Ch03 - Fourier Representations of Signals LTI Systems PDF

CHAPTER
Fourier Representations of Signals & LTI Systems
3.1 Introduction
1. A signal can be represented as a weighted superposition of complex
sinusoids.
x(t) or x[n] y(t) or y[n]
2. LTI system: LTI System
Output = A weighted superposition
of the system response to each
complex sinusoid.
3. Four distinct Fourier representations
3.2 Complex Sinusoids and Frequency Response of LTI
Systems
 Frequency response  The response of an LTI system to a sinusoidal input.
 Discrete-time LTI system
1. Impulse response of discrete-time LTI system = h[n], input = x[n] = e j n
2. Output:
 
yn   hk xn  k    hk e j  n  k 
k   k  
Signals and Systems_Simon 1

Haykin & Barry Van Veen
CHAPTER
Figure 3.1 (p. 196)

The output of a complex sinusoidal input to an LTI system is a complex sinusoid of
the same frequency as the input, multiplied by the frequency response of the system.

y n  e jn
 h  k  e  jk
 H  e j
 e jn
k 
Complex scaling factor
3. Frequency response:

A function of frequency 
j
H (e )   h[k ]e
k 
 j
(3.1)
 Continuous-time LTI system

1. Impulse response of continuous-time LTI system = h(t), input = x(t) = e j  t
2. Output:  
y (t )   h( )e j ( t  ) d  e j t  h( )e  j d
  (3.2)
 H ( j )e j t
CHAPTER

H ( j )   h( )e j d (3.3)

 Polar form complex number c = a + jb:
c  c e j arg{c} where c  a 2  b2 and arg{c}  tan 1 

b
a
 Polar form for H (j):
H ( j )  H ( j ) e j arg{H ( j )}
where H ( j )  Magnitude response and argH ( j )  Phase respnse
 
j  t  argH  j 
y  t   H  j  e
Example 3.1 RC Circuit: Frequency response
The impulse response of the system relating to the input voltage to the voltage
across the capacitor in Fig. 3.2 is derived in Example 1.21 as
1  t / RC Find an expression for the frequency response, and
h( t )  e u(t ) plot the magnitude and phase response.
RC
CHAPTER
<Sol.> Frequency response:
  1 
   j  
H  j   u  e
1 1 

 j 
e RC
d   e  RC 
d
RC 
RC 0

 1 
1 1  j   Figure 3.2 (p. 197)
 e  RC 
RC circuit for Example 3.1.
RC  1 
 j   Magnitude response:
 RC  0
1 1

1
0  1
RC  1  H  j   RC
 j    2
 RC   1 
2   
1  RC 
 RC Phase response:
argH  j    arctanRC 
1
j 
RC
Fig. 3.6 (a) and (b)

CHAPTER
H  j  arg H  j 

2

4



 4


2
Figure 3.3 (p. 198)
Frequency response of the RC circuit in Fig. 3.2. (a) Magnitude response.
(b) Phase response.
 Eigenvalue and eigenfunction of LTI system
[A] Continuous-time case:
1. Eigenrepresentation: Fig. 3.4 (a) H  t    t 

CHAPTER
 (t )  (t ) e j t
H ( j )e j t
e jn H  e j  e jn
H H H
 [n]   [n]
Figure 3.4 (p. 198)
Illustration of the eigenfunction property of linear systems. The action of
the system on an eigenfunction input is multiplication by the corresponding
eigenvalue. (a) General eigenfunction (t) or [n] and eigenvalue .
(b) Complex sinusoidal eigenfunction e jt and eigenvalue H(j).
(c) Complex sinusoidal eigenfunction e jn and eigenvalue H(ej).
Eigenfunction: (t )  e j t
Eigenvalue:   H ( j )
2. If ek is an eigenvector of a matrix A with eigenvalue k, then
Ae k  k e k
Arbitrary input = weighted superpositions of eigenfunctions
Convolution operation  Multiplication
Ex. Input: Output: M
y t    a k H  j e
M
xt    a k e jk t
j t k
k 1 k 1
CHAPTER
[B] Discrete-time case:

1. Eigenrepresentation: Fig. 3.4 (c) H ( [n])   [n]
Eigenfunction:  [n]  e jn
Eigenvalue:   H ( e j )
2. Input: Output:
y[n]   ak H  e j  e jk n
M M
x[n]   ak e j k n
k 1 k 1
3.3 Fourier Representations for Four classes of Signals

 Four distinct Fourier representations: Table 3.1.

CHAPTER
3.3.1 Periodic Signals: Fourier Series Representations
1. x[n] = discrete-time signal with fundamental period N. DTFS of x[n] is

x[n]  A[k ]e jk0n (3.4)
k
o = 2/N  Fundamental
frequency of x[n]
2. x(t) = continuous-time signal with fundamental period T. FS of x(t) is
x(t )   A[k ]e jk0t (3.5)
o = 2/T  Fundamental
k frequency of x(t)
 “^” denotes approximate value. A[k] = the weight applied to the kth harmonic.
 e jk0t is the k th harmonic of e j0t .
3. The complex sinusoids exp(jkon) are N-periodics in the frequency index k.
<pf.> e j  N  K 0n  e jN0n e jk0n  e j 2n e jN0n  e jk0n
There are only N distinct complex sinusoids of the form exp(jk0n)

should be used in Eq. (3.4).
N 1
Eq. (3.4) x(t )   A[k ]e jk 0n (3.6)
k 0
 Symmetries in x[k]: k =  (N  1)/2 to (N  1)/2

CHAPTER
4. Continuous-time complex sinusoid exp(jk0t) with distinct frequencies k0
are always distinct.
FS of continuous-time periodic signal x(t) becomes

x(t )  
k 
A[k ]e jk0t (3.7)
 Mean-square error (MSE) between the signal and its series representation:
Discrete-time case:
1 N 1
MSE   x[n]  x[n] dt
2
(3.8)
N n 0
Continuous-time case:
1 T 2
MSE   x(t )  x(t ) dt (3.9)
T 0
3.3.2 Nonperiodic Signals: Fourier-Transform Representations
1. FT of continuous-time signal:
 1 
X(j)/(2) = the weight applied to a sinusoid
x t   X  j e d 
j t
of frequency  in the FT representation.
2 

CHAPTER
2. DTFT of discrete-time signal:


X(e j )/(2) = the weight applied to the
 


1
x n  X e j e jn d  sinusoid ej n in the DTFT representation.
2 
3.4 Discrete-Time Periodic Signals: The Discrete-Time

Fourier Series
1. DTFS pair of periodic signal x[n]:
N 1 Fundamental period = N;
x[n]   X [k ]e jk 0 n
(3.10) Fundamental frequency = o = 2/N
k 0
N 1 Fourier coefficients; Frequency domain

1
X [k ] 
N
 x[n]e
n 0
 jk 0n
(3.11) representation
 Notation: x  n 
DTFS ; 0
 X k 
Example 3.2 Determining DTFS Coefficients
Find the frequency domain representation of the signal depicted in Fig. 3.5.
<Sol.>
1. Period: N = 5 o = 2/5
CHAPTER
Figure 3.5 (p. 203)

Time-domain signal for Example 3.2.
2. Odd symmetry n =  2 to n = 2
3. Fourier coefficient:
2

1
X k   x  n  e  jk 2 n / 5
5 n 2

 x 2e jk 4 / 5  x 1e jk 2 / 5  x0e j 0  x1e  jk 2 / 5  x2e  jk 4 / 5
1
5

1 1 1
X [k ]  {1  e jk 2 / 5  e  jk 2 / 5 }
5 2 2
(3.12)
1
 {1  j sin(k 2 / 5)}
5
CHAPTER
 One period of the DTFS coefficients X[k], k =  2 to k = 2:
sin 4 / 5
X  2 
1
j  0.232e  j 0.531
5 5
sin 2 / 5
X  1   j
1
 0.276e  j 0.760
5 5
X 0   0.2e j 0
1
Fig. 3.6.
5
sin 2 / 5
X 1   j
1
 0.276e j 0.760
5 5
sin 4 / 5
X 2   j
1
 0.232e j 0.531
5 5 Eq. (3.11)
3. Calculate X[k] using n = 0 to n = 4:
X k  
1
5

x0e j 0  x1e  jk 2 / 5  x2e  jk 4 / 5  x3e  jk 6 / 5  x4e  jk 8 / 5 
1 1 1 
 1  e  jk 2 / 5  e  jk 8 / 5  e jk 8 / 5  e jk 2 e jk 2 / 5  e jk 2 / 5
5 2 2 
CHAPTER
X [k ]  Magnitude spectrum of x[n]
X [k ] Even function
arg X [k ]
arg X [k ]  Phase spectrum of x[n]
Odd function
Figure 3.6 (p. 204)

Magnitude and phase of the DTFS coefficients for the signal in Fig. 3.5.
CHAPTER
Example 3.3 Computation of DTFS by Inspection
Determine the DTFS coefficients of x[n] = cos (n/3 + ), using the method of
inspection.
<Sol.>
1. Period: N = 6 o = 2/6 = /3
2. Using Euler’s formula, x[n] can be expressed as
 
j ( n  )  j ( n  )  
e 3
e 3
1  j  j 3 n 1 j j 3 n (3.13)
x[n]   e e  e e
2 2 2
3. Compare Eq. (3.13) with the DTFS of Eq. (3.10) with o = /3, written by
summing from k =  2 to k = 3:
3
x n   X [
k 2
k ]e jk n / 3
 X [2]e  j 2 n / 3  X [1]e  j n / 3  X [0]  X [1]e j n / 3  X [2]e j 2 n / 3  X [3]e j n

(3.14)
Equating terms in Eq.(3.13) with those in Eq. (3.14) having
equal frequencies, k/3, gives

CHAPTER
 j

 e / 2, k  1

x n  X  k    e j / 2,
DTFS ;
 3
k 1
 0, otherwise on  2  k  3

4. Magnitude spectrum and phase spectrum of X[k]: Fig. 3.8.
Figure 3.8 (p. 206)

Magnitude and phase
of DTFS coefficients
for Example 3.3.

CHAPTER
Example 3.4 DTFS Representation of An Impulse Train
Find the DTFS coefficients of the N-periodic impulse train

x  n     n  lN 
l 
as shown in Fig. 3.9.

<Sol.>
It is convenient to evaluate Eq.(3.11) over the interval n = 0 to n = N  1 to obtain
1 N 1 1
X  k      n  e jkn 2 / N 
N n 0 N
Figure 3.9 (p. 207)

A discrete-time impulse train with period N.
CHAPTER
 The Inverse DTFS
The similarity between Eqs. (3.10) and (3.11) indicates that the same
mathematical methods can be used to find the time-domain signal
corresponding to a set of DTFS coefficients.
Example 3.5 The Inverse DTFS
Use Eq. (3.10) to determine the time-domain signal x[n] from the DTFS
coefficients depicted in Fig. 3.10.
<Sol.>
1. Period of DTFS coefficients = 9 o = 2/9
2. It is convenient to evaluate Eq.(3.11) over the interval k =  4 to k = 4 to obtain
4
x n   X k e
k 4
jk 2 n / 9
 e j 2 / 3e j 6 n / 9  2e j / 3e j 4 n / 9  1  2e j / 3e j 4 n / 9  e  j 2 / 3e j 6 n / 9

 2cos  6 n / 9  2 / 3  4cos  4 n / 9   / 3  1
Example 3.6 DTFS Representation of A Square Wave
Find the DTFS coefficients for the N-periodic square wave given by
CHAPTER
/3
Figure 3.10 (p. 208)  /3

Magnitude and phase of
DTFS coefficients for
Example 3.5. 
CHAPTER
1, M n M
xn  
0, M  n  N  M
That is, each period contains 2M + 1 consecutive ones and the remaining
N  (2M +1) values are zero, as depicted in Fig. 3.11. Note that this definition
requires that N > 2M +1.
<Sol.> Figure 3.11 (p. 209)
Square wave for Example 3.6.
1. Period of DTFS coefficients = N o = 2/N

2. It is convenient to evaluate Eq.(3.11) over the interval n =  M to n = N  M  1.
N  M 1
1
X [k ] 
N

n  M
x[n ]e  jk0n
M
1

N
e
n  M
 jk 0n

CHAPTER
2M
1
X [k ] 
N
e
m0
 jk 0 ( m  M )
Change of variable on the
(3.15) index of summation: m = n + M
1  jk 0 M 2 M  jk 0m
 e
N

m0
e
3. For k = 0,  N,  2N, …, we have

e jko  e jko  1
1 2M 2M  1
(3.15) X  k   1 , k  0,  N ,  2 N ,
N m0 N
3. For k  0,  N,  2N, …, we may sum the geometric series in Eq. (3.15) to
obtain
e jk0 M  1  e jk0 M (2 M 1) 
X [k ]    jk 0  , k  0,  N ,  2 N , ...... (3.16)
N  1 e 
1  e 0
jk  2 M 1 / 2
  1  e jk0  2 M 1  1  e jk0  2 M 1 / 2  e jk0  2 M 1 / 2 
X k    jk 0 / 2  jk 0   jk 0 / 2  jk 0 / 2 
N e  1  e  N  e  e 
k  0,  N ,  2 N , ......
CHAPTER
1 sin k 0 2M  1 / 2

X k   , k  0,  N ,  2 N ,
N sin k 0 / 2
The numerator and
denominator of above Eq. are
4. Substituting o = 2/N, yields
divided by 2j.
 1 sin  k  2 M  1 / N 
 , k  0,  N ,  2 N ,
X k    N sin  k / N 

  2 M  1 / N , k  0,  N ,  2 N ,
L’Hôpital’s
1 sin  k  2 M  1 / N  2 M  1 Rule
lim 
k  0,  N ,  2 N , N sin  k / N  N
For this reason, the expression for X[k] is commonly written as
1 sin k 2M  1 / N 
X k   The value of X[k] for k = 0,  N,  2N, …,
N sin k / N  is obtained from the limit as k  0.
Fig. 3.12.

CHAPTER
Figure 3.12 (p. 211)

The DTFS coefficients for the square
wave shown in Fig. 3.11, assuming a
period N = 50: (a) M = 4. (b) M = 12.

CHAPTER
 Symmetry property of DTFS coefficient: X[k] = X[ k].

1. DTFS of Eq. (3.10) can be written as a series involving harmonically related
cosines.
2. Assume that N is even and let k range from  ( N/2 ) + 1 to N/2. Eq. (3.10)
becomes
N /2
xn   X k e jk 0 n
k   N / 2 1
 X me 
N / 2 1
xn  X 0  X N / 2e jk0 n / 2
 jk 0 n
 X  me  jk0n
m 1
3. Use X[m] = X[ m] and the identity No = 2 to obtain

N / 2 1
 e jm0n  e jm0n  e jn = cos(n)
x  n   X 0  X  N / 2 e j n
  2 X  m   since sin(n) =
m1  2 
N / 2 1
0 for integer n.
 X 0  X  N / 2 cos  n    2 X mcos  m n 
m1
0
4. Define the new set of coefficients

CHAPTER
 X  k , k  0, N / 2
B k   
2 X  k , k  1, 2, , N / 2 1
and write the DTFS for the square wave in terms of a series of harmonically
related cosines as
N /2
x[n]   B[k ]cos(k 0 n) (3.17)
k 0
 A similar expression may be derived for N odd.

