Discrete-Time Signals and Systems

主講人：虞台文
Discrete-Time Signals and Systems

Introduction
The Taxonomy of Signals
 Signal: A function that conveys information

Amplitude
Continuous Discrete
Continuous continuous-time
analog signals
signals
Time
discrete-time
Discrete digital signals
signals
Signal Process Systems
Facilitate the extraction • Filters
of desired information • Parameter estimation
e.g.,

Signal
Signal
signal Processing
Processing output
System
System
Signal Process Systems
continuous-time signal continuous-time signal
analog
analog output
signal system
system

discrete-time signal discrete-time signal

discrete-
discrete-
signal time
time output
system
system

digital signal digital signal

digital
digital output
signal system
system
Signal Process Systems
 A important class of systems

Linear
Linear Shift-Invariant
Shift-Invariant Systems.
Systems.
 In particular, we’ll discuss

Linear
Linear Shift-Invariant
Shift-Invariant Discrete-Time
Discrete-Time Systems.
Systems.
Discrete-Time Signals and Systems

Discrete-Time Signals---

Sequences
Representation by a Sequence

 Discrete-time system theory

– Concerned with processing signals that are
represented by sequences.

xx {{xx((nn)},
)}, 
  nn

x(n)

3 4 5 6 7 n
-8 -7 -6 -5 -4 -3 -2 -1 1 2 8 9 10
 Important Sequences
 Unit-sample sequence (n)
 Sometime call (n)
 a discrete-time impulse; or
 an impulse

(n)
11 nn 00
((nn))  
00 nn  00 n
-8 -7 -6 -5 -4 -3 -2 -1 1 2 3 4 5 6 7 8 9 10
Important Sequences
 Unit-step sequence u(n)

11 nn 00  Fact:

uu((nn))  
00 nn 00 ((nn)) uu((nn))uu((nn11))

u(n)

n
-8 -7 -6 -5 -4 -3 -2 -1 1 2 3 4 5 6 7 8 9 10
Important Sequences
 Real exponential sequence

xx((nn))  aa nn

x(n)
...
... n
-8 -7 -6 -5 -4 -3 -2 -1 1 2 3 4 5 6 7 8 9 10
Important Sequences
 Sinusoidal sequence
xx((nn))  AAcos(
cos(nn
00  ))

x(n)

n
Important Sequences
 Complex exponential sequence
((jj00))nn
xx((nn))  ee
Important Sequences
 A sequence x(n) is defined to be periodic with
period N if

xx((nn))  xx((nn  N
N)) for
for all
all N
N
j 0 n
 Example: consider x ( n)  e
x(n)  e j0 n  e j0 ( n  N )  e j0 N e j0n  x(n  N )
2 k 2 must be a rational
 0 N  2 k N 
0 0 number
Energy of a Sequence
 Energy of a sequence is defined by

nn
EE   || xx((nn)) ||
22

nn

Operations on Sequences
 Sum
xx yy {{xx((nn)) yy((nn)}
)}
 Product
xxyy {{xx((nn))yy((nn)}
)}
 Multiplication
xx {{xx((nn)}
)}
 Shift
yy((nn)) xx((nnnn00))
Sequence Representation
Using delay unit

xx((nn)) 
xx((kk))((nnkk))
kk


x(n)
a-3
a1
2 7 n
-8 -7 -6 -5 -4 -3 -2 -1 1 3 4 5 6 8 9 10

a2 a7

xx((nn))aa33((nn33))aa11((nn11))aa22((nn33))aa77((nn77))
Discrete-Time Signals and Systems

Linear Shift-Invariant
Systems
Systems

x(n) TT[[]] y(n)=T[x(n)]

Mathematically modeled as a
unique transformation or
operator.
Linear Systems

x(n) TT[[]] y(n)=T[x(n)]

ax11((nn))bx
TT[[ax bx22((nn)])] aT
aT[[xx11((nn)])]bT
bT[[xx22((nn)])]
Examples:

x(n) TT[[]] y(n)=T[x(n)]

Ideal Delay System y (n)  x(n  nd )

k M
1
Moving Average y ( n)   x(n  k )
M 1  M 2  1 k  M1
n
Accumulator y ( n)   x(k )
k  
Examples:

x(n) Are
Are these
TT[[]system
these system
]
linear?
linear?
y(n)=T[x(n)]

