integral (4) suggests the formal definition:
The funciotn F(w) is called a Fourier Transform of function f(x) if:

F(w) : F{ f (t)}:
The function
iwt
e
 f (t)dt
(5)


1
iwx
F 1{F(w)}:
e
F(w)dw

2 
(6)
 an inverse Fourier transform of F(w).
Is called

5
Example 1
The Fourier transform of
is
1 | t |1
f (t) : 
0 | t |1

sinw

2
F{ f (t)}   e iwt dt   w

1

 2
1
w0
w 0
The inverse Fourier transform is
2  iwx sin w
2
e
dw


2 
w



1



0

 coswx
0
sin w
dw 
w
sin(w(x 1))
1
dw 
w



0
 1 | x | 1

sin(w(x 1))
dw  1/2 x  1
w

x 1
 0
6
Fourier Integral
If f(x) and f’(x) are piecewise continuous in every finite interval, and f(x) is absolutely
integrable on R, i.e.
converges, then



1
1
iwx
iwt
[ f (x)  f (x)] 
e  e f (t)dtd w

2
2  


Remark: the above conditions are sufficient, but not necessary.

7
Properties of Fourier transform
1
Linearity:
For any constants a, b the following equality holds:
F{af (t)  bg(t)}  aF{ f (t)}  bF{g(t)}  aF (w)  bG(w)
Proof is by substitution into (5).
2
Scaling:
For any constant c, the following equality holds:
1
w
F{ f (ct)} 
F( )
|c | c
8
Properties of Fourier transform 2
3
Time shifting:
F{ f (t  t0 )}  e iwt0 F(w)
Proof:
4.


F{ f (t  t0 )}   f (t  t0 )e iwt dt  e iwt0  f (u)e iwudu



Frequency shifting:
F{e iwt0 f (t)}  F(w  w0 )

Proof:

F{e iw0 t f (t)}   e iwt0 f (t)e iwt dt  F(w  w0 )


9
Properties of Fourier transform 3
5.
Symmetry:
F{F(t)}  2f (w)
Proof:
The inverse Fourier transform is

therefore

1
iwt
f (t)  F { f (w)} 
F(w)e
dw

2 
1


1
itw
2f (w) 
F(t)e
dt  F{F(t)}

2 
10
Properties of Fourier transform 4
6.
Modulation:
1
F{ f (t)cos(w0 t)}  [F(w  w0 )  F(w  w0 )]
2
1
F{ f (t)sin(w0 t)}  [F(w  w0 )  F(w  w0 )]
2
Proof:
Using Euler formula, properties 1 (linearity) and 4 (frequency shifting):

1
F{ f (t)cos(w0 t)}  [F{e iw0 t f (t)} F{e iw0 t f (t)}]
2
1
 [F(w  w0 )  F(w  w0 )]
2
11
Differentiation in time
7.
Transform of derivatives
Suppose that f(n) is piecewise continuous, and absolutely integrable on R. Then
F{ f (n ) (t)}  (iw)n F(w)
In particular
and
F{ f '' (t)} 
F{ f ' (t)}  iwF(w)

w 2 F(w)
Proof:
From the definition of F{f(n)(t)} via integrating by parts.

12
Example 2
The property of Fourier transform of derivatives can be used for solution of differential
y 4y  H(t)e 4 t
equations:
0 t  0
H(t)  
1 t  0

4t
F{ y}  4F{y}  F{H(t)e } 
1
4  iw
Setting F{y(t)}=Y(w), we have


1
iwY(w)  4Y(w) 
4  iw
13
Example 2
Then
1
1
Y (w) 

(4  iw)(4  iw)
16  w 2
Therefore
1 4|t|
y(w)  F {Y(w)}   e
8
1


14
Frequency Differentiation
F{t n f (t)}  i n F (n ) (w)
In particular
F{tf (t)}  iF (w)
and
F{t f (t)}  F(w)
2

Which can be proved from the definition of F{f(t)}.