Example 3.7 Building a Square Wave From DTFS Coefficients
The contribution of each term to the square wave may be illustrated by defining
the partial-sum approximation to x[n] in Eq. (3.17) as
J
x J [n]   B[k ]cos(k 0 n) (3.18)
k 0
where J  N/2. This approximation contains the first 2J + 1 terms centered on k

= 0 in Eq. (3.10). Assume a square wave has period N = 50 and M = 12. Evaluate
one period of the Jth term in Eq. (3.18) and the 2J + 1 term approximation x J [n]
for J = 1, 3, 5, 23, and 25.
CHAPTER
Figure 3.14a (p. 213)

Individual terms in the DTFS expansion of a square wave (left panel)
and the corresponding partial-sum approximations J[n] (right panel).
The J = 0 term is 0[n] = ½ and is not shown. (a) J = 1. (b) J = 3. (c) J =
5. (d) J = 23. (e) J = 25.
CHAPTER
Signals and Systems_Simon

Figure 3.14b (p. 213) 26
CHAPTER
<Sol.>
1. Fig. 3.14 depicts the J th term in the sum, B[J] cos(Jon), and one period of
x J [n] for the specified values of J.
2. Only odd values for J are considered, because the even-indexed coefficients
B[k] are zero when N = 25 and M = 12.
3. The approximation improves as J increases, with x[n] represented exactly
when J = N/2 = 25.
4. The coefficients B[k] associated with values of k near zero represent the low-
frequency or slowly varying features in the signal, while the coefficients
associated with values of k near  N/2 represent the high frequency or
rapidly varying features in the signal.
Example 3.8 Numerical Analysis of the ECG
Evaluate the DTFS representations of the two electrocardiogram (ECG)
waveforms depicted in Figs. 3.15 (a) and (b). Fig. 3.15 (a) depicts a normal ECG,
while Fig. 3.15 (b) depicts a heart experiencing ventricular tachycardia. The
discrete-time signals are drawn as continuous functions, due to the difficulty of
depicting all 2000 values in each case. Ventricular tachycardia is a series
cardiac rhythm disturbance (i.e., an arrhythmia) that can result in death.
CHAPTER
Figure 3.15 (p. 214)

Electrocardiograms
for two different
heartbeats and the
first 60 coefficients of
their magnitude
spectra.
(a) Normal heartbeat.
(b) Ventricular
tachycardia.
(c) Magnitude
spectrum for the
normal heartbeat.
(d) Magnitude
spectrum for
ventricular
tachycardia.

CHAPTER
Figure 3.15
(p. 214)
Electrocardiograms
for two different
heartbeats and the
first 60 coefficients
of their magnitude
spectra.
(c) Magnitude
spectrum for the
normal heartbeat.
(d) Magnitude
spectrum for
ventricular
tachycardia.

CHAPTER
It is characterized by a rapid, regular heart rate of approximately 150 beats per
minute. Ventricular complexes are wide (about 160 ms in duration) compared
with normal complexes (less than 110 ms) and have an abnormal shape. Both
signals appear nearly periodic, with only slight variations in the amplitude and
length of each period. The DTFS of one period of each ECG may be computed
numerically. The period of the normal ECG is N = 305, while the period of the
ECG showing ventricular tachycardia is N = 421. One period of each waveform
is available. Evaluate the DTFS coefficients of each, and plot their magnitude
spectrum.
<Sol.>
1. The magnitude spectrum of the first 60 DTFS coefficients is depicted in Fig.
3.15 (c) and (d).
2. The normal ECG is dominated by a sharp spike or impulsive feature.
3. The DTFS coefficients of the normal ECG are approximately constant,
exhibiting a gradual decrease in amplitude as the frequency increase.
Fairly small magnitude!
4. The ventricular tachycardia ECG contains smoother features in addition to
sharp spikes, and thus the DTFS coefficients have a greater dynamic range,
with the low-frequency coefficients containing a large proportion of the total
power.
CHAPTER
3.5 Continuous-Time Periodic Signals: The Fourier Series

 Fourier series pair  Exponential FS
1. Continuous-time periodic signals are represented by the Fourier series (FS).
2. A signal with fundamental period T and fundamental frequency o = 2/T,
the FS pair of x(t) is

x(t )   X [k ]e
k 
jk0t
(3.19)
Frequency domain
1 T

X [k ]  x (t ) e  jk0t
dt representation of x(t)
(3.20)
T 0
3. Notation: The variable k determines the

frequency of the complex sinusoid
x  t  
FS ;0
 X k  associated with X[k] in Eq. (3.19).
 Mean-square error (MSE)
1. Suppose we define

xˆ  t    X k  e
k 
jko t
and choose the coefficients X[k] according to Eq.(3.20.)

CHAPTER
2. If x(t) is square integrable,

1 T
 x t  2
dt  
T 0
then the MSE between x(t) and x̂  t  is zero, or, mathematically,

 2
xt   xt  dt  0
1 T
T 0
x(t )  xˆ  t  at all values of t; it simply implies that there is zero power

in their difference.
3. Pointwise convergence of x(t )  xˆ  t  is guaranteed at all values of t
except those corresponding to discontinuities if the Dirichlet conditions are
satisfied:  x(t) is bounded.
 x(t) has a finite number of maximum and minima in one period.
 x(t) has a finite number of discontinuities in one period.
 If a signal x(t) satisfies the Dirichlet conditions and is not continuous, then
x̂  t  converges to the midpoint of the left and right limits of x(t) at each
discontinuity.
CHAPTER
Example 3.9 Direct Calculation of FS Coefficients
Determine the FS coefficients for the signal x(t) depicted in Fig. 3.16.
<Sol.>
1. The period of x(t) is T = 2,
so o = 2/2 = .
2. One period of x(t): x(t) = e  2t,
0  t  2.
3. FS of x(t):
1 2 2t  jk t 1 2  2 jk t Figure 3.16 (p. 216)
X k    e e dt   e dt Time-domain signal for Example 3.9.
2 0 2 0
2
4
1 
X k  
2  2  jk 
e
  2  jk t

1
4  jk 2
1  e4e jk 2  
1 e
4  jk 2
0
Fig. 3.17.
e  jk2 = 1
 The Magnitude of X[k]  the magnitude spectrum of x(t);
the phase of X[k]  the phase spectrum of x(t).

CHAPTER
Figure 3.17
(p. 217)
Magnitude
and phase
spectra for
Example 3.9.

CHAPTER
Example 3.10 FS Coefficients for An Impulse Train
Determine the FS coefficients for the signal x(t) defined by

x t      t  4l 
l 
<Sol.>
1. Fundamental period of x(t) is T = 4, each period contains an impulse.
2. By integrating over a period that is symmetric about the origin  2 < t  2, to
obtain X[k]:
1 2 1
X k     jk  / 2 t
4 2
 t e dt 
4
3. The magnitude spectrum is constant and the phase spectrum is zero.
 Inspection method for finding X[k]: Whenever x(t) is expressed in terms of
sinusoid, it is easier to obtain X[k] by inspection. The method of inspection is
based on expanding all real sinusoids in terms of complex sinusoids and
comparing each term in the resulting expansion to the corresponding terms
of Eq. (3.19).

CHAPTER
Example 3.11 Calculation of FS Coefficients By Inspection
Determine the FS representation of the signal
xt   3 cost / 2   / 4
<Sol.>
1. Fundamental period of x(t) is T = 4, and o = 2/4 = /2. Eq. (3.19) is written
as 
x(t )  
X [k ]e jk t / 2
k 
(3.21)
j  t / 2  / 4   j  t / 2  / 4 
e e 3 3
x t   3  e j / 4e j t / 2  e j / 4e j t / 2
2 2 2
2. Equating each term in this expression to the terms in Eq. (3.21) gives the FS
coefficients: 3  j / 4
2 e , k  1
 Magnitude and
3
X [k ]   e j / 4 , k  1 (3.22) phase spectrum:
2 Fig. 3.18.
0, otherwise

CHAPTER
/4
 /4
Figure 3.18 (p. 219)
Magnitude and phase spectra for Example 3.11.
 Time-domain representation obtained from FS coefficients
Example 3.12 Inverse FS
Find the time-domain signal x(t) corresponding to the FS coefficients
X k   1 / 2 e jk / 20
k
Assume that the fundamental period is T = 2.
<Sol.>
1. Fundamental frequency: o = 2/T = . From Eq. (3.19), we obtain
 
x  t    1/ 2  e   1/ 2  e jk / 20e jk t
k jk / 20 jk t k
e
k 0 k 1

CHAPTER
 
  1 / 2 e jk / 20 jkt
 1 / 2 e  jl / 20e  jlt
k l
e
k 0 l 1
xt  
1 1
 1
1  1 / 2e j t  / 20
1  1 / 2e  j t  / 20
2. Putting the fraction over a common denominator results in
xt  
3
5  4 cost   / 20
Example 3.13 FS for A Square Wave
Determine the FS representation of the square wave depicted in Fig. 3.21.
Figure 3.21 (p. 221)
<Sol.>
CHAPTER
1. Fundamental frequency: o = 2/T.

2. X[k]: by integrating over  T/2  t  T/2
1 T /2 1 T0  jk0t
X k    x t  e  jk0 t
dt   e dt
T  T / 2 
T 0 T
T0
1  jk0t
 e , k 0
Tjk0  T0
2  e jk0T0  e jk0T0 
  , k  0
Tk0  2j 
2sin  k 0T0 
 , k 0
Tk0
For k = 0, we have
X 0   dt  0
1 T0 2T
T T0 T
3. By means of L’Hôpital’s rule, we have
CHAPTER
2 sin k 0T0  2T0

lim 
k 0 Tk 0 T
2 sin k 0T0 
X k  
Tk 0
4. X[k] is real valued. Using o = 2/T gives X[k] as a function of the ratio To/T:
2sin(k 2 T0 / T )
X [k ]  (3.23)
k 2
5. Fig. 3.22 (a)-(c) depict X[k],  50  k  50, for To/T =1/4, To/T =1/16, and To/T =
1/64, respectively.
 Definition of Sinc function:
sin( u )
sinc(u )  (3.24) Fig. 3.23.
u
1) Maximum of sinc function is unity at u = 0, the zero crossing occur at
integer values of u, and the amplitude dies off as 1/u.
2) The portion of this function between the zero crossings at u =  1 is
manilobe of the sinc function.
CHAPTER
Figure 3.22a&b
(p. 222)
The FS
coefficients,
X[k], –50  k  50,
for three square
waves. (see Fig.
3.21.) (a) To/T =
1/4 . (b) To/T =
1/16. (c) To/T =
1/64.

CHAPTER
Figure 3.22c
(p. 222)

CHAPTER
Figure 3.23 (p. 223)

Sinc function sinc(u) = sin(u)/(u)
CHAPTER
3) The smaller ripples outside the mainlobe are termed sidelobe.
4) The FS coefficients in Eq. (3.23) are expressed as
2T0  2T 
X k   sinc  k 0 
T  T 
 Fourier series pair  Trigonometric FS
1. Trigonometric FS of real-valued signal x(t):

x(t )  B[0]   B[k ]cos(k0t )  A[k ]sin(k0t ) (3.25)
k 1
FS coefficients:
B[0] = X[0] represents the time-
1 T
B[0] 
T 0
x(t )dt averaged value of the signal.
2 T
B[k ]   x(t ) cos(k0t )dt (3.26)
T 0
2 T
A[k ]   x(t )sin(k0t )dt
T 0

CHAPTER
2. Relation between exponential FS and trigonometric FS coefficients:

B[k ]  X [k ]  X [k ] Euler’s Formula
for k  0 (3.27)
A[k ]  j ( X [k ]  X [k ])
Ex. Find trigonometric FS coefficients of the square wave studied in Example
3.13.
<Sol.>
1. Substituting Eq. (3.20) into Eq. (3.27), gives
B[0]  2T0 / T
2sin(k 2 T0 / T )
B[k ]  , k 0 (3.28)
k
A[k ]  0 Because x(t) is an even function
2. Trigonometric FS expression of x(t):

x(t )   B[k ]cos(k0t ) (3.29)
k 0

CHAPTER
Example 3.14 Square Wave Partial-Sum Approximation
Let the partial-sum approximation to the FS in Eq. (3.29), be given by
 J
x J t    Bk cosk 0 t 
k 0
This approximation involves the exponential FS coefficients with indices  J 

k  J. Consider a square wave with T = 1 and To/T = ¼. Depict one period of the
Jth term in this sum, and find

x J  t  for J  1, 3, 7, 29, and 99.
<Sol.>
1. The individual terms and partial-sum approximation are depicted in Fig. 3.25.
2. Each partial-sum approximation passes through the average value (1/2) of
the discontinuity, the approximation exhibits ripple.
3. This ripple near discontinuities in partial-sum FS approximations is termed
the Gibbs phenomenon.
4. As J increase, the ripple in the partial-sum approximations becomes more
and more concentrated near the discontinuities.

CHAPTER
Figure 3.25a (p. 225)

Individual terms (left panel) in the FS expansion of a square wave and the
corresponding partial-sum approximations J(t) (right panel). The square
wave has period T = 1 and Ts/T = ¼. The J = 0 term is 0(t) = ½ and is not
shown. (a) J = 1.

CHAPTER
Figure 3.25b-3 (p. 226)

(b) J = 3. (c) J = 7. (d) J = 29.
CHAPTER
Figure 3.25e (p. 226)

(e) J = 99.
Example 3.15 RC Circuit: Calculating The Output By Means of FS
Let us find the FS representation for the output y(t) of the RC circuit depicted in
Fig. 3.2 in response to the square-wave input depicted in Fig. 3.21, assuming
that To/T = ¼, T = 1 s, and RC = 0.1 s.
Figure 3.21 (p. 221)
Square wave for
Example 3.13.
Figure 3.2 (p. 197)

RC circuit for Example 3.1.
CHAPTER
<Sol.>
1. If the input to an LTI system is expressed as a weighted sum of sinusoid,
then the output is also a weighted sum of sinusoids.
2. Input: 4. Frequency response of the system H(j):
y  t    Y k   H  jk0  X k 
 FS ;0
xt    X k e
k  
jk0t
5. Frequency response of the RC circuit:

3. Output:
H  j  
1 / RC

j  1 / RC
yt    H  jk X k e
k  
0
jk0t
6. Substituting for H(jko) with RC = 0.1 s and o = 2, and To/T = ¼, gives
sin k / 2
Y k  
10
j 2k  10 k
7. We determine y(t) using the approximation
100 The magnitude and phase spectra for the
y(t )  
k 100
Y [k ]e jk0t
(3.30) range  25  k  25: Fig. 3.26 (a) and (b).
Waveform of y(t): Fig. 3.26(c).
CHAPTER
Figure 3.26 (p. 228)

The FS coefficients
Y[k], –25  k  25,
for the RC circuit
output in response to
a square-wave input.
(a) Magnitude
spectrum. (b) Phase
spectrum. c) One
period of the input
signal x(t) dashed
line) and output
signal y(t) (solid line).
The output signal y(t)
is computed from the
partial-sum
approximation given
in Eq. 3.30).

CHAPTER
Figure 3.26 (p. 228)

The FS coefficients
Y[k], –25  k  25,
for the RC circuit
output in response to
a square-wave input.
(a) Magnitude
spectrum. (b) Phase
spectrum. c) One
period of the input
signal x(t) dashed
line) and output
signal y(t) (solid line).
The output signal y(t)
is computed from the
partial-sum
approximation given
in Eq. (3.30).