Ideal Delay System y (n)  x(n  nd )

k M
1
Moving Average y ( n)   x(n  k )
M 1  M 2  1 k  M1
n
Accumulator y ( n)   x(k )
k  
Examples:

x(n) TT[[]] y(n)=T[x(n)]

A Memoryless System y (n)  [ x(n)]

2

Is
Is this
this system
system linear?
linear?
 Linear Systems


x(n)
xx((nn))  xx((kk))((nnkk))
 TT[[]] y(n)=T[x(n)]  
yy((nn))TT  xx((kk))((nnkk))
kk 
k k
 


 
y ( n)   x(k )T [(n  k )]   x(k )h (n)
k   k  
k

k impulse
時時
間間
時之

n
於

之
輸
出
值
Shift-Invariant Systems

x(n) y(n)=T[x(n)]
TT[[]]
x(n-k) y(n-k)
x(n) y(n)

x(n-1) y(n-1)

x(n-2) y(n-2)
Shift-Invariant Systems

x(n) y(n)=T[x(n)]
TT[[]]
x(n-k) y(n-k)
x(n) y(n)

輸入
x(n-1)
輸入 // 輸出關係僅
輸出關係僅 y(n-1)
與時間差有關
與時間差有關
x(n-2) y(n-2)
 Linear Shift-Invariant Systems


x(n)
xx((nn))  xx((kk))((nnkk))
 TT[[]] y(n)=T[x(n)]  
yy((nn))TT  xx((kk))((nnkk))
kk 
k k
 


 
y ( n)   x(k )T [(n  k )]   x(k )h(n  k )
k   k  
k impulse
時時
間間
時之

僅
與
時
間
差
有
關
n
於

之
輸
出
值
 Linear Shift-Invariant Systems


x(n)
xx((nn))  xx((kk))((nnkk))
 TT[[]] y(n)=T[x(n)]  
yy((nn))TT  xx((kk))((nnkk))
kk 
k k
 


 
y ( n)   x(k )T [(n  k )]   x(k )h(n  k )
k   k  
k impulse
時時
間間
時之

僅
與
時
間
差
有
關
n
於

之
輸
出
值
Impulse Response
x(n)=(n) h(n)=T[(n)
TT[[]] ]
0 0

0 0
 Convolution Sum
(n) h(n
TT[[]] )
x(n) y(n)

y ( n)   x(k )h(n  k )  x(n) * h(n)
k  
convolution
AAlinear
linearshift-invariant
shift-invariantsystem
systemisiscompletely
completely
characterized
characterizedbybyits
itsimpulse
impulseresponse.
response.
Characterize a System

x(n) h(n)
h(n) x(n)*h(n)
Properties of Convolution Math

y ( n)   x(k )h(n  k )  x(n) * h(n)
k  


y ( n)   h ( k ) x ( n  k )  h ( n) * x ( n)
k  

xx((nn))**hh((nn))  hh((nn))**xx((nn))
Properties of Convolution Math

x(n) hh11(n)
(n) hh22(n)
(n) y(n)

x(n) hh22(n)
(n) hh11(n)
(n) y(n)

x(n) hh1(n)*h (n)

1(n)*h22(n)
y(n)

These
These systems
systems are
are identical.
identical.
Properties of Convolution Math

hh11(n)
(n)

x(n) + y(n)
hh2(n)
2(n)

x(n) hh1(n)+h (n)

1(n)+h22(n)
y(n)

These
These two
two systems
systems are
are identical.
identical.
Example
x ( n)  u ( n)  u ( n  N )

a n n0
h( n)   y(n)=?
0 n0
0 1 2 3 4 5 6

0 1 2 3 4 5 6
Example

y ( n)  x ( n) * h( n )   x(k )h(n  k )
k  

x(k)
k
0 1 2 3 4 5 6

h(k) k
0 1 2 3 4 5 6

h(0k)
k
0 1 2 3 4 5 6
Example

y ( n)  x ( n) * h( n )   x(k )h(n  k )
k  

x(k)
k
0 1 2 3 4 5 6
compute y(0)
h(0k)
k
0 1 2 3 4 5 6
compute y(1)
h(1k)
k
0 1 2 3 4 5 6
How to computer y(n)?
Example
Two
Two conditions
conditions have
have to
to be
be considered.
considered.