15
Fourier Transform
A CT signal x(t) and its frequency domain, Fourier transform signal,
X(jw), are related by

X ( jw )   x(t )e  jwt dt

x(t ) 
1
2
This is denoted by:



X ( jw )e jwt dw
analysis
synthesis
F
x(t )  X ( jw )
For example:
 at
F
e u (t ) 
1
a  jw
Often you have tables for common Fourier transforms
The Fourier transform, X(jw), represents the frequency content of
x(t).
It exists either when x(t)->0 as |t|->∞ or when x(t) is periodic (it
generalizes the Fourier series)
Linearity of the Fourier Transform
The Fourier transform is a linear function of x(t)
F
x1 (t )  X 1 ( jw )
F
x2 (t )  X 2 ( jw )
F
ax1 (t )  bx2 (t )  aX 1 ( jw )  bX 2 ( jw )
This follows directly from the definition of the Fourier
transform (as the integral operator is linear) & it easily
extends to an arbitrary number of signals
Like impulses/convolution, if we know the Fourier transform
of simple signals, we can calculate the Fourier transform
of more complex signals which are a linear combination
of the simple signals
Fourier Transform of a Time Shifted Signal
We’ll show that a Fourier transform of a signal which has a
simple time shift is:
F{x(t  t0 )}  e jwt0 X ( jw)
i.e. the original Fourier transform but shifted in phase by –wt0
Proof
Consider the Fourier transform synthesis equation:


w
X
(
j
w
)
e
dw
 
w
w
e
X
(
j
w
)
e
dw


 
x(t ) 
1
2
x(t  t0 ) 
1
2

1
2

X ( jw )e jwt dw

j ( t  t0 )


 j t0
j t

but this is the synthesis equation for the Fourier transform
e-jw0tX(jw)
Example: Linearity & Time Shift
Consider the signal (linear sum of two time
shifted rectangular pulses)
x(t)  0.5x1(t  2.5)  x 2 (t  2.5)
x1(t)
where x1(t) is of width 1, x2(t) is of width 3,
centred on zero (see figures)
 Using the FT of a rectangular pulse L10S7
X1( jw )  2sin(w /2) w
t
x2(t)
X 2 ( jw )  2sin(3w /2) w
 Then using the linearity and time shift
Fourier transform properties



 j 5w / 2 sin(w /2)  2sin(3w /2)
X( jw )  e

w 


t
x (t)
t
Fourier Transform of a Derivative
By differentiating both sides of the Fourier transform
synthesis equation with respect to t:
dx(t ) 1
 2
dt


jw X ( jw )e jwt dw

Therefore noting that this is the synthesis equation for the
Fourier transform jwX(jw)
dx (t ) F
 jwX ( jw )
dt
This is very important, because it replaces differentiation in
the time domain with multiplication (by jw) in the
frequency domain.
We can solve ODEs in the frequency domain using
algebraic operations (see next slides)
Convolution in the Frequency Domain
We can easily solve ODEs in the frequency domain:
F
y (t )  h(t ) * x(t )  Y ( jw )  H ( jw ) X ( jw )
Therefore, to apply convolution in the frequency domain, we
just have to multiply the two Fourier Transforms.
To solve for the differential/convolution equation using Fourier
transforms:
1. Calculate Fourier transforms of x(t) and h(t): X(jw) by H(jw)
2. Multiply H(jw) by X(jw) to obtain Y(jw)
3. Calculate the inverse Fourier transform of Y(jw)
H(jw) is the LTI system’s transfer function which is the Fourier
transform of the impulse response, h(t). Very important in
the remainder of the course (using Laplace transforms)
This result is proven in the appendix
Example 1: Solving a First Order ODE
Calculate the response of a CT LTI system with impulse response:
h(t )  ebt u(t )
b0
to the input signal:
x(t )  eat u(t )
a0
Taking Fourier transforms of both signals:
1
1
H ( jw ) 
, X ( jw ) 
b  jw
a  jw
gives the overall frequency response:
1
Y ( jw ) 
(b  jw )(a  jw )
to convert this to the time domain, express as partial fractions:
assume