CHAPTER
8. The circuit attenuates the amplitude of X[k] when k 1.

The degree of attenuation increases as the frequency ko increases.
The circuit also introduces a frequency-dependent phase dependent shift.
Example 3.16 DC-to-AC Conversion
A simple scheme for converting direct current (dc) to alternating current (ac)
is based on applying a periodic switch to a dc power source and filtering out
or removing the higher order harmonics in the switched signal. The switch in
Fig. 3.27 changes position every 1/20 second. We consider two cases: (a) The
switch in either open or closed; (b) the switch reverses polarity. Fig. 3. 28 (a)
and (b) depict the output waveforms for these two cases. Define the
conversion efficiency as the ratio of the power in the 60-Hz component of the
output waveform x(t) to the available dc power at the input. Evaluate the
conversion efficiency for each case.
<Sol.>
1. The FS for the square wave x(t) with T = 1/60 second and o = 2/T = 120
rad/s: Fig. 3.28(a), where
2 A sin k / 2
Bk   Ak   0
A
B 0  ,k  0
2 k
CHAPTER
Figure 3.27 (p. 229)
Switching power supply for DC-to-AC conversion.
Figure 3.28 (p. 229)

Switching power supply
output waveforms with
fundamental frequency
0 = 2/T = 120. (a) On-
off switch. (b) Inverting
switch.

CHAPTER
2. The harmonic at 60 Hz in the trigonometric FS representation of x(t) has

amplitude given by B[1] and contains power B[1]2/2.
The dc input has power A2.
3. Conversion efficiency for Fig. 3.28 (a):
 B1 / 2
2
Ceff   2 /  2  0.20
A2
4. The FS for the signal x(t) shown in Fig. 3.28(b):
B0  0
4 A sin k / 2
Bk   ,k  0
k
Ak   0
5. Conversion efficiency for Fig. 3.28 (b):
 B1 / 2
2
Ceff   8 /  2  0.81
A2

CHAPTER
3.6 Discrete-Time Nonperiodic Signals: The Discrete-Time

Fourier Transform (DTFT)
1. The DTFT is used to represent a discrete-time nonperiodic signal as a
superposition of complex sinusoids.
2. DTFT representation of time-domain signal x[n]:
1  DTFT
x[n] 
2 

X (e j )e jn d  (3.31)
pair

X (e )  j
 x[n]e
n 
 jn
(3.32) Frequency-domain representation x[n]
x  n   X  e j 
3. Notation: DTFT
4. Condition for convergence of DTFT:

If x[n] is of infinite duration, then the sum converges only for certain classes
of signals. If x[n] is absolutely summable, i.e.,

The sum in Eq. (3.32) converges uniformly to a

n 
x n  
continuous function of .
CHAPTER
 If x[n] is not absolutely summable, but does satisfy (i.e., if x[n] has finite
energy),
 It can be shown that the sum in Eq. (3.32)
 xn  
2
converges in a mean-square error sense,
n   but does not converge pointwise.
Example 3.17 DTFT of An Exponential Sequence
Find the DTFT of the sequence x[n] = nu[n].
<Sol.>
1. DTFT of x[n]: This sum diverges for   1
    une
 
X e j n  jn
  n e  jn
n   n 0
2. For  < 1, we have


X (e )    e j  
n 1
j
,  1 (3.33)
n 0 1 e  j
3. If  is real valued, Eq. (3.33) becomes
X e j  
1 Euler’s Formula
1   cos   j sin 
CHAPTER
4. Magnitude and phase spectra:

Even
 j 1

1
1   cos  
X e
 
function
 1  2 cos 
1/ 2 1/ 2
  sin
2 2 2 2
  sin  
  
arg X e j   arctan  Odd
 1   cos   function
Fig. 3.29 for  = 0.5 and  = 0.9.
Example 3.18 DTFT of A Rectangular Pulse
Let
1, n  M
x n  
0, n  M
as depicted in Fig. 3.30 (a). Find the DTFT of x[n].
<Sol.>
1. DTFT of x[n]:
X e    1e
M
j  jn
.
n  M

CHAPTER
Figure 3.29
(p.232)
The DTFT of an
exponential signal
x[n] = ()nu[n]. (a)
Magnitude
spectrum for  =
0.5. (b) Phase
spectrum for  =
0.5. (c) Magnitude
spectrum for
 = 0.9. (d) Phase
spectrum for  =
0.9.

CHAPTER
Figure 3.30
(p. 233)
Example 3.18.
(a) Rectangular
pulse in the time
domain. (b)
DTFT in the
frequency
domain.
p

CHAPTER
X e   e
2M 2M
Change of variable
j
m0
 j ( m  M )
e j M
 e jm
m0
m=n+M
 jM 1  e j 2( M 1)
e ,   0,  2 ,  4 ,
 1 e  j
 2 M  1, =0,  2 ,  4 ,

2. The expression for X(e j ) when   0,  2 ,  4 ,
j
X (e )  e j M

e j 2 M 1 / 2 e j 2 M 1 / 2  e j 2 M 1 / 2 
e  j / 2  e j / 2  e  j / 2 

 e
j  2 M 1 / 2
e
 j  2 M 1 / 2

e j / 2  e  j / 2
sin    2 M  1 / 2  L’Hôptital’s Rule
lim  2 M  1;
0,2 ,4 , , sin   2 
sin    2M  1 / 2  With understanding that X(e j ) for
j
X (e )    0,  2 ,  4 , is obtained as
sin   2  limit.
CHAPTER
3. Graph of X(e j ): Fig. 3.30(b).

Example 3.19 Inverse DTFT of A Rectangular Pulse
Find the inverse DTFT of
Figure 3.31 (p. 234)
  
 
1, W
X e j   , Example 3.19. (a) Rectangular
0, W  n   pulse in the frequency domain. (b)
which is depicted in Fig. 3.31 (a). Inverse DTFT in the time domain.
<Sol.>
1. Note that X(e j ) is specified only for   <   .

p
2. Inverse DTFT x[n]: 1
W
x  n   e jn d 
2 W

CHAPTER
1 jn W
 e W , n  0
2 nj
1
 sin Wn  , n  0.
n
3. For n = 0, the integrand is unity and we have x[0] = W/. Using L’Hôpital’s
rule, we easily show that
1 W
lim sin Wn   ,
n 0 n 
and thus we usually write
1
x  n  sin Wn 
n
as the inverse DTFT of X(e j ), with the understanding that the value at n =
0 is obtained as limit.
4. We may also write
W
x  n  sin c Wn /   , Fig. 3.31 (b).

CHAPTER
Example 3.20 DTFT of The Unit Impulse
Find the DTFT of x[n] = [n].
<Sol.>
1. DTFT of x[n]:

X e j
    n e  jn
1 Figure 3.32 (p. 235)
n 
Example 3.20. (a) Unit impulse
in the time domain. (b) DTFT of
 n 
DTFT
 1. unit impulse in the frequency
2. This DTFT pair is depicted in Fig. 3. 32. domain.
Example 3.21 Inverse DTFT of A Unit Impulse Spectrum

p
Find the inverse DTFT of X(e j ) = (),   <   .
<Sol.>
1. Inverse DTFT of X(e j ):
CHAPTER
1 
x  n       e jn d . Sifting property of impulse function
2 
1

DTFT
  ,      .
2
 We can define X(e j ) over all  by writing it as an infinite sum of delta
functions shifted by integer multiples of 2.
p
2. This DTFT pair is depicted in Fig. 3. 33.
Figure 3.33 (p. 236)

 Example 3.21. (a) Unit impulse in the
j
X e         k 2 . frequency domain. (b) Inverse DTFT in
k 
the time domain.
 Dilemma: The DTFT of x[n] = 1/(2) does not converge, since it is not a
square summable signal, yet x[n] is a valid inverse DTFT!
CHAPTER
Example 3.22 Moving-Average Systems: Frequency Response
Consider two different moving-average systems described by the input-output
equations
y1  n 
1
2
 x  n  x  n  1 and y2  n 
1
2
 x  n  x  n  1
The first system averages successive inputs, while the second forms the
difference. The impulse responses are
1 1 1 1
h1  n    n    n  1 and h2  n    n    n  1
2 2 2 2
Find the frequency response of each system and plot the magnitude responses.
<Sol.>
1. The frequency response is the DTFT of the impulse response.
2. For system # 1  h1[n] :
Frequency response:
  
1 1  j
e
j j 
H1 (e j )   e j j
e
2 2
j e 2
cos   .
2 2 H1 (e )  e 2
2
2
CHAPTER
Magnitude response: Phase response:

 
H1  e j
  cos   , arg H1 (e j )   .
2 2
3. For system # 2  h2[n] :
Frequency response:
 
 j j 
1 1 j e 2
e 2 j 
H 2 (e )   e j  je
j 2
 je 2
sin   .
2 2 2j 2
Magnitude response: Phase response:
  
H2 e j
  sin   , 

 2  , 0
2 arg H 2 ( e j )   2
4. Fig. 3.36 (a) and (b) depict the   2   ,   0

 2
magnitude responses of the two
systems on the interval   <   .

CHAPTER
p
Figure 3.36 (p. 239)

The magnitude responses of two simple discrete-time systems.
(a) A system that averages successive inputs tends to attenuate high
frequencies. (b) A system that forms the difference of successive inputs
tends to attenuate low frequencies.
Example 3.23 Multipath Communication Channel: Frequency Response
The input-output equation introduced in Section 1.10 describing a discrete-time
model of a two-path propagation channel is
y  n  x  n  ax  n  1.
In Example 2.12, we identified the impulse response of this system as h[n] = [n]
+ a[n  1] and determined that the impulse response of the inverse system was
hinv[n] = ( a)nu[n]. The inverse system is stable, provided that a< 1.
H  e j   1  a e 
 j arga
.
CHAPTER
Compare the magnitude responses of both systems for a = 0.5 e j/3 and a =
0.9 e j2/3.
<Sol.>
1. The frequency response is the DTFT of the impulse response.
2. Frequency response of the system modeling two-path propagation:
H  e j   1  ae j . Using a  a e j arga
H  e j   1  a e 
 j arga
. Apply Euler’s formula
H  e j   1  a cos    arg a  j a sin    arg a .

3. Magnitude response:
cos2   sin 2   1
H e   1  a cos    arga 
12
 a sin    arg a
2
j 2 2

 1  a  2 a cos    arg a 
2 12
,
4. The frequency response of the inverse system may be obtained by replacing

 with  a in Eq. (3.33).
CHAPTER
H inv  e j  
1 Expressing a in polar form
 j
, a 1
1  ae
 e    1  a cos    arg a  

1 2
 a sin    arg a
2
inv j 2 2
H
 
1 2
 1  a  2 a cos    arg a
2
,
 The frequency response of the inverse system is the inverse of the frequency
response of the original system.
5. Magnitude response of the inverse system:
H inv  e j  
1
.
1  a  2 a cos    arg a 
2 12
6. Fig. 3.37 depicts the magnitude response of H(e j ) for both a = 0.5 e j/3 and
a = 0.9 e j2/3 on the interval   <   .
7. The magnitude response approaches a maximum of 1 + a when  = arg{a}
and a minimum of 1  a when  = arg{a}  .
8. The magnitude response of the corresponding inverse system: Fig. 3.38.
CHAPTER
Figure 3.37 (p. 241)

Magnitude response of the system in Example 3.23 describing multipath
propagation. (a) Echo coefficient a = 0.5ej/3. (b) Echo coefficient a = 0.9ej2/3.

CHAPTER
Figure 3.38 (p. 241)

Magnitude response of the inverse system for multipath propagation in
Example 3.23. (a) Echo coefficient a = 0.5ej/3. (b) Echo coefficient a = 0.9ej/3
 If a = 1, then the multipath model applies zero gain to any sinusoid with
frequency  = arg{a}  .
 The multipath system cannot be inverted when a= 1.

CHAPTER
3.7 Continuous-Time Nonperiodic Signals: The Fourier

Transform (FT)
1. The Fourier transform (FT) is used to represent a continuous-time non-
periodic signal as a superposition of complex sinusoids.
2. FT representation of time-domain signal x(t):
1 
x(t ) 
2 

X ( j )e jt d (3.35)
 Frequency-domain representation of
X ( j )   x(t )e  j t
dt (3.36) the signal x(t)

3. Notation for FT pair:

x  t  
FT
X  j  .
4. Convergence condition for FT:
1) Approximation:
1 
xˆ  t    X  j  e jt d ,
2 
2) Squared error between x(t) and x̂  t  : the error energy, given by

CHAPTER

 x  t   xˆ  t  dt ,
2

is zero if x(t) is square integrable, i.e., if


 x  t  dt  .
2

 Zero squared error does not imply pointwise convergence [i.e., x  t   xˆ  t  ]

at all values of t.
There is zero energy in the difference of x(t ) and xˆ  t 
3. Pointwise convergence of x(t )  xˆ  t  is guaranteed at all values of t
except those corresponding to discontinuities if x(t) satisfies the Dirichlet
conditions for nonperiodic signals:
 x(t) is absolutely integrable:

 x  t  dt  .

 x(t) has a finite number of maximum, minima, and discontinuities in any

finite interval.
 The size of each discontinuity is finite.

CHAPTER
Example 3.24 FT of A Real Decaying Exponential
Find the FT of x(t) = e  a t u(t), shown in Fig. 3.39(a).
<Sol.>
1. For a  0, since x(t) is not absolutely integrable, i.e.,

0
e at dt  , a  0 The FT of x(t) does not converge for a  0.
2. For a > 0, the FT of x(t) is
 
X  j    e at u  t  e j t dt   e
  a  j t
dt
 0
1 1
e 
 a  j t 
 
a  j a  j
0
3. Magnitude and phase of X(j):

1
X  j   1
a 2
 
2 2
Fig. 3.39 (b) and (c).
arg  X  j    arctan  a  ,

CHAPTER
Figure 3.39 (p. 243)

Example 3.24. (a)
Real time-domain
exponential signal.
(b) Magnitude X(j) 
spectrum.
(c) Phase spectrum.
arg{X(j)}
/4
/4

 /4
 /2
CHAPTER
Example 3.25 FT of A Rectangular Pulse
Consider the rectangular pulse depicted in Fig. 3.40 (a) and defined as
1, T0  t  T0
x t   
0, t  T0
Find the FT of x(t).
<Sol.>
1. The rectangular pulse x(t) is absolutely integrable, provided that To < .
2. FT of x(t):

X  j    x  t  e  j t dt   e  j t dt
T0
  T0
With understanding that
the value at  = 0 is
1  j t 2
 e T0
 T0  sin T0  ,   0 obtained by evaluating
j  a limit.
3. For  = 0, the integral simplifies to 2To.
2 2
lim sin T0   2T0 . X  j   sin T0  ,
 0  
4. Magnitude spectrum:
CHAPTER
sin T0 
X  j   2 , Fig. 3. 40 (b).