y ( n)  x ( n) * h( n )   x(k )h(n  k )
k   n<N
n<N and
and nN.
nN.
x(k)
k
0 1 2 3 4 5 6
compute y(0)
h(0k)
k
0 1 2 3 4 5 6
compute y(1)
h(1k)
k
0 1 2 3 4 5 6
How to computer y(n)?
Example

y ( n)  x ( n) * h( n )   x(k )h(n  k )
k  

n<N
 ( n 1) 1
n n
1  a a n
 a
y (n)   a n  k a n  a  k a n 1

k 0 k 0 1 a 1  a 1

nN
N n N
N 1 N 1
1  a a n
 a
y (n)   a n  k a n  a  k a n 1

k 0 k 0 1 a 1  a 1
 Example

y ( n)  x ( n) * h( n )   x(k )h(n  k )
k  

5
4
n<N
 ( n 1) 1
3 n n
1  a a n
 a
2 y (n)   a n  k a n  a  k a n 1

1
0
k 0 k 0 1 a 1  a 1
0 5 10 15 20 25 30 35 40 45 50

nN
N n N
N 1 N 1
1  a a n
 a
y (n)   a n  k a n  a  k a n 1

k 0 k 0 1 a 1  a 1
Impulse Response of
the Ideal Delay System

Ideal Delay System y (n)  x(n  nd )

By letting x(n)=(n) and y(n)=h(n),
hh((nn))((nnnndd))

(n
(n nndd))
0 1 2 3 4 5 6 nd
Impulse Response of
the Ideal Delay System

你必須知道 x(n) * (n  nd )  x(n  nd )

(n nd) 扮演如下功能：

• Shift; or
• Copy

(n
(n nndd))
0 1 2 3 4 5 6 nd
Impulse Response of
the Moving Average
k M
1
Moving Average y ( n)   x(n  k )
M 1  M 2  1 k  M1
MM
11
hh((nn))   ((nnkk))
M1 M
M
1 M2 11kkMM1
2 1

 11

hh((nn))M M11nnM
M M22
M11MM2211
00 otherwise
otherwise


... ...
M1  0  M2
你能以(n
你能以 (n k) 解釋嗎??
k)解釋嗎
Impulse Response of
the Accumulator
n
Accumulator y ( n)   x(k )
k  

nn
hh((nn)) 
((kk))uu((nn))
kk


...
0
你能 解釋嗎 ??
你能解釋嗎
Discrete-Time Signals and Systems

Stability and Causality

Stability
 Stable systems --- every bounded input
produce a bounded output (BIBO)
 Necessary and sufficient condition for a BIBO

 
SS   || hh((kk))|| 

kk

Prove
Necessary Condition for Stablility

 Show that if x is bounded and S < , then y is

bounded.

 
| y (n) |  h(k )x(n  k )  M  | h(k ) |  
k   k  

where M = max x(n)

Prove
Sufficient Condition for Stablility

 Show that if S = , then one can find a

bounded sequence x such that y is unbounded.

 h* (  n)
 h( n)  0
Define x ( n)   | h(  n) |
0 h( n)  0

 
| h ( k ) |2
y (0)   x( k )h(k )   S
k   k   | h( k ) |
Causality
 Causal systems --- output for y(n0) depends only
on x(n) with n n0.
 A causal system whose impulse response h(n)
satisfies

hh((nn))  00 for
for nn  00
Example:
 Show that the linear shift-invariant system with
impulse response h(n)=anu(n) where |a|<1 is sta
ble.

 
11
SS   || h
h ((kk))|| 
 
kk
a
a 
 
 

kk00 kk00 11aa
Discrete-Time Signals and Systems

Linear Constant-Coefficient
Difference Equations
N-th Order Difference Equations
NN MM


aa yy((nnkk))
kk00
kk bb xx((nnkk))
kk00
kk

Examples:
Ideal Delay System y (n)  x(n  nd )
1 k M
Moving Average y ( n)  
M  1 k 0
x(n  k )

Accumulator y (n)  y (n  1)  x(n)

Compute y(n)
NN MM


aa yy((nnkk))
kk00
kk bb xx((nnkk))
kk00
kk

NN MM
aakk bbkk
yy((nn)) yy((nnkk)) xx((nnkk))
kk11 aa0 kk00 aa0
0 0
The Ideal Delay System
yy((nn))  xx((nnnndd)) hh((nn))  ((nnnndd))

x(n) Delay Delay ... Delay y(n)

x(n) nd sample delays y(n)

The Moving Average
11 kkMM
yy((nn))   xx((nnkk))
M 11 kk00
M

1 k M
h( n)  
M  1 k 0
( n  k )