1 
1
1


Y ( jw ) 

ba
b  a  (a  jw ) (b  jw ) 
Therefore, the CT system response is:
y(t )  b1a eat u(t )  ebt u(t )
Example 2: Design a Low Pass Filter
Consider an ideal low pass filter in frequency domain:
1 | w | wc
H(jw)
H ( jw )  
0 | w | wc
 X ( jw ) | w | wc
Y ( jw )  
| w | wc
 0
wc
wc
w
The filter’s impulse response is the inverse Fourier transform
w c j wt
h(t)
sin(wc t )
1
h(t )  2  e dw 
wc
t
0
t
which is an ideal low pass CT filter. However it is non-causal, so
this cannot be manufactured exactly & the time-domain
oscillations may be undesirable
We need to approximate this filter with a causal system such as 1st
order LTI system impulse response {h(t), H(jw)}:
F
1
1 y (t )
 at
a
 y (t )  x(t ),
e u (t ) 
t
a  jw
Convolution
The convolution of two functions f(t) and g(t) is defined as:

f *g 
f (u)g(t  u)du

Theorem:
Proof:

F{ f * g}  F{ f }F{g}
1

F{ f (t)g(t)} 
[F * G](w)
2



iwt
F{ f * g}   e  f (u)g(t  u)dudt 







iwu
iw(t u)
g(t  u)d(t  u)du  F{ f }F{g}
 e f (u) e





24
Appendix: Proof of Convolution Property
y(t) 


x(t )h(t  t )dt

Taking Fourier transforms gives:

Y( jw ) 
 




 jwt
x(
t
)h(t

t
)d
t
e
dt

Interchanging the order of integration, we have

Y( jw ) 
 x(t )




 jwt
h(t

t
)
e
dt dt

By the time shift property, the bracketed term is e-jwtH(jw), so

Y( jw ) 

 x(t )e
 jwt
H( jw )dt


 H( jw )  x(t )e  jwt dt

 H( jw )X( jw )
Summary
The Fourier transform is widely used for designing filters. You can
design systems with reject high frequency noise and just retain
the low frequency components. This is natural to describe in the
frequency domain.
Important properties of the Fourier transform are:
F
1. Linearity and time shifts ax(t )  by(t )  aX ( jw )  bY ( jw )
2. Differentiation
3. Convolution
dx (t ) F
 jwX ( jw )
dt
F
y (t )  h(t ) * x(t )  Y ( jw )  H ( jw ) X ( jw )
Some operations are simplified in the frequency domain, but there
are a number of signals for which the Fourier transform does not
exist – this leads naturally onto Laplace transforms. Similar
properties hold for Laplace transforms & the Laplace transform
is widely used in engineering analysis.
Lecture 11: Exercises
Theory
1. Using linearity & time shift calculate the Fourier transform of
x(t )  5e3(t 1)u(t 1)  7e3(t 2)u(t  2)
2. Use the FT derivative relationship (S7) and the Fourier
series/transform expression for sin(w0t) (L10-S3) to evaluate the FT of
cos(w0t).
3. Calculate the FTs of the systems’ impulse responses
y (t )
y (t )

3
y
(
t
)

x
(
t
)
3
 y (t )  x(t )
a) t
b)
t
4. Calculate the system responses in Q3 when the following input signal
is applied
5t
x(t )  e u(t )
Matlab/Simulink
5. Verify the answer to Q1 using the Fourier transform toolbox in Matlab
6. Verify Q3 and Q4 in Simulink
7. Simulate a first order system in Simulink and input a series of
sinusoidal signals with different frequencies. How does the response
depend on the input frequency (S12)?