5. Phase spectrum: 6. X(j) in terms of sinc function:
 0, sin T0    0 X  j   2T0 sinc T0   .
arg  X  j   
 , sin T0    0
As To increases, the nonzero time
extent of x(t) increases, while X(j) v  ; p  
becomes more concentrated about
the frequency origin.
Figure 3.40 (p. 244)

Example 3.25. (a) Rectangular pulse in the
time domain. (b) FT in the frequency domain.
CHAPTER
Example 3.26 Inverse FT of A Rectangular Pulse
Find the inverse FT of the rectangular spectrum depicted in Fig. 3. 42 (a) and
given by 1, W    W
 Figure 3.42 (p. 246)
X  j    Example 3.26. (a) Rectangular spectrum in
0,  W
the frequency domain.
<Sol.>
(b) Inverse FT in the time domain.
v  ; p  
1. Inverse FT of x(t):
1 1 j t 1
x t  
W
 e j t d    sin(Wt ), t  0
W
e W
2 W j t t
2. When t = 0, the integral simplifies to W/, i.e.,
CHAPTER
1
lim sin Wt   W  , With understanding that the value
t 0  t
at t = 0 is obtained as a limit.
3. Inverse FT is usually written as
1 W  Wt 
x t   sin Wt  , or x t   sinc   , Fig. 3. 42 (b).
t   
As W increases, the frequency-domain reprsentation becomes less
concentrated about  = 0, while the time-domain representation X(j)
becomes more concentrated about t = 0.
 Duality between Example 3.25 and 3.26.
Example 3.27 FT of The Unit Impulse
Find the FT of x(t) = (t).
<Sol.>
1. x(t) does not satisfy the Dirichlet conditions, since the discontinuity at the
origin is infinite.
2. FT of x(t): 
X  j      t  e
 j t
dt  1


CHAPTER
 Notation for FT pair:   t   1,

FT
Example 3.28 Inverse FT of An Impulse Spectrum

Find the inverse FT of X(j) = 2().
<Sol.>
1. Inverse FT of X(j):
1 
x t    2   e j t d  1 Duality between
2  Example 3.27 and 3.28
2. Notation of FT pair:
1 
FT
2  
Example 3.29 Characteristic of Digital Communication Signals
In a simple digital communication system, one signal or “symbol” is
transmitted for each “1” in the binary representation of the message, while a
different signal or symbol is transmitted for each “0”. One common scheme,
binary phase-shift keying (BPSK), assumes that the signal representing “0” is
the negative of the signal and the signal representing “1” is the positive of the
signal. Fig. 3.44 depicts two candidate signals for this approach: a rectangular
pulse defined as
CHAPTER
 A, t  T0 / 2
xr  t   
 0, t  T0 / 2
and a raised-cosine pulse defined as Figure 3.44 (p. 249)
Pulse shapes used in
 Ac / 2   2cos  2 t / T0   ,
 t  T0 / 2
xc  t   
BPSK communications.

 0, t  T0 / 2 (a) Rectangular pulse.
(b) Raised cosine pulse.
The transmitted BPSK signals for communicating a sequence of bits using each
pulse shape are illustrated in Fig. 3.45. Note that each pulse is To seconds long,
CHAPTER
so this scheme has a transmission rate of 1/To bits per second. Each user’s
signal is transmitted within an assigned frequency band, as depicted in Chapter
5. In order to prevent interference with users of other frequency bands,
governmental agencies place limits on the energy of a signal that any user
transmits into adjacent frequency bands. Suppose the frequency band
assigned to each user is 20 kHz wide. Then, to prevent interference with
adjacent channels, we assume that the peak value of the magnitude spectrum
of the transmitted signal outside the 20-kHz band is required to be  30 dB
below the peak in-band magnitude spectrum. Choose the constant Ar and Ac so
that both BPSK signals have unit power. Use the FT to determine the maximum
number of bits per second that can be transmitted when the rectangular and
raised-cosine pulse shapes are utilized.
<Sol.>
1. BPSK signals are not periodic, but their magnitude squared is To periodic.
2. Powers in rectangular pulse and raised-cosine pulse:
1 T0 / 2 2
Pr   Ar dt Ar2
T0 T0 / 2
and
CHAPTER
Figure 3.45 (p. 249)

BPSK signals constructed by using (a) rectangular pulse shapes and (b)
raised-cosine pulse shapes.
CHAPTER
1
 A 4 1  2cos  2 t 2  
T0 2

2
Pc  2
c dt
T0  T0 2
Ac2
1  2cos  2 t T0   1 2  1 2cos  4 t T0   dt
T0 2

4T0  T0 2
3 Ac2

8
Hence, unity transmission power is obtained by choosing Ar = 1 and Ac = 8 / 3 .
3. FT of the rectangular pulse xr(t):
sin  T0 2 
X r  j   2

In terms of Hz rather than rad/sec ( = 2f): X r ( j )  X r ( jf )
sin  fT0 
X r  jf   2
f
Normalized Spectrum of 20log10  X r ( jf ) / To  dB: Fig. 3.46.

CHAPTER
Figure 3.46
(p. 250)
Spectrum of
rectangular
pulse in dB,
normalized by
T0.
The normalized
by To removed
the
dependence of
the magnitude
on To.

CHAPTER
4. From Fig. 3.46, we must choose To so that the 10 th zero crossing is at 10 kHz
in order to satisfy the constraint that the peak value of the magnitude
spectrum of the transmitted signal outside the 20-kHz band allotted to this
user be less than  30 dB.
f = k/To 1000 = 10/To To = 10  3
Data transmission rate = 1000 bits/sec.
5. FT of raised-cosine pulse xc(t):
X c  j  
1 8 T0 2
2 3  T0 2
1  2cos  2 t T0   e  j t
dt
Euler’s formula
2 T0 2  j t 1 2 T0 2  j   2 T0 t 1 2 T0 2  j   2 T0 t
X c  j    e dt   e dt   e dt

3 0 T 2 
2 3 0 T 2 
2 3 0 T 2
Each of three integrals is of the form

T0 2 sin  T0 2  Substituting the appropriate

 j t
e dt 2 value of  for each integral
T0 2 
2 sin T0 2  2 sin    2 T0  T0 2  2 sin    2 T0  T0 2 
X c  j   2  
3  3   2 T0 3   2 T0
CHAPTER
In terms of f (Hz), the above equation becomes
2 sin  fT0  2 sin   f  1 T0  T0  2 sin   f  1 T0  T0 
X c  jf    0.5  0.5
3 f 3   f  1 T0  3   f  1 T0 
1) First term corresponds to the spectrum of the rectangular pulse.
2) The second and third terms are the versions of 1st term shifted by  1/To.
Fig. 3.47 for To = 1.
Normalized Magnitude Spectrum of 20log10  X c ( jf ) / To  dB: Fig. 3.48.
The normalized by To removed the dependence of the magnitude on To.
3) In this case, the peak value of the first sidelobe is below  30 dB, so we may
satisfy the adjacent channel interference specifications by choosing the
mainlobe to be 20 kHz wide:
f = k/To 10,000 = 2/To To = 2  10  4 seconds.
Data transmission rate = 5000 bits/sec.

CHAPTER
Figure 3.47 (p.

252)
The spectrum of
the raised-cosine
pulse consists of
a sum of three
frequency-shifted
sinc functions.

CHAPTER
Figure 3.48
(p. 252)
Spectrum of
the raised-
cosine pulse in
dB, normalized
by T0.

CHAPTER
3.8 Properties of Fourier Representation

1. Four Fourier
representations:
Table 3.2.

CHAPTER
2. Signals that are periodic in time have discrete frequency-domain

representations, while nonperiodic time signals have continuous frequency-
domain representations: Table 3.3.
3.9 Linearity and Symmetry Properties

1. Linearity property for four Fourier representations:
z  t   ax  t   by  t  
FT
Z  j   aX  j   bY  j 
z  t   ax  t   by  t  FS ;o
  Z  k   aX  k   bY  k 
z  n   ax  n   by  n  
DTFT
 Z  e j   aX  e j   bY  e j 
z  n   ax  n   by  n  DTFS ;o
  Z  k   aX  k   bY  k 
CHAPTER
Example 3.30 Linearity in The FS
Suppose z(t) is the
periodic signal depicted
in Fig. 3.49(a). Use the
linearity and the results
of Example 3.13 to
determine the FS
coefficients Z[k].
<Sol.>
1. Signal z(t):
3 1
z t   x t   y t 
2 2
where x(t) and y(t):
Fig. 3.49(b) and (c).
2. X[k] and Y[k]:

CHAPTER
 X  k   1  k   sin  k 4 
From Example 3.13.
x  t  
FS ;2
y  t  
FS ;2
 Y  k   1  k   sin  k 2 
Linearity
3. FS of z(t):
z  t  
FS ;2
 Z  k    3  2k   sin  k 2   1  2k   sin  k 2 
3.9.1 Symmetry Properties: Real and Imaginary Signals
 Symmetry property for real-valued signal x(t):
1. FT of x(t):
*
 
 
X ( j )   x(t )e dt   x* (t )e j t dt
*  j t
(3.37)
   
2. Since x(t) is real valued, x(t) = x(t). Eq. (3.37) becomes


X *
 j    x  t  e j (  )t dt X(j) is complex-conjugate symmetric
X * ( j )  X ( j ) (3.38)
Re X ( j )  ReX ( j ) and ImX ( j )   ImX ( j )
CHAPTER
 The conjugate symmetry property for DTFS:

X * [k ]  X [ N  k ]
Because the DTFS coefficients are N periodic, and thus
X [k ]  X [ N  k ]
 The symmetry conditions in all four Fourier representations of real-valued
signals are indicated in Table 3.4.

CHAPTER
 Complex-conjugate symmetry in FT:
A simple characterization of the output of an LTI system with a real-
valued impulse response when the input is a real-valued sinusoid.
 Continuous-time case:
1. Input signal of LTI system:
Euler’s formula
x  t   A cos  t   
x  t    A / 2  e j t     A / 2  e j  t  
2. Real-valued impulse response of LTI system is denoted by h(t).
3. Output signal of LTI system:
y(t )  H ( j ) ( A / 2)e j (t  argH ( j ))  H ( j ) ( A / 2)e j (t  argH ( j ))
Exploiting the symmetry conditions:
Applying Eq. (3.2)
H ( j )  H ( j ) and linearity
argH ( j )   argH ( j )

y  t   H  j  A cos  t    argH  j  
CHAPTER
 The LTI system modifies the amplitude of the input sinusoid by H(j)and
the phase by arg{H(j)}.
v  ; f  
Fig. 3.50.
Figure 3.50 (p. 257)

A sinusoidal input to an LTI system results in a sinusoidal output of the
same frequency, with the amplitude and phase modified by the system’s
frequency response.
 Discrete-time case:
1. Input signal of LTI system:
x[n] = Acos( n  )
CHAPTER
2. Real-valued impulse response of LTI system is denoted by h[n].

3. Output signal of LTI system:

y  n   H  e j  A cos n    arg H  e j   
 The LTI system modifies the amplitude of the input sinusoid by H(e j)and
the phase by arg{H(e j)}.
 x(t) is purely imaginary:
1. x(t) =  x(t).
2. Eq.(3.37) becomes

X *
 j     x  t  e j (  )t dt
X * ( j )   X ( j ) (3.39)
Re X ( j )   Re X ( j ) and ImX ( j )  ImX ( j )
3.9.2 Symmetry Properties: Even and Odd Signals
1. x(t) is real valued and has even symmetry.
x(t) = x(t) and x ( t) = x(t) x(t) = x( t)
CHAPTER
2. Eq.(3.37) becomes
 Change of variable  =  t
X *
 j    x  t  e  j (  t )
dt

X *  j     x   e j d  X  j 

3. Conclusion:
1) The imaginary part of X(j) = 0: X  ( j )  X ( j )
If x(t) is real and even, then X(j) is real.
2) If x(t) is real and odd, then X(j) =  X(j) and X(j) is imaginary.
3.10 Convolution Properties
 The convolution property is a consequence of complex sinusoids being
eigenfunctions of LTI system.
3.10.1 Convolution of Nonperiodic Signals — Continuous-time case
1. Convolution of two nonperiodic continuous-time signals x(t) and h(t):

y  t   h  t   x  t    h   x  t   d

2. FT of x(t  ):
CHAPTER
1 
x t     X  j  e j ( t  )d
2 
3. Substituting this expression into the convolution integral yields

  1  
y  t    h     X  j  e jt e  j d   d

 2  
1    X  j  e jt d 
 
2   
 j
 h  e d

The inner integral over  as the FT of h(), or H(j). Hence, y(t) may be written
as 
1
y t    H  j  X  j  e j t d
2 
and we identify H(j)X(j) as the FT of y(t).

y(t )  h(t )* x(t ) 
FT
Y ( j )  X ( j )H ( j ) (3.40)
Example 3.31 Solving A Convolution Problem in The Frequency Domain
Let x(t) = (1/(t))sin(t) be the input to a system with impulse response h(t) =
(1/(t))sin(2t). Find the output y(t) = x(t)  h(t).
CHAPTER
<Sol.> x(t) = (1/(t))sin(t)

1. Solving problem: h(t)
y(t) = x(t)  h(t)
Time domain  Extremely Difficult
Frequency domain  Quite Simple
Convolution property
2. From Example 3.26, we have
1,   1,   2
x  t   X  j   
FT
and h  t   H  j   
FT
0,   0,   2
Since y  t   h  t   x  t  
FT
Y  j   X  j  H  j 
1,  
Y  j    y(t) = (1/(t))sin(t)
0,  
Example 3.32 Finding Inverse FT’s by Means of The Convolution Property
Use the convolution property to find x(t), where
4
x  t  
FT
X  j   sin 2  
 2

CHAPTER
<Sol.>
2
1. Write X(j) = Z(j) Z(j), where Z  j   sin  

2. Convolution property:
z  t   z  t  
FT
Z  j   Z  j 
Fig. 3.52 (a)
x t   z t   z t 
1, t 1
3. Using the result of Example 3.25: z  
t  
FT
Z  j 
0, t 1
4. Performing the convolution with itself gives the triangular waveform depicted
in Fig. 3.52 (b) as the solution for x(t).
Figure 3.52 (p. 261)
Signals for
Example 3.32. (a)
Rectangular pulse
z(t). (b)
Convolution of z(t)
with itself gives x(t).
CHAPTER
 Convolution of Nonperiodic Signals — Continuous-time case

If x n  
DTFT
 X  e j  and h n  
DTFT
 H  e j 
y[n]  x[n]* h[n] 

DTFT
Y (e j )  X (e j ) H (e j ) (3.41)
3.10.2 Filtering
1. Multiplication in frequency domain  Filtering.
2. The terms “filtering” implies that some frequency components of the input
are eliminated while others are passed by the system unchanged.
3. System Types of filtering: Fig. 3.53 (a), (b), and (c)
1) Low-pass filter
2) High-pass filter The characterization of DT filter is
3) Band-pass filter based on its behavior in the
4. Realistic filter: frequency range   <    because
1) Gradual transition band its frequency response is 2 -periodic.
2) Nonzero gain of stop band
5. Magnitude response of filter: 20log H  j  or 20log H  e j  [dB]

CHAPTER
v  ; p  
Figure 3.53 (p. 263)