1
  u (n)  u (n  M  1) 
M 1
1
  (n)  (n  M  1)  * u (n)
M 1
The Moving Average
11 kkMM
yy((nn))   xx((nnkk))
M 11 kk00
M

11
hh((nn))  ((nn))((nnMM 11)) **uu((nn))
M 11
M
Attenuator
1 ++ Accumulator
M 1
_ system

M+1 sample
delay
Discrete-Time Signals and Systems

Frequency-Domain
Representation of
Discrete-Time Signals and
Systems
Sinusoidal and Complex
Exponential Sequences

 Play an important role in DSP


y ( n)   h(k ) x ( n  k )
k  

x(n)  e jn LTI

LTI  

h ( k ) e j ( nk )

k  

  
   h(k )e  jk e jn
h(n)
h(n)  k   

j jn
 H ( e )e
Frequency Response

j n H ( e j ) j jn
e H (e )e
eigenvalue
eigenfunction


H (e )    h ( k )e
j
j)  jkjk
H ( e h ( k ) e
kk

Frequency Response


H((ee )) 
hh((kk))ee
j jkjk
H j

kk


j
j)  H (e j
j)  jH (e j
H
H (e )  HRR (e )  jH (e j))
( e

j j HH((eejj))
H((ee ))||H
H j
H((ee ))||ee
j

phase
magnitude
Example:
The Ideal Delay System

yy((nn))  xx((nnnndd)) hh((nn))  ((nnnndd))

 
H ( e j )   h (
k  
k ) e  jn
   (
k  
k  nd ) e  jn
 e  jnd

j
magnitude | H (e ) | 1

phase H (e j )  nd
 Example:
The Ideal Delay System

x(n)  A cos(0 n  ) y (n)  A cos[0 (n  nd )  ]

H (e j )  e  jnd
A A A j  j  0 n  j  0 n d A  j   j 0 n j 0 n d
x ( n )  e j  e j 0 n  e  j e  j 0 n y (n)  e e e  e e e
2 2 2 2
A A
 e j  e j 0 ( n  n d )  e  j e  j 0 ( n  n d )
2 2
Periodic Nature of
Frequency Response
 j j j((22mm))
H((ee )) H
H H((ee
j ))
H((ee )) 
hh((kk))ee
j jkjk
H j

kk
 m00,,11,,22,,
m 

H (e j (  2  )
)  h ( k )e
k  
 jk (  2  )


  h (
k  
k ) e  jk

 H ( e j )
Periodic Nature of
Frequency Response
 j j j((22mm))
H((ee )) H
H j
H((ee ))
H((ee )) 
hh((kk))ee
j jkjk
H j

kk
 m00,,11,,22,,
m 

| H ( e j ) |

4 3 2   2 3 4 

Periodic Nature ofGenerally, we choose
Frequency Response
To represent one period in
frequency domain.
 j j j((22mm))
H((ee )) H
H j
H((ee ))
H((ee )) 
hh((kk))ee
j jkjk
H j

kk
 m00,,11,,22,,
m 

| H ( e j ) |

4 3 2   2 3 4 

Periodic Nature of
Frequency Response
 j j j((22mm))
H((ee )) H
H j
H((ee ))
H((ee )) 
hh((kk))ee
j jkjk
H j

kk
 m00,,11,,22,,
m 
| H ( e j ) |

  
High Low High
Frequency Frequency Frequency
 Ideal Frequency-Selective Filters
| H ( e j ) |
Lowpass Filter 1

 c c  
Bandstop Filter | H ( e j ) |
1

 b a a b  
| H ( e j ) |
Highpass Filter 1

 b a a b  
 Moving Average

11 kkMM j
H (e )   h( k )e  jk

yy((nn))   xx((nnkk)) k  
M 11 kk00
M 1 M 1  1  e  j( M 1) 
 
M  1 k 0
e  jk
 
M 1  1 e  j


11 MM
hh((nn))  ((nnkk))
M 11kk00
M

1  e  j( M 1) / 2 (e j( M 1) / 2  e  j( M 1) / 2 ) 
 
 j / 2 j / 2  j / 2
M 1  e (e e ) 
1   jM / 2 (e j( M 1) / 2  e  j( M 1) / 2 ) 
  e j / 2  j / 2