Frequency response of ideal continuous- (left panel) and discrete-time (right
panel) filters. (a) Low-pass characteristic. (b) High-pass characteristic. (c) Band-
pass characteristic.
CHAPTER
6. The edge of the passband is usually defined by the frequencies for which
the response is  3 dB, corresponding to a magnitude response of (1/ 2 ).
 Unity gain = 0 dB
7. Energy spectrum of filter: Y  j   H  j  X  j 
2 2 2
The  3 dB point corresponds to frequencies at which the filter

passes only half of the input power.
 3 dB point Cutoff frequency
Example 3.33 RC Circuit: Filtering
For the RC circuit depicted in Fig. 3.54, the impulse response for the case where
yC(t) is the output is given by
1  t RC
hC  t   e u t 
RC
Since yR(t) = x(t)  yC(t), the impulse response for
the case where yR(t) is the output is given by
1  t RC
hR  t     t   e u t  Figure 3.54 (p. 264)
RC RC circuit with input x(t) and outputs yc(t) and yR(t).
CHAPTER
Plot the magnitude responses of both systems on a linear scale and in dB, and
characterize the filtering properties of the systems.
<Sol.>
1
1. Frequency response corresponding to hC(t): H C j  
i RC  1
i RC
2. Frequency response corresponding to hR(t): H R  j  
i RC  1
3. Magnitude response in linear scale: Fig. 3-55 (a) and (b).
Magnitude response in dB scale: Fig. 3-55 (c) and (d).
4. System corresponding to yC(t):
Low-pass filter H C  j  and H R  j 
Cutoff frequency = c = 1/(RC)
System corresponding to yR(t):
High-pass filter
Cutoff frequency = c = 1/(RC)

CHAPTER
v
Figure 3.55 (p. 265)

RC circuit magnitude responses as a function of normalized frequency RC. (a)
Frequency response of the system corresponding to yC(t), linear scale.
(b) Frequency response of the system corresponding to yR(t), linear scale.
CHAPTER
v
Figure 3.55a&c (p. 265)

RC circuit magnitude responses as a function of normalized frequency RC.
(c) Frequency response of the system corresponding to yC(t), dB scale.
(d) Frequency response of the system corresponding to yR(t), dB scale, shown on
the range from 0 dB to –25 dB.
CHAPTER
The convolution property implies that the frequency response of a system

may be expressed as the ratio of the FT or DTFT of the output to the input .
Y ( j )
For CT system: H ( j )  (3.42)
X ( j ) Frequency response
j Y (e j  )
For DT system: H (e )  (3.43)
X (e j  )
Example 3.34 Identifying a System, Given Its Input and Output
The output of an LTI system in response to an input x(t) = e  2 t u(t) is y(t) =
e  t u(t) . Find the frequency response and the impulse response of this system.
<Sol.>
1 1
1. FT of x(t) and y(t): X  j   and Y  j  
j  2 j  1
Y  j  j  2
H  j   H  j  
X  j  j  1

CHAPTER
3. Impulse response:
 j  1  1 1
H  j       1  h  t     t   et u  t 
 j  1  j  1 j  1
 Recover the input of the system from the output:
CT case DT case
X  j   H inv  j  Y  j  and X  e j   H inv  e j  Y  e j 
where H inv  j   1/ H ( j) and H inv  e j   1/ H (e j )

◆ An inverse system is also known as an equalizer, and the process of
recovering the input from the output is known as equalization.
◆ Causality restriction  An exact inverse system is difficult or impossible
to be implemented!
Ex. Time delay in communication system need an equalizer to introduce a time
advance.
Equalizer is a noncausal system and
Approximation!
cannot in fact be implemented!
CHAPTER
Example 3.35 Multipath Communication Channel: Equalization

Consider again the problem addressed in Example 2.13. In this problem, a
distorted received signal y[n] is expressed in terms of a transmitted signal x[n]
as
y n  x n   ax n  1 , a 1
Use the convolution property to find the impulse response of an inverse
system that will recover x[n] from y[n].
<Sol.>
1. In Example 2.13, we have y[n] = x[n]  h[n], where the impulse response is
1, n0

h[n]  a, n 1
0, otherwise

2. The impulse response of an inverse system, hinv[n], must satisfy
hinv n  h n    n 

DTFT
 H inv  e j  H  e j   1
CHAPTER
H inv  e j  
1 Frequency response of inverse system
H  e j 
3. DTFT of h[n]:
h n  
DTFT
 H  e j   1  ae j
H inv  e j  
1
1  ae  j
The frequency response of the inverse system is then obtained as
hinv n    a  u n 
n
3.10.3 Convolution of Periodic Signals

 Basic Concept:
Periodic h Unstable system

CHAPTER
1. Define the periodic convolution of two CT signals x(t) and z(t), each having
period T, as
y  t   x  t  # z  t    x   z  t    d
T
where the symbol  denotes that integration is performed over a single period
of the signals involved.
2
FS ;
y (t )  x(t ) # z(t )  Y [k ]  TX [k ]Z [k ]
T (3.44)
◆ Convolution in Time-Domain  Multiplication in Frequency-Domain

Example 3.36 Convolution of Two Periodic Signals
Evaluate the periodic convolution of the sinusoidal signal
z  t   2cos  2 t   sin  4 t 
with the periodic square wave x(t) depicted in Fig. 3.56.
<Sol.>
1. Both x(t) and z(t) have fundamental period T = 1.
2
2. Let y  t   x  t  # z  t  FS ;
y (t )  x(t ) # z(t )  Y [k ]  TX [k ]Z [k ]
T

CHAPTER
Figure 3.56 (p. 268)

3. Coefficients of FS representation of z(t):

 1, k  1
 1 2 j k 2

Z k   
 1  2 j  k  2
 0 otherwise
Coefficients of FS representation of x(t):
2sin  k 2  Example 3.13

X k  
2k
CHAPTER
3. FS Coefficients of y(t):
1/  k  1
Y k   X k  Z k    y  t    2   cos  2 t 
 0 otherwise
 Explanation for the origin of the Gibbs phenomenon:
Example 3.14:
1. A partial-sum approximation to the FS representation for x(t) may be obtained
by using the FS coefficients
Xˆ J k   X k W k 
T=1
where
1  J  k  J
W k   
0 otherwise
2. In the time domain, xˆ j (t ) is the periodic convolution of x(t) and
sin  t  2 J  1  T=1
w t  
sin  t 
(3.45)

CHAPTER
3. One period of the

signal w(t) is depicted
in Fig. 3.57 for J = 10.
4. The periodic convolution
of x(t) and w(t) is the area
under time-shifted version
of w(t) on t< ½ .Time-
shifting the sidelobes of w(t)
into and out of the interval
t< ½ introduces ripples
the partial-sum
approximation xˆ j (t ) .
5. As J increases, the mainlobe
and sidelobe widths in w(t)
decrease, but the size of the Figure 3.57 (p. 270)
sidelobe does not change. The signal z(t) defined in Eq. (3.45) when J = 10.
This is why the ripples in xˆ j (t ) becomes more concentrated near
discontinuity of x(t), but remain the same magnitude.
CHAPTER
 Define the discrete-time convolution of two N-periodic sequences x[n] and

z[n] as
N 1
y  n   x  n  # z  n    x k  z n  k 
DTFS of y[n]
k 0
2
DTFS ;
y[n]  x[n] # z[n] 
 Y [k ]  NX [k ]Z [k ]
N (3.46)
◆ Convolution in Time-Domain  Multiplication in Frequency-Domain

◆ The convolution properties of all four Fourier representation are summarized
in Table 3.5.
Table 3.5 Convolution Properties
x(t )  z(t ) 
FT
 X ( j) Z ( j)
FS ; 0
x(t ) # z(t )    TX [k ]Z [k ]
x[n]  z[n] 
FT
 X (e j )Z (e j )
DTFS ; 0
x[n] # z[n]   NX [k ]Z [k ]


xt   X  j  je
d 1
CHAPTER
jt
d
dt 2 
3.11 Differentiation and Integration Properties

 Differentiation and integration are operations that apply to continuous functions.
3.11.1 Differentiation in Time
1. A nonperiodic signal x(t) and its FT, X(j), is related by
 Differentiating both 
xt    X  j e xt    
1 d 1

jt jt
d X j je d
2 
sides with respect to t dt 2 
d
x(t ) 
FT
 j X ( j )
dt
◆ Differentiation of x(t) in Time-Domain  (j)  X(j) in Frequency-Domain
d 
F  x (t )   ( j )  X ( j )  0  0
 dt   0
Example 3.37 Verifying the Differentiation Property

The differentiation property implies that
j
d  at
dt

e u  t   FT
  Verify this result by differentiating and
a  j taking the FT of the result.
CHAPTER
<Sol.> 1. Using the product rule for differentiation, we have

d at
dt

e u  t   aeat u  t   e at  t   ae  at u  t     t 
2. Taking the FT of each term and using linearity, we may write

a  j
d  at
dt

e u  t  
FT
 
a  j
 1 
FT

a  j
 Frequency response of a continuous system
1. System equation in terms of differential equation:
N
dk M
dk Taking FT of N M
        a  j  Y  j    b  j  X  j 
k k
a k k
y t bk x t k k
k 0 dt k 0 dt k both sides k 0 k 0
M
 bk  j 
k
Y  j 
 k 0
X  j 
M
 bk  j 
N
 a  j 
k
k
H  j  
k
k 0
k 0 (3.47)
N
 a  j 
k
2. Frequency response of the system: k
k 0
CHAPTER
Example 3.38 MEMS Accelerometer: Frequency Response and Resonance

The MEMS accelerometer introduced in Section 1.10 is described by the
differential equation
d2 n d
2
y t   y t    2
n yt   xt 
dt Q dt
Find the frequency response of this system and plot the magnitude response
in dB for n = 10,000 rads/s for (a) Q = 2/5, (b) Q = 1, and (c) Q = 200.
<Sol.>
H  j  
1
 j 2   n  j    n2
Q
2. Magnitude response for (a) Q = 2/5, (b) Q = 1, and (c) Q = 200: Fig. 3.58.
 Discussion: Refer to textbook, p. 273.
n = 10,000 rads/s  Resonant condition!
CHAPTER
v
Figure 3.58 (p. 273)

Magnitude of frequency response in dB for MEMS accelerometer for n = 10,000
rad/s, Q = 2/5, Q = 1, and Q = 200.
CHAPTER
 The Differentiation property for a periodic signal

1. The FS representation of a periodic signal x(t):
 
xt    X k  jk 0 e jk0t
Differentiating both d
x t    X k  e jk0t
k  sides with respect to t dt k  
2. Differentiation property of a periodic signal:

d FS ; 0
x  t    jk0 X  k 
dt
◆ Differentiation of x(t) in Time-Domain  (jk0)  X[k] in Frequency-Domain
d 
F  x (t )   ( jk0 )  X [k ] k 0  0
 dt  k 0
Example 3.39 Differentiation Property
Use the differentiation property to find the FS representation of the triangular
wave depicted in Fig. 3. 59 (a).
<Sol.> Fig. 3. 59 (b)
d
1. Define a waveform: z ( t )  y (t )
dt
CHAPTER
Figure 3.59 (p. 274)

Signals for Example 3.39. (a) Triangular wave y(t). (b) The derivative of y(t) is the
square wave z(t).
2. Comparing z(t) in Fig. 3. 59 (b) and the sqare wave x(t) with T0/T = ¼ in
Example 3.13, we have
z(t )  4 x(t )  2 Z [k ]  4 X [k ]  2 [k ]
 0, k 0


z t  FS ; 0
  Z  k    4sin  k 

  2 , k 0

 k

CHAPTER
3. The differentiation property implies that

Except for k = 0
1
Z [k ]  jk0Y [ y ] Y [ y]  Z [k ]
jk0
The quantity Y[0] is the average value of y(t) and is determined by
inspection Fig. 3.59 (b) to be T/2.
Therefore,  T / 2, k 0

 k
y t  FS ; 0
  Y  k    2T sin  

  2 , k 0

 jk 
2 2
3.11.2 Differentiation in Frequency

1. FT of signal x(t):

X  j     jtx t e  jt dt
 Differentiating both d
X  j    xt e  jt
dt

sides with respect to  d 
2. Differentiation property in frequency domain:

d
 jtx  t  
FT
 X  j 
d
CHAPTER
◆ Differentiation of X(j) in Frequency-Domain  ( jt)  x(t) in Time-Domain

Example 3.40 FT of a Gaussian Pulse
Use the differentiation-in-time and differentiation-in-frequency properties to
determine the FT of the Gaussian pulse, depicted in Fig. 3.60 and defined by

g (t )  1/ 2 e  t 2 / 2
D.E. of g(t)
<Sol.> p
1. The derivative of g(t) with respect to t:
d
g (t )  (t / 2 )e t 2 / 2
 tg (t ) (3.48)
dt
2. Differentiation-in-time property:
d
g  t  
FT
 jG  j 
dt
(3.48) tg (t ) 
FT
jG( j ) Figure 3.60 (p. 275)
(3.49)
Gaussian pulse g(t).
d
3. Differentiation-in-time property:  jtg  t  
FT
 G  j 
d
CHAPTER
1 d
tg (t ) 
FT
G( j ) (3.50)
j d
Since the left-hand sides of Eqs. (3.49) and (3.50) are equal, then we have
D.E. of G(j)
G j   G j 
d
d
G j   ce  2 2
4. The integration constant c is determined by noting that (see Appendix A-4)
1/ 2  e dt  1

G  j0   t 2 / 2
The FT of a Gaussian pulse is

also a Gaussian pulse!
1/ 2  e  e 
t 2 / 2  2
 FT /2
3.11.3 Integration
 The operation of integration applies only to continuous dependent variables.
In both FT and FS, we may integrate with respect to time.
In both FT and DTFT, we may integrate with respect to frequency.
CHAPTER
 Case for nonperiodic signal

1. Define
d
yt    x d
t
y (t )  x(t ) (3.51)
 dt
2. By differentiation property, we have
1 Indeterminate at  = 0 X(j0) = 0
Y ( j )  X ( j ) (3.52)
j
3. When the average value of x(t) is not zero, then it is possible that y(t) is not
square integrable.
FT of y(t) may not converge!
W can get around this problem by including impulse in the transform, i.e.
1
Y ( j )  X ( j )  c ( ) c can be determined by
j the average value of x(t)
C =  X(j0)
4. General form of integration property:

CHAPTER
t 1
 x( )d  X ( j )   X ( j 0) ()
FT
(3.53)
 j
Ex. Demonstration for the integration property by deriving the FT of the unit step
1. Unit step signal:
u t      d
t

2. FT of  (t):  (t ) 
FT
 1
1
3. Eq. (3.53) suggests that: u  t  
FT
 U  j      
j
 Check:
1. Unit step: 1 1
u (t )   sgn(t ) (3.54)
2 2
Fig. 3.61
2. Signum function:  1, t  0

sgn  t    0, t  0
 1, t  0


CHAPTER
Figure 3.61 (p. 279)

Representation of a step function as the sum of a constant and a signum function.
3. Using the results of Example 3.28:
1
FT
  ( )
2
The transform of sgn(t) may be derived using the differentiation property
because it has a zero time-average value.
Let
sin t   2 t  jS  j   2
d
 S ()
sgn(t ) FT
dt
We know that S(j0) = 0. This knowledge removes the indeterminacy at  =
0 associated with differentiation property, and we conclude that

CHAPTER
This relationship can be written as S(j) =
2 j ,   0
S  j    2/j  with the understanding that S(j0) = 0.
 0,   0
4. Applying the linearity property to Eq. (3.54) gives the FT of u(t):
1
u  t  
FT
 U  j       This agrees exactly with the transform of the unit
j step obtained by using the integration property.
Table 3.6 Commonly Used Differentiation and Integration Properties.

d
x (t )   j X ( j )
FT
◆ The differentiation dt
and integration d FS ; 0
x  t    jk 0 X  k 
properties of Fourier dt
representation are d
summarized in  jtx  t  
FT
 X  j 
d
Table 3.6.
 jnx  n  
DTFT d
d
X e j  
t 1
 x ( )d  X ( j )   X ( j 0) ( )
FT
 j
CHAPTER
3.12 Time- and Frequency-Shift Properties

3.12.1 Time-Shift Property
1. Let z(t) = x(t  t0) be a time-shifted version of x(t).
2. FT of z(t):
 
Z  j    z t e  jt dt   xt  t 0 e  jt dt
 
3. Change variable by  = t  t0:


x e  j  d  e  jt0 X  j 

Z  j    x e  j  t0 
d  e  jt0

 
◆ Time-shifting of x(t) by t0 in Time-Domain

 (e  j t0)  X(j) in Frequency-Domain
Z ( j)  X ( j) and arg{Z ( j)}  arg{X ( j)}  0t

◆ The time-shifting properties of four Fourier representation are summarized
in Table 3.7.

CHAPTER
Table 3.7 Time-Shift Properties of Fourier Representations
x  t  t0  
FT
e jt0 X  j 
x  t  t0  
FT ; 0
 e jk0t0 X  k 
x n  n0  
DTFT
 e jn0 X  e j 
x n  n0  
DTFS ; 0
 e jk0n0 X k 
Example 3.41 Finding an FT Using the Time-Shift Property

Use the FT of the rectangular pulse x(t) depicted in Fig. 3.62 (a) to determine
the FT of the time-shift rectangular pulse z(t) depicted in Fig. 3. 62 (b).
<Sol.>
1. Note that z(t) = x(t  T1).
2. Time-shift property:
Z ( j)  e jT1 X ( j)
CHAPTER
3. In Example 3.25, we obtained
X  j   sin T0  Z  j   e  jT1 sin T0 

2 2
 
 Frequency response of a discrete-time system
1. Difference equation:
N M
 a yn  k    b xn  k 
k 0
k
k 0
k
 e jk Z  e j 
Taking DTFTof both
z n  k  
DTFT
sides of this equation
  Y e    b e  X e 
N M
 ak e
k 0
 j k j
k 0
k
 j k j
 
M
 bk e  j
k
 
Y e j
 k 0
 
 a e 
j N
X e  j k
k
k 0
CHAPTER
 
M
 bk e  j
k
 
H e j  k 0
(3.55)
 a e 
N
 j k
k
k 0
3.12.2 Frequency-Shift Property The problem is to

express the inverse
1. Suppose that: FT of Z(j) = X(j(   ))
x(t ) 
FT
X ( j) and z(t ) 
FT
Z ( j) in terms of x(t).
2. By the definition of the inverse FT, we have
 
z t   Z  j e X  j    e jt d
1 1
 
jt
d 
2  2 
Substituting variables  =    into above Eq. gives

 
z t   X  j e X  j e jt d  e jt xt 
1 j   t 1
 
jt
d  e
2  2 
◆ Frequency-shifting of X(j) by  in Frequency-Domain [i.e. X(j(   ))]

 (e   t)  x(t) in Time-Domain
CHAPTER
◆ The frequency-shift properties of Fourier representations are

summarized in Table 3.8.
Table 3.8 Frequency-Shift Properties of Fourier Representations
e j t x  t  
FT
X  j     
e jk00t x  t  
FS ; 0
 x  k  k0 
e jn x[n] 
DTFT
 X e j   
e jk00n x[n] 
FS ; 0
 X  k  k0 
Example 3.42 Finding an FT by Using the Frequency-Shift Property
Use the frequency-shift property to determine the FT of the complex sinusoidal
pulse:
e ,
j10t
t 
z t   
 0, t 
<Sol.>
1. Express z(t) as the product of a complex sinusoid ej10t and a rectangular pulse
CHAPTER
1, t 
xt   
0, t 
2. Use the results of Example 3.25, we have
2
x  t  
FT
X  j   sin  

2
Frequency-shift property x  t  
FT
X  j   sin  

3. FT of z(t):
X  j   10  sin    10  
2
e j10t x  t  
FT
z  t  
FT
  10
Example 3.43 Using multiple Properties to Find an FT
Find the FT of the signal
xt  
d
dt
e ut * e u t  2
3t t
<Sol.>
1. To solve this problem, it needs to use three properties: differentiation in time,
convolution, and time shifting.
X  j   jW  j V  j 
CHAPTER
2. Let w(t )  e3t u(t ) and v(t )  et u(t  2)
xt   wt  * vt 

d
dt
X  j   jW  j V  j 
Convolution and differentiation property
3. From the transform pair:
W  j  
1 1
e at u  t  
FT
a  j 3  j
4. Use the same transform pair and the time-shift property to find V(j) by first
writing  j 2
V  j   e 2
e
vt   e 2 e t 2 ut  2
1  j
5. FT of x(t):
je  j 2
X  j   e 2
1  j 3  j 

CHAPTER
3.13 Finding Inverse Fourier Transforms by Using Partial-

Fraction Expansions
A ratio of two
Frequency response
polynomials in j or e j 
3.13.1 Inverse Fourier Transform
1. FT X(j) in terms of a ratio of polynomials in j:
bM  j     b1  j   b0 B j 
M
X  j   
 j N  a N 1  j N 1   a1  j   a0 A j 
Assume that M < N. If M  N, then we may use long division to express X(j)
in the form
Partial-fraction expansion
B j 
   A j 
M N
X  j   f j k 
k
is applied to this term
k 0
Applying the differentiation property
2. Let the roots of the denominator
and the pair  (t ) FT
1 to these
A(j) be dk, k = 1, 2, …, N. These
terms
roots are found by replacing j
with a generic variable v and determining the roots of the polynomial
CHAPTER
v N  a N 1v N 1    a1v  a0  0
5. For M < N, we may the write
M
 bk  j 
k
X  j   k 0
N
Ck, k = 1, 2, …, N are determined
  j  d 
k 1
k
by the method of residues
Assuming that all the roots dk, k = 1, 2, …, N, are distinct, we may write
N
Ck
X  j    Appendix B
k 1 j  d k
In Example 3.24, we derived the transform pair
1
edt u  t  
FT This pair is valid even if d is complex,
j  d provided that Re{d} < 0.
Assuming that the real part of each dk, k = 1, 2, …, N, is negative, we use

linearity to write
CHAPTER
N N
Ck
x  t    Ck e u  t  
dk t FT
X  j   
k 1 k 1 j  d k
Example 3.44 MEMS Accelerometer: Impulse Response
Find the impulse response for the MEMS accelerometer introduced in Sec. 1.10,
assuming that n = 10,000 rads/s, and (a) Q = 2/5, (b) Q = 1, and (c) Q = 200.
<Sol.>
 Frequency response for the MEMS accelerometer:
1
H  j  
n
 j    j   n2
2
Q
Case (a): n = 10,000 rads/s and Q = 2/5
H  j  
1
 j 2  25000 j   100002
CHAPTER
2. Partial-fraction expansion:
1 C1 C2
 
 j 2  25000 j   100002 j  20000 j  5000
The roots of the denominator polynomial are d1 =  20,000 and d2 =  5,0000.
Coefficients C1 and C2:
1
C1   j  20,000
 j   25000  j   10000
2 2
j 20,000
1
  1/15000
j  5,000 j 20,000
1
C2   j  5000 
 j   25000  j   10000
2 2
j 5,000
1
  1/15,000
j  20,000 j 5,000

CHAPTER
 1 / 15000
H  j  
1 / 15000

j  20000 j  5000
h  t   1/15000  e5000t  e20000t  u  t 
Case (b): n = 10,000 rads/s and Q = 1
H  j  
1
 j 2  10000 j   100002
2. Partial-fraction expansion: d1  5000  j5000 3
The roots of the denominator polynomial are
d 2  5000  j5000 3
H  j  

 j / 10000 3  

j / 10000 3 
j  5000  j5000 3 j  5000  j5000 3

CHAPTER
 
h  t    j 10000 3 e 5000t e j 5000 3t
 e 5000t e  j 5000 3t

  
 1/ 5000 3 e 5000t sin 5000 3t u  t  
Case (c): n = 10,000 rads/s and Q = 200
The roots of the denominator
H  j  
1 polynomial are
 j 2  50 j   100002 d1  25  j10,000
2. Impulse response: d 2  25  j10,000
ht   1 / 10000e 25t sin10000t ut 
 Impulse response for (a) Q = 2/5, (b) Q = 1, and (c) Q = 200: Fig. 3.64 (a) ~ (c).
1. For both Q = 2/5 and Q = 1, the impulse response is approximately zero for
t > 1 ms.
2. When Q = 200, the accelerometer exhibits resonant nature.
A sinusoidal oscillation of n = 10,000 rads/sec.

CHAPTER
Figure 3.64 (p. 289)

Impulse response of MEMS
accelerometer.

CHAPTER
3.13.2 Inverse Discrete-Time Fourier Transform

1. Suppose X(e j ) is given by a ratio of polynomial in e j , i.e.
 M e  jM    1e  j   0
X e  
j Normalized to 1
 N e  jN   N 1e  j N 1    1e  j  1
2. Factor the denominator polynomial as
 
N
Ne  jN
  N 1e  j  N 1
    1e  j
 1   1  d k e  j
k 1
Replace e j  with the generic variable v: Find dk, the roots of

v N  1v N 1   2 v N 2     N 1v   N  0 this polynomial
3. Partial-fraction expansion:
Assuming that M < N and all the dk are distinct, we may express X(e j )as
   C
N
j k
X e  j
1 d ek 1 k
Since
CHAPTER
1 Expansions for
 d k  u n  
n DTFT
repeated roots are
1  d k e  j
treated in Appendix B.
The linearity property implies that
N
xn   C k d k  un
n
k 1
Example 3.45 Inverse by Partial-Fraction Expansion

Find the inverse DTFT of 5
 e  j  5
 
X e j 
1
6
1
1  e  j  e  j 2
<Sol.> 6 6
1. Characteristic polynomial: 1 1
v2  v   0
6 6
2. The roots of above polynomial: d1 =  1/2 and d2 = 1/3.
3. Partial-Fraction Expansion:
CHAPTER
5
 e  j  5
6 C1 C2
 
1 1 1 1
1  e  j  e  j 2 1  e  j 1  e  j
6 6 2 3
Coefficients C1 and C2
5 5
 e  j  5  e  j  5
 1  6
C1   1  e  j   6 4
 2  1  1 e  j  1 e  j 2 1  j
1 e
6 6 e j 2 3 e j 2
5 5
 e  j  5  e  j  5
 1  6
C2   1  e  j   6 1
 3  1  1 e  j  1 e  j 2 1  j
1 e
6 6 e  j
3 2 e  j  3
xn  4 1 / 2 un  1 / 3 un

n n

CHAPTER
3.14 Multiplication Property

 Non-periodic continuous-time signals
1. Non-periodic signals: x(t), z(t), and y(t) = x(t)z(t). Find the FT of y(t).
2. FT of x(t) and z(t):
 
xt   X  jv e dv z t   Z  j e jt d
1 1
 
jvt
and
2  2 
 
yt   X  jv Z  j e j  v t ddv
1
2 2  
Change variable:  =   v Inner Part: Z(j)  X(j)
 1  jt
 
1  
y t     2  X  jv  Z j    v  dv  e d 
2
3. FT of y(t): Outer Part: FT of y(t)
1
y (t )  x(t ) z(t ) 
FT
Y ( j)  X ( j) * Z ( j) (3.56)
2
Scaled by 1/2
CHAPTER

where X  j  * Z  j    X  jv Z  j   v dv

◆ Multiplication of two signals in Time-Domain

 Convolution in Frequency-Domain  (1/2)
 Non-periodic continuous-time signals Find the
1. Non-periodic DT signals: x[n], z[n], and y[n] = x[n]z[n]. DTFT of
2. DTFT of y[n]: y[n].
1
y[n]  x[n]z[n] 
DTFT
 Y (e j )  X (e j ) # Z (e j ) (3.57)
2
where the symbol  denotes periodic convolution.
Here, X(e j ) and X(e j ) are 2-periodic, so we evaluate the convolution over
a 2 interval:
 
X  j  # Z  j    X  e j  Z e j  d


◆ Multiplication of two signals in Time-Domain

 Convolution in Frequency-Domain  (1/2)

CHAPTER
 Multiplication property can

be used to study the effects
of truncating a time-domain
signal on its frequency-
domain.
Windowing !
 Truncate signal x(t) by a
window function w(t) is
represented by
y(t) = x(t)w(t)
Fig. 3. 65 (a)
Time Interval
T0  t  T0
Figure 3.65a (p. 293)
The effect of windowing.
(a) Truncating a signal in time by
using a window function w(t).
CHAPTER
1. FT of y(t):
1
y  t  
FT
Y  j   X  j  *W  j 
2
2. If w(t) is the rectangular window depicted
in Fig. 3. 65 (b), we have p;v
W  j   sin T0 
2

Smoothing the details
in X(j) and introducing
oscillation near
discontinuities in X(j).
Figure 3.65b (p. 293)

(b) Convolution of the signal
and window FT’s resulting from
truncation in time.
CHAPTER
 The smoothing is a consequence of the 2/T0 width of the mainlobe of W(j),

while the oscillations near the discontinuities are due to the oscillations in
the sidelobes of W(j).
Example 3.46 Truncating the Impulse Response
The frequency response H(e j) of an ideal discrete-time system is depicted in
Fig. 3. 66 (a). Describe the frequency response of a system whose impulse
response is the ideal system impulse response truncated to the interval  M 
n  M.
<Sol.>
1. The ideal impulse response is the inverse DTFT of H(e j ).
Using the result of Example 3.19, we write
1 n 
h n  1, n M
 n  2 
sin
w n  
2. Let ht[n] be the truncated impulse response: 0, otherwise
h  n  , n M
ht  n    ht [n]  h[n]w[n]
 0, otherwise
CHAPTER
v  , u  
Figure 3.66 (p. 294)

The effect of truncating the impulse
response of a discrete-time system.
(a) Frequency response of ideal system.
(b) F() for  near zero.
CHAPTER
v  , u  
Ht(e j ) is the area under

F( ) between  =  /2
and  = /2.
Figure 3.66 (p. 294)

The effect of truncating the impulse response of a discrete-time
system. (c) F() for  slightly greater than /2.(d) Frequency
response of system with truncated impulse response.
CHAPTER
3. Let ht [n] 

DTFT
 H t (e j )
Using the multiplication property in Eq. (3.57), we have

H t  e j  
1
2   
H  e j   W e j   d

4. Since
1,   /2 On the basis of
He  j
 Example 3.18
0,  / 2    
sin       2 M  1 / 2 
and 
W e j    sin       / 2 
 /2 Discussion: refer to
 
Ht e j

1
2  F  d
 p. 295 in textbook.
 /2
where we have defined


F    H  e j W e j  

W e

 
j   
,   /2

 0,   /2
CHAPTER
 Multiplication of periodic time-domain signals corresponds to convolution

of the Fourier representations.
Same fundamental
FT ; 2 / T
y(t )  x(t ) z(t )  Y [k ]  X [k ]* Z [k ] (3.58) periods of x(t) and z(t)
where
 Non-periodic convolution
X [k ]  Z [k ]   X [ m] Z [ k  m]
m 
of the FS coefficients
 x(t) 與 z(t)基本週期不同時，FS係數 X[k]與 Z[k]需取兩信號基本週期的最小