M 1 (e e ) 
h(n)
1   jM / 2 sin[( M  1) / 2] 
0 0 M   e 
M 1  sin( / 2) 
Moving Average
j 1   jM / 2 sin[( M  1) / 2] 
H (e )   e 
M 1 sin( / 2) 

j 1 sin[( M  1) / 2]
| H (e ) |
M 1 sin( / 2)
 • M=4
• Lowpass
Moving Average • Try larger M

j 1   jM / 2 sin[( M  1) / 2] 
H (e )   e 1.5

M 1 1
sin( / 2) 
0.5

1 sin[( M  1) / 2]
-4 -3 -2 -1 0 1 2 3 4
-0.5
j
| H (e ) |
M 1 sin( / 2) 1

0.5

0
-4 -3 -2 -1 0 1 2 3 4
-0.5

-1
Discrete-Time Signals and Systems

Representation of
Sequences by
Fourier Transform
Fourier Transform Pair
Synthesis
11 
xx((nn))  XX((eejj))eejjnndd Inverse Fourier Transform
22  (IFT)

Analysis
nn
XX((eejj)) 
 x ( n )e
x ( n ) e j
jnn Fourier Transform
nn
(FT)

 1 

j j n
Prove x ( n) 
2  
X ( e ) e d

nn
XX((ee )) 
j j 
j
xx((nn))ee
nn

jnn n=m 

e j  ( n  m ) d  2 

1 
2  
j jn
x ( n)  X ( e ) e d

1  m 
 
2 m 
 
 x ( m ) e  jm jn
e d

1 m  
   e d
2 m  
x ( m ) e  jm jn

1 m  
 
2 m  
x ( m )  e j  ( n  m ) d

 1 

j j n
Prove x ( n) 
2  
X ( e ) e d

nn
XX((ee )) 
j j 
j
xx((nn))ee
nn
jnn nm 
e j( n  m ) d
 1 

j ( n  m )
 e dj(n  m)
1  j ( n  m)  

2  
j jn
x ( n)  X ( e ) e d

1
j ( n  m)
  
e j( n  m )   
1  m 
 
2 m 
 x ( m ) e  jm jn
e d
 
 
1
 e j( n  m )  e  j ( n  m )
1 m   j ( n  m)
   e d
2 m  
x ( m ) e  jm jn

1
  2 j sin (n  m)
1 m  
j ( n  m)
 
2 m  
x ( m )  e j  ( n  m ) d
 2 sin (n  m)
 0
( n  m)
 1 

j j n
Prove x ( n) 
2  
X ( e ) e d

nn
XX((ee )) 
j
j
 x ( n )e
x ( n ) e j
jnn
nn


1 
2  
j jn
x ( n)  X ( e ) e d

1  m 
 
2 m 
 
 x ( m ) e  jm jn
e d


1 m 
  e d
2 m  
x ( m )

e  jm jn
= x(n)
1 m  
 
2 m  
x ( m )  e j  ( n  m ) d

Notations
Synthesis Inverse Fourier Transform
11 
22 
xx((nn)) XX((eejj))eejjnndd (IFT)

x(n)  F -1[ X (e j )]
Analysis
nn Fourier Transform
XX((eejj)) 
 x ( n )e
x ( n ) e j
jnn (FT)
j
nn
 X (e )  F [ x(n)]
j
xx((nn))

FF
XX((ee ))
j
Real and Imaginary Parts

Fourier Transform (FT)

nn
XX((ee ))  
j
j
 e
x
x[[nn]e
]
j
jnn is a complex-valued function

nn


j j j
X (e )  X R (e )  jX I (e )
Magnitude and Phase
j j j
X (e )  X R (e )  jX I (e )

j X ( e j )
X I (e )
magnitude phase | X ( e j ) |

X ( e j )

j X R ( e j )
j j jX ( e
X (e ) | X (e ) | e )
Discrete-Time Signals and Systems

Symmetry Properties
of Fourier Transform
Conjugate-Symmetric and
Conjugate-Antiymmetric Sequences

 Conjugate-Symmetric Sequence

xxee((nn))  xx ((nn))
**
ee
Called an even sequence
if it is real.