公倍數(Least common multiple)週期。
Example 3.47 Radar Range Measurement: Spectrum of RF Pulse Train
The RF pulse train used to measure the range and introduced in Section 1.10
may be defined as the product of a square wave p(t) and a sine wave s(t), as
shown in Fig. 3.68. Assume that s(t) = sin (1000t/T). Find the FS coefficients of
x(t).
<Sol.>
1. Suppose that s(t) = p(t) x(t). Multiplication property X [k ]  P[k ]  S[k ]

CHAPTER
Figure 3.68 (p. 297)

The RF pulse is expressed as the product of a periodic square wave and a sine wave.
CHAPTER
2. Fundamental period: Fundamental frequency

For p(t): 0 = 2/T. s(t) = sin (1000t/T) = sin(5000t ).
3. FS coefficients of s(t):
 1/(2 j ), k  500

S [k ]   1/(2 j ), k  500
0,
 otherwise
S[k ]  1/(2 j ) [k  500]  1/(2 j) [k  500]
4. FS coefficients of p(t):
1) Result of Example 3.13
sin(koTo )
P[k ]  e jkToo 2) Time-shift property
k
5. FS coefficients of x(t):
1 sin((k  500)oTo ) 1  j ( k 500)Too sin((k  500)oTo )
X [k ]  e j ( k 500)Too  e .
2 (k  500) 2 (k  500)
6. Fig. 3.69 depicts the magnitude spectrum for 0  k  1000.
CHAPTER
Figure 3.69
(p. 298)
FS magnitude
spectrum for
0  k  1000.
The result is
depicted as a
continuous
curve, due to
the difficulty
of displaying
1000 stems

CHAPTER
 Multiplication property for discrete-time periodic signals:

DTFT ; 2 / T
y[n]  x[n]z[n]   Y [k ]  X [k ] # Z [k ] (3.59)
N 1
X [k ] # Z [k ]   X [m] Z [k  m ]
where Periodic convolution of
m 0
DTFS coefficients
◆ Fundamental period = N.
◆ The multiplication properties of all four Fourier representations are
Table 3.9 Multiplication Properties of Fourier Representations
1
x ( t ) z (t ) FT
X ( j )  Z ( j )
2
FS ; o
x ( t ) z (t )   X [k ]  Z [k ]
1
x[n ]z[n ] 
DTFT
X ( e j ) # Z ( e j )
2
DTFT ; o
x[n]z[n ]   X [k ] # Z [k ]
CHAPTER
3.15 Scaling Property

1. Let z(t) = x(at).
2. FT of z(t) : Changing variable:

Z ( j)   z(t )e  jt

dt   x(at )e  jt dt  = at
 
 (1/ a )  x( )e  j ( / a ) d , a  0

Z ( j )  


(1/ a )  x( )e  j ( / a ) d , a  0
 

Z ( j)  (1/ a )  x( )e j ( / a ) d ,

z(t )  x(at ) 

FT
(1/ a ) X ( j / a). (3.60)
◆ Scaling in Time-Domain  Inverse Scaling in Frequency-Domain
Signal expansion or compression!
Refer to Fig. 3.70.

CHAPTER
v
Figure 3.70 (p. 300)

The FT scaling property. The figure assumes that 0 < a < 1.

CHAPTER
Example 3.48 Scaling a Rectangular Pulse

Let the rectangular pulse Use the FT of x(t) and the scaling property to
find the FT of the scaled rectangular pulse
1, t 1
x(t )   1, t 2
0, t 1 y (t )  
0, t 2
<Sol.>
1. Substituting T0 = 1 into the result of Example 3. 25 gives
2
X ( j )  sin( )
 Scaling property and a = 1/2
2. Note that y(t) = x(t/2).
 2  2
Y ( j )  2 X ( j 2 )  2   sin(2 )  sin(2 ) Fig. 3.71
 2  
 Substituting T0 = 2 into the result of Example 3.25 can also give the answer!
Example 3.49 Using Multiple Performance to Find an Inverse FT
Find x(t) if d  e j 2 
X ( j )  j  
d 1  j ( / 3) 
CHAPTER
Figure 3.71
(p. 301)
Application of the
FT scaling
property in
Example 3.48. (a)
Original time
signal. (b)
Original FT. (c)
Scaled time
signal y(t) = x(t/2).
(d) Scaled FT
Y(j) = 2X(j2).
v

CHAPTER
<Sol.>
1. Differentiation in frequency, time shifting, and scaling property to be used to
solve the problem.
2. Transform pair: 1
s(t )  e u(t )  S ( j ) 
t FT
1  j
X ( j )  j
d
d

e j 2 S ( j / 3)
3. Treatment procedure: scaling  time-shift  differentiation
Define Y(j) = S(j/3).
y(t )  3s(3t )  3e3t u(3t )  3e3t u(t )
Define W(j) = e j2 Y(j/3).
Differentiation
w(t )  y(t  2)  3e3(t 2)u(t  2) property
Since d
X ( j )  j W ( j ) x(t )  tw(t )  3te3(t 2)u(t  2)
d

CHAPTER
 Scaling property in FS:

1. If x(t) is a periodic signal, then z(t) = x(at) is also periodic.
2. If x(t) has fundamental period T, then z(t) has fundamental period T/a.
3. If the fundamental frequency of x(t) is 0, then the fundamental frequency of
z(t) is 0/a.
4. FS coefficients of z(t):
a T /a

 jkaot
Z [k ]  z (t ) e dt
T 0
FS ; a0
z(t )  x(at )    2Z [k ]  X [k ], a0 (3.61)
 Scaling property in Discrete-Time Domain: z[n] = x[pn]

1. First of all, z[n] = x[pn] is defined only for integer values of p.
2. If p> 1, then scaling operation discards information.
Refer to Problem 3.80.

CHAPTER
3.16 Parseval Relationships

 The energy or power in the time-domain representation of a signal is equal to
the energy or power in the frequency-domain representation.
◆ Case for CT nonperiodic signal: x(t)
1. Energy in x(t): The integral inside
 the braces is the FT
Wx  
2
x(t ) dt
 of x(t).
x (t )  x (t ) x  (t )
2
2. Note that
1 
 X  ( j )e jt d

Express x*(t) in terms of its FT X(j): x (t ) 
2 
 1  
 
Wx   x(t )   X ( j )e jt d  dt

 2  
Wx 
1  
2 
X ( j ) 


x 
(t )e  jt

dt d
1  
Wx 
2 
X ( j ) X ( j )d
CHAPTER
 1 
 x(t ) dt   X ( j ) d
2 2
(3.62)
 2 
Normalization!
◆ Energy in Time-Domain Representation
 Energy in Frequency-Domain Representation  (1/2)
◆ The Parseval Relationships of all four Fourier representations are
Table 3.10 Parseval Relationships for the Four Fourier Representations
Representations Parseval Relationships
 1 
 x(t ) dt    d
2 2
FT X ( j )
 2 
1 T
 k 

T 0

2 2
FS x (t ) dt X [ k ]
1 
 n


2 j 2
DTFT x [ n ]  X ( e ) d
2  
1
  k 0
N 1 2 N 1 2
DTFS x[ n ]  X [ k ]
N n 0
CHAPTER
Example 3.50 Calculating Energy in a Signal

Let
sin(Wn)
x[n] 
n
Use Parseval’s theorem to evaluate
 
sin 2 (Wn) Direct calculation in time-domain
  x[n]  
2
n  n   2n 2 is very difficult!

<Sol.>
1. Using the DTFT Parseval relationship in Table 3.10, we have
1 

2
 j
X (e ) d 
2 
2. Since  W
j
1,
x[n]  DTFT
X (e )  
0, W    
1 W

2  W
1d   W / 

CHAPTER
3.17 Time-Bandwidth Product

1. Preview: 1, t  To
x (t )   
FT
X ( j )  2sin(To ) /  Fig. 3.72
0, t  To
v  , p  
Figure 3.72 (p. 305)

Rectangular pulse illustrating the inverse relationship between the time and
frequency extent of a signal.
2. x(t) has time extent 2T0. X(j) is actually of infinite extent frequency.
3. The mainlobe of sinc function is fallen in the interval of < /T0, in which
contains the majority of its energy.
CHAPTER
4. The product of the time extent T0 and mainlobe width 2/T0 is a constant.
5. Compressing a signal in time leads expansion in the frequency domain and
vice versa.
Signla’s time-bandwidth product !
6. Bandwidth: The extent of the signal’s significant frequency content.
1) A mainlobe bounded by nulls.
Ex. Lowpass filter  One half the width of mainlobe.
2) The frequency at which the magnitude spectrum is 1/2 times its peak values.
7. Effective duration of signal x(t):
1/ 2
  t 2 x(t ) 2 dt 

Td     (3.63)
 dt 
 
2
x (t )

8. Effective bandwidth of signal x(t):
1/ 2
   2 X ( j ) 2 d 

Bw      (3.64)
 
   
2
X ( j ) d 
CHAPTER
9. The time-bandwidth product for any signal is lower bounded according to

the relationship
Td Bw  1/ 2 (3.65)
We cannot simultaneously decrease the duration and bandwidth of
a signal.
Uncertainty principle !
Example 3.51 Bounding the Bandwidth of a Rectangular Pulse
Let 1, t  To Use the uncertainty principle to place a lower
x(t )   bound on the effective bandwidth of x(t).
0, t  To
<Sol.>
1. Td of x(t):
1/ 2
 t dt 
To

2
Td   Too 
T

 To dt 
CHAPTER
1/ 2
 To t 2dt 
T 
Td   Too  [(1/(2To ))(1/ 3)t 3 To 1/ 2
]  Td / 3
   To
 To 
dt
2. The uncertainty principle given by Eq. (3.65) states that

Bw  1/(2Td ) Bw  3 /(2To ) Rectangular in time-domain  sinc
function in frequency-domain
3.18 Duality
3.18.1 The Duality Property of the FT Difference in the factor 2 and the
1. FT pair: sign change in the complex sinusoid.
1  
x(t ) 
2 
X ( j )e jt d and X ( j )   x(t )e jt dt

2. General equation:
1 
y (v ) 
2  
z ( )e jv dv (3.66)
Choose  = t and  =  the Eq. (3.66) implies that

CHAPTER
v  ; p  
Figure 3.73 (p. 307)

Duality of rectangular pulses and sinc functions.

CHAPTER
1 
y (t ) 
2 

z ( )e jt d
y(t ) 
FT
z() (3.67)
3. Interchange the role of time and frequency by letting  =   and  = t, then
Eq. (3.66) implies that
1 
y ( ) 
2  
z (t )e jt dt
z(t ) 
FT
2 y( ) (3.68)
Fig. 3.74
4. If we are given by an FT pair
z(t ) 
FT
2 y( ) (3.69) f (t ) 
FT
F ( j) (3.70)
Example 3.52 Applying Duality
Find the FT of 1
x(t ) 
1  jt
CHAPTER
v  ; p  
Figure 3.74 (p. 309)

The FT duality property.
CHAPTER
<Sol.> 1
1. Note that: f (t )  e t u(t ) 
FT
F ( j ) 
1  j
2. Replacing  by t, we obtain
1
F ( jt )  x(t) has been expressed as F(jt).
1  jt
3. Duality property:
X ( j)  2 f ( )  2 eu( ) Difference in the factor

3.18.2 The Duality Property of the DTFS N and the sign change
in the complex
1. DTFS Pair for x[n]: sinusoidal frequency.
N 1
1 N 1
x[n]   X [k ]e jk O n
and X [k ]   x[n]e jk O n
k 0 N n 0
2. The DTFS duality is stated as follows: if
CHAPTER
2
DTFS ; N = time index; k =
x[n]  N
X [k ] (3.71)
frequency index
2
DTFS ; 1
X [n]  N
x[k ] (3.72)
N
3.18.3 The Duality Property of the DTFT and FS
1. FS of a periodic continuous time signal z(t):

z (t )   Z[k ]e
k 
jkot
2. DTFT of an nonperiodic discrete-time signal x[n]:


j
X (e )   x[n]e
k 
 jn
Require
3. Duality relationship between z(t) and X(e j) T = 2
 In the DTFT  t in the FS
Assumption: 0 = 1
n In the DTFT   k in the FS
CHAPTER
4. FS coefficients Z[k]:
1   In the DTFT  t in the FS
  z(t )e
 jkt
Z [k ]  dt n In the DTFT   k in the FS
2 
5. DTFT representation of x[n]:

1 
x[n] 
2 

X ( e j ) e j n  d 
6. Duality property between the FS and the DTFT: if
x[n] 
DTFS
 X (e j ) (3.73)
X (e jt ) 
FS ;1
 x[k ] (3.74)
◆ The Duality Properties of Fourier representations are summarized in
Table 3.11.
Example 3.53 FS-DTFT Duality
Use the duality property and the results of Example 3.39 to determine the
inverse DTFT of the triangular spectrum X(e j) depicted in Fig. 3. 75 (a).
<Sol.>
CHAPTER
Table 3.11 Duality Properties of Fourier Representations
FT f (t ) 
FT
F ( j ) F ( jt ) 
FT
2 f (  )
DTFS ; 2 / N DTFS ; 2 / N
DTFS x[n]   X [k ] X [n ]   (1/ N ) x[ k ]
FS-DTFT x[n ] 
DTFS
 X ( e j ) X (e jt ) 
FS ;1
 x[k ]
p
Figure 3.75 (p. 311)

Example 3.53. (a) Triangular spectrum. (b) Inverse DTFT.
1. Define a time function z(t) =X(e jt).
2. Duality property of Eq. (3.74): if
CHAPTER
z(t ) 
FS ;1
 Z [k ] x[n]  Z [n]
Hence, we seek the FS coefficients Z[k] associated with z(t).
3. Assuming that T = 2, z(t) is a time-shifted version of the triangular wave y(t),
i.e.
z(t )  y(t   / 2)
 , k 0

Z [k ]  e jk / 2Y [k ]   4 j k 1 sin(k / 2)
 , k 0
k 2
4. From x[n] = Z [ n], we obtain
 , n0 Fig. 3.75 (b)


x[n]   4( j )n 1 sin(n / 2)
 , n0
 n 2

CHAPTER
3.19 Exploring Concepts with MATLAB

3.19.1 Frequency Response of LTI Systems from Impulse Response
 Discrete-time system x[n] LTI System y[n]
h[n]
1. The frequency response of a discrete-time LTI system with finite-duration
impulse response may be determined with the use of a finite-duration input
sinusoid that is sufficiently long drive the system to a steady state.
2. Suppose that h[n] = 0 for n < kh and n > lh.
Let input be the finite-duration sinusoid:
v[n]  e j  n  u[n]  u[n  lv ]
3. System output: y[n]  h[n]* v[n]
lb
  h[k ]e
k  kb
j ( n  k )
, lb  n  kh  lv
 h[n]* e jn , lb  n  kh  lv
 H (e j )e jn , lb  n  kh  lv
CHAPTER
Hence, the system output in response to a finite-duration sinusoidal input

corresponds to the output in response to an infinite-duration sinusoidal
input on the interval lh  n < kh + lv.
4. The magnitude and phase response of the system are determined from y[n],
lh  n < kh + lv, by noting that
j  
j ( n  arg H ( e j ) )
y[n]  H (e ) e , lb  n  kb  lb
1) Magnitude: y[n]  H (e j ) , lh  n  kh  lv
2) Phase: arg  y[n]  n  arg H (e j ) , lh  n  kh  lv

Example
Consider the system with impulse response
1 1
h2 [n]   [n]   [n  1]
2 2
Let us determine the frequency response and 50 values of the steady-state
output of this system for the input frequencies  = /4 and 3/4.
<Sol.>
CHAPTER
1. kh = 0 and lh = 1. to obtain 50 values of the sinusoidal steady-state response,

we require that lv  51.
2. MATLAB commands: Obtaining the
>> omega1=pi/4;omega2=3*pi/4; output signals
>> v1=exp(j*omega1*[0:50]);
>> v2=exp(j*omega2*[0:50]);
>> h=[0.5,-0.5];
>> y1=conv(v1,h);y2=conv(v2,h);
3. Plots for the output signals:
1) MATLAB commands for plotting the real and imaginary components of y1:
>> subplot(2,1,1)
>> stem([0:51],real(y1))
>> xlabel('Time');ylabel('Amplitude');
>> title('real(y1)')
Fig 3. 76 (a) and (b)
>> subplot(2,1,2)
>> stem([0:51],imag(y1))
>> xlabel('Time');ylabel('Amplitude');
>> title('imag(y1)')
CHAPTER
real(y1)
0.5
Figure 3.76
Amplitude
(p. 314)
Sinusoidal 0
steady-state
response
computed with -0.5
the use of 0 10 20 30 40 50 60
MATLAB. The Time
values at times 1 imag(y1)
0.5
through 50
represent the
Amplitude
sinusoidal steady-
state response. 0
-0.5
0 10 20 30 40 50 60
Time
CHAPTER
2) MATLAB commands for obtaining the magnitude and phase components from
any element of the vectors of y1 and y2:
Ex. Find the fifth element:
>> H1mag=abs(y1(5)) The phase response
H1mag = is measured in radians.
0.3827
>> H2mag=abs(y2(5))
H2mag =
0.9239
>> H1phs=angle(y1(5))-omega1*5
H1phs =
-5.8905
>> H2phs=angle(y2(5))-omega2*5
H2phs =
-14.5299
 The angle command always returns a value between   and  radians.
 Command: angle(y1(n))-omega1*n may result in answers that differ by integer
multiples of 2 when different values of n are used.
CHAPTER
3.19.2 The DTFS Fast Fourier transform

1. DTFS pair:
N 1 N 1
1
x[n]   X [k ]e jk 0n
(3.10) and X [k ]   x[ n ]e jk 0n
(3.11)
k 0 N n 0
2. MATLAB commands fft and ifft may be used to evaluate the DTFS.
1) Given a length-N vector x representing one period of an N periodic signal x[n],
the command
MATLAB run from 0 to N  1
>> X=fft(x)/N
produces a length-N vector X containing the DTFS coefficients X[k].
◆ The division by N is necessary because fft evaluates the sum in Eq. (3.11) without
dividing by N.
2) Given DTFS coefficients in a vector X, the command
>> x=ifft(X)*N
produces a length-N vector x that represents one period of the time-domain
waveform.
CHAPTER
Example
Consider using MATLAB to solve Problem 3.3 (a) for the DTFS coefficients.
The signal is
 3 
x[n]  1  sin  n /12  
 8 
<Sol.>
>> x=ones(1,24)+sin([0:23]*pi/12+3*pi/8);
1. Period = 24. 2. MATLAB command:
>> X=fft(x)/24
Output:
X=
Columns 1 through 5
1.0000 0.4619 - 0.1913i 0.0000 + 0.0000i -0.0000 + 0.0000i -0.0000 + 0.0000i
Columns 6 through 10
-0.0000 - 0.0000i 0.0000 - 0.0000i 0.0000 - 0.0000i 0.0000 + 0.0000i -0.0000 - 0.0000i
-0.0000 - 0.0000i -0.0000 - 0.0000i 0 -0.0000 + 0.0000i -0.0000 + 0.0000i
-0.0000 + 0.0000i 0.0000 - 0.0000i 0.0000 + 0.0000i 0.0000 + 0.0000i -0.0000 + 0.0000i
-0.0000 - 0.0000i -0.0000 - 0.0000i 0.0000 - 0.0000i 0.4619 + 0.1913i

CHAPTER
3. Explicit answer in rectangular form:

 1, k 0
0.4619  j 0.1913, Answer to
 k 1 Problem 3.3 (a)
X [k ]  
0.4619  j 0.1913, k  23
 0, otherwise on 0  k  23
 Using indices  11 k  12, we may also write the answer by specifying one period
as
 1, k 0
0.4619  j 0.1913, k 1

X [k ]  
0.4619  j 0.1913, k  1
 0, otherwise on  11  k  12
4. Using ifft, we may reconstruct the time-domain signal and evaluate the first four
values of the reconstructed signal with the command:
>> xrecon=ifft(X)*24;
>> xrecon(1:4);
CHAPTER
ans =
1.9239-0.0000i 1.9914+0.0000i 1.9914+0.0000i 1.9239-0.0000i
Example
The partial-sum approximation used in Example 3.7 is easily evaluated in MATLAB:
>> k=1:24;
>> n=-24:25;
>> B(1)=25/50; % coeff for k=0
>> B(2:25)=2*sin(k*pi*25/50)./(50*sin(k*pi/50));
>> B(26)=sin(25*pi*25/50)./(50*sin(25*pi/50)); % coeff for k =N/2
>> xjhat(1,:)=B(1)*cos(n*0*pi/25);
% term in sum for k=0
%accumulate partial sums
>> for k=2:26
xjhat(k,:)=xjhat(k-1,:)+B(k)*cos(n*(k-1)*pi/25);
end
 This set of commands produces a matrix xjhat whose (J + 1)st row corresponds to
xˆ J [n ]
CHAPTER
3.19.3 The FS
 Trigonometric FS expression of x(t) in Example 3.14:

x(t )   B[k ]cos(k0t ) (3.29)
k 0
where
B[0]  2T0 / T
2sin(k 2 T0 / T )
B[k ]  , k 0 (3.28)
k
Let the partial-sum approximation to the FS in Eq. (3.29), be given by
 J
x J  t    B  k  cos  k0t 
k 0
1. We must use sufficiently closely spaced samples to capture the details in xˆ J (t )

2. Requirement:
The highest-frequency term in sum: cos( J max0t ) is well approximated by the
sampled signal: cos( J max0nTs )

CHAPTER
3. With plot command, a good approximation to continuous cosine can be achieved

by the order of 20 samples per period.
Ts  T /(20 J max )
Note that the total number of samples in one period is then 20 Jmax.
4. Assuming Jmax = 99 and T = 1, the MATLAB command for computing the partial-
sum from the given B[k] is as below:
>> t=[-(10*jmax-1):10*jmax]*(1/(20*99));
維度有一些錯誤!
>> xjhat(1,:) = B(1)*cos(t*0*2*pi/T);
>> for k=2:100
xjhat(k,:) = xjhat(k-1,:)+B(k)*cos(t*(k-1)*2*pi/T);
end
5. The row of xjhat represents samples of a continuous-valued function.
We should use plot instead of stem to display xjhat.
The partial-sum for J = 5 can be displayed by the command:
plot(t, xjhat(6,:))

CHAPTER
3.19.4 Frequency Response of LTI Systems Described by Differential or

Difference Equations
1. MATLAB commands:
freqs Frequency response of LTI system described by differential eq.
freqz Frequency response of LTI system described by difference eq.
2. MATLAB command:
h=freqs(b,a,w) returns the values of the CT system frequency response given by
Eq. (3.47) at the frequencies specified in the vector w, where
b  [bM , bM 1,..., b0 ]
Coefficients of the differential equation.
a  [aN , aN 1,..., a0 ]
M
 bk  j 
k
H  j   k 0
N (3.47)
 a  j 
k
k
k 0

CHAPTER
Example
The frequency response of the MEMS accelerometer depicted in Fig. 3.58 for Q =1
is obtained via the MATLAB commands:
-155
>> w=[0:100:100000];
>> b=1; -160
>> a=[1 10000 10000*10000]; -165
>> H=freqs(b,a,w);
>> plot(w,20*log10(abs(H))) -170
-175
3. MATLAB command:
h=freqz(b,a,w) returns the -180
values of the DT system -185

frequency response given
-190
by Eq. (3.55) at the
frequencies specified in the -195
vector w, where
-200
b  [bM , bM 1,..., b0 ] 0 2 4 6 8 10
4
x 10
Coefficients of the difference equation.
a  [aN , aN 1,..., a0 ]
CHAPTER
 The entries of w must lie between 0 and 2.
b e 
M
 j k
k
H e j
 k 0
(3.55)
 a e 
N
 j k
k
DT periodic
k 0 signals
3.19.5 Time-Bandwidth Product
1. Command fft can be used to evaluate DTFS and explore the time-bandwidth
product property.
2. Example 3.6
1, n M
1) Time-domain signal: x[n]  
0, M  n  N  M
1, n M
2) DTFS coefficients: x[n]  
0, M  n  N  M
3) Parameter definitions:
Td = 2M +1 =the nonzero portion of one period of x[n]
CHAPTER
Bw  N/(2M +1) = Bandwidth = the “frequency” of the first null of X[k]
 The time-bandwidth product for the square wave, Td Bw  N, is independent of M.
4) MATLAB command:
>> N=50; Fig. 3. 12 (b)
>> M=12;
>> x=[ones(1,M+1),zeros(1,N-2*M-1),ones(1,M)];
0.5
>> X=fft(x)/N;
>> k=[0:N-1]; %frequency index 0.4
>> stem(k,real(fftshift(X)))
0.3
fft  find DTFS coefficients
stem  display the results 0.2
real  suppress any small

0.1
imaginary components
fftshift  reorder the elements of 0
the vector X to generate
the DTFS coefficients -0.1
centered on k = 0
-0.2
0 10 20 30 40 50

CHAPTER
>> N=50;
>> M=4; Fig. 3. 12 (a)
>> x=[ones(1,M+1),zeros(1,N-2*M-1),ones(1,M)];
>> X=fft(x)/N;
>> k=[0:N-1]; %frequency index 0.2
>> stem(k,real(fftshift(X)))
0.15
0.1
0.05
-0.05
0 10 20 30 40 50

CHAPTER
3. Effective duration and bandwidth for the discrete-time periodic signals:

1 1
 ( N 1) / 2
2 2  ( N 1) / 2
2 2
   
2 2
n x[ n ]  k X [ k ] 
Td   ( N 1) / 2  Bw   ( N 1) / 2 
n  ( N 1) / 2 k  ( N 1) / 2
(3.75) (3.76)
 2   2 
  x[ n ]    X [ k ] 
 n  ( N 1) / 2   n  ( N 1) / 2 
 MATLAB function for >> function TBP=TdBw(x)
computing the product >> %Compute the Time-Bandwidth product using the DTFS
TdBw: >> %One period must be less than 1025 points
>> N=1025;
1) The length of input vector >> M=(N-max(size(x)))/2;
x = odd and centers on >> xc=[zeros(1,M),x,zeros(1,M)];
middle >> % center pulse within a period
2) .* = element-by-element >> n=[-(N-1)/2:(N-1)/2];
>> n2=n.*n;
product
>> Td=sqrt((xc.*xc)*n2'/(xc*xc'));
3) * = inner product >> X=fftshift(fft(xc)/N); % evaluate DTFS and center
4) ' = complex-conjugate >> Bw=sqrt(real((X.*conj(X))*n2'/(X*X')));
transpose >> TBP=Td*Bw;
CHAPTER
4. Use the function TdBw to evaluate the time-bandwidth product for two
rectangular, raised cosine, and Gaussian pulse train as follows:
>> x=ones(1,101); % 101 point rectangular pulse
>> TdBw(x)
ans =
788.0303
>> x=ones(1,301); % 301 point rectangular pulse
>> TdBw(x)
ans =
1.3604e+003
>> x=0.5*ones(1,101)+cos(2*pi*[-50:50]/101); % 101 point rectangular pulse
>> TdBw(x)
ans =
277.7327
>> x=0.5*ones(1,301)+cos(2*pi*[-150:150]/301); % 301 point rectangular pulse
>> TdBw(x)
ans =
443.0992

CHAPTER
>> n=[-500:500];
>> x=exp(-0.001*(n.*n)); % Narrow Gaussian pulse
>> TdBw(x)
ans =
81.5669
>> x=exp(-0.0001*(n.*n)); % Broad Gaussian pulse
>> TdBw(x)
ans =
81.5669
 Note that the Gaussian pulse trains have the smallest time-bandwidth product.
 Time-bandwidth product is identical for both the narrow and broad Gaussian pulse
trains .

CHAPTER


Ch03 - Fourier Representations of Signals LTI Systems PDF

Uploaded by

Copyright:

Available Formats

You might also like

Ch03 - Fourier Representations of Signals LTI Systems PDF

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Ch03 - Fourier Representations of Signals LTI Systems PDF

Uploaded by

Copyright:

Available Formats

CHAPTER

Fourier Representations of Signals & LTI Systems

Signals and Systems_Simon 1

Figure 3.1 (p. 196)

 Continuous-time LTI system

c  c e j arg{c} where c  a 2  b2 and arg{c}  tan 1 

Signals and Systems_Simon 4

Signals and Systems_Simon 5

[B] Discrete-time case:

3.3 Fourier Representations for Four classes of Signals

Signals and Systems_Simon 7

There are only N distinct complex sinusoids of the form exp(jk0n)

 Symmetries in x[k]: k =  (N  1)/2 to (N  1)/2

Signals and Systems_Simon 9

2. DTFT of discrete-time signal:

3.4 Discrete-Time Periodic Signals: The Discrete-Time

N 1 Fourier coefficients; Frequency domain

Figure 3.5 (p. 203)

Figure 3.6 (p. 204)

 X [2]e  j 2 n / 3  X [1]e  j n / 3  X [0]  X [1]e j n / 3  X [2]e j 2 n / 3  X [3]e j n

Signals and Systems_Simon 14

Figure 3.8 (p. 206)

as shown in Fig. 3.9.

Figure 3.9 (p. 207)

 e j 2 / 3e j 6 n / 9  2e j / 3e j 4 n / 9  1  2e j / 3e j 4 n / 9  e  j 2 / 3e j 6 n / 9

Figure 3.10 (p. 208)  /3

1. Period of DTFS coefficients = N o = 2/N

Signals and Systems_Simon 19

3. For k = 0,  N,  2N, …, we have

1 sin k 0 2M  1 / 2

Signals and Systems_Simon 21

Figure 3.12 (p. 211)

Signals and Systems_Simon 22

 Symmetry property of DTFS coefficient: X[k] = X[ k].

3. Use X[m] = X[ m] and the identity No = 2 to obtain

4. Define the new set of coefficients

 A similar expression may be derived for N odd.

where J  N/2. This approximation contains the first 2J + 1 terms centered on k

Figure 3.14a (p. 213)

Signals and Systems_Simon

Figure 3.15 (p. 214)

Signals and Systems_Simon 28

Signals and Systems_Simon 29

3.5 Continuous-Time Periodic Signals: The Fourier Series

3. Notation: The variable k determines the

and choose the coefficients X[k] according to Eq.(3.20.)

2. If x(t) is square integrable,

then the MSE between x(t) and x̂  t  is zero, or, mathematically,

x(t )  xˆ  t  at all values of t; it simply implies that there is zero power

Signals and Systems_Simon 33

Signals and Systems_Simon 34

Signals and Systems_Simon 35

Signals and Systems_Simon 37

2. Putting the fraction over a common denominator results in

1. Fundamental frequency: o = 2/T.

2 sin k 0T0  2T0

Signals and Systems_Simon 41