 Conjugate-Antisymmetric Sequence

xxoo((nn))  xx ((nn))

**
oo
Called an odd sequence if
it is real.
Sequence Decomposition
 Any sequence can be expressed as the sum of a conju
gate-symmetric one and a conjugate-antisymmetric on
e, i.e.,

xx((nn))  xxee((nn)) xxoo((nn))

Conjugate Conjugate
Symmetric Antiymmetric

xxee((nn)) 121[[xx((nn))xx**((nn)])] xxoo((nn)) 121[[xx((nn))xx**((nn)])]

2 2
 Function Decomposition
 Any function can be expressed as the sum of a conjug
ate-symmetric one and a conjugate-antisymmetric one,
i.e.,
jj jj jj
XX((ee ))  XXee((ee )) XXoo((ee ))
Conjugate Conjugate
Symmetric Antiymmetric

XXe ((eejj)) 121[[XX((eejj))XX**((eejj)])] XXo ((eejj)) 121[[XX((eejj))XX**((eejj)])]

e 2 o 2
Conjugate-Symmetric and
Conjugate-Antiymmetric Functions

 Conjugate-Symmetric Function

j j Called an even function if

XXee((ee ))  XX ((ee
j **
ee ))
j
it is real.

 Conjugate-Antisymmetric Function

j j
XXoo((ee ))  XX ((ee
j **
oo
j )) Called an odd function if
it is real.
Symmetric Properties
j
xx((nn))

FF
XX((ee ))
j
j xx((nn))

XX((ee
FF j ))
 

 x(n)e
n  
 jn
  x ( n )e
n  
jn
 X (e  j )

magnitude magnitude

   
 phase  phase

 
Symmetric Properties
j
xx((nn))
FF
XX((ee ))
j
j xx**((nn))

XX ((ee
FF ** j ))
*
  
  x ( n )e 
 

 x * (n)e    x ( n ) e j n   X ( e )
 jn jn * *  j

n   n    n  

magnitude magnitude

   
 phase  phase

 
 Symmetric Properties
j ** j
xx((nn))
FF
X ( e
 X (e ) j) xx**((nn))
FF
X (
 X (e ) e j)

magnitude magnitude

   
 phase  phase

 
 Symmetric Properties
j
xx((nn))
FF
X ( e
 X (e ) j)

j Re{x(n)}  12 [ x(n)  x * (n)]

)}
FF
Re{
Re{xx((nn)} XXee((ee )) j
1
2 [ x(n)  x * (n)] 
F j
2 [ X (e )  X ( e
1 *  j
)]

j
)} FF
jjIm{
Im{xx((nn)} XXoo((ee ))j

j Im{x(n)}  12 [ x(n)  x * (n)]

j  j
1
2 [ x(n)  x * (n)] 
F
2 [ X (e )  X (e
1 *
)]
 Symmetric Properties
j
xx((nn))
FF
X ( e
 X (e ) j)

j xe (n)  12 [ x(n)  x * (n)]

xxee((nn))
FF

XXRR((ee ))
j
1
2 [ x( n)  x * ( n)] 
F 1 j * j
2 [ X (e )  X (e )]

j xo ( n)  12 [ x(n)  x * ( n)]
xxoo((nn))
FF

 jX
jXII((ee ))
j
1
2 [ x(n)  x * (n)] 
F 1 j * j
2 [ X (e )  X (e )]
Symmetric Properties for
Real Sequence x(n)
xx((nn))
FF
XX((ee ))
j
j  xx**((nn))
 FF
XX ((ee
** j
j ))

magnitude
Facts:
1. real part is even X R (e j )  X R (e  j )
j  j
  2. Img. part is odd X I (e )   X I (e )
 phase j  j
|
3. Magnitude is even IX ( e ) || X I ( e )|

4. Phase is odd X (e j )  X (e  j )


Discrete-Time Signals and Systems

Fourier Transform
Theorems
Linearity
j j
ax((nn))by
by((nn)) aX((ee ))bY
FF
ax aX bY((ee )) j j

  

 [ ax
n  
( n )  by ( n ) ]e  jn
 a  x
n 1
( n ) e  j n
 b  y
n 1
( n ) e  jn

j j
 aX (e )  bY (e )
Time Shifting  Phase Change
j j
xx((nnnndd))

ee
FF jnndd XX((ee ))
j


F [ x(n  nd )]   x (
n  
n  nd )e  j n


  x (
n  
n )e  j ( n  n d )


 e  jn d  x
n  
( n )e  j n

 j n d j
e X (e )
Frequency Shifting Signal Modulation

j jj((0 ))
ee j0 nn
0 xx((nn))
 FF
XX((ee 0 ))

F [e j0 n x(n)]   e j 0 n

n  
x ( n ) e  j n


  x (
n  
n ) e  j (   0 ) n

 X ( e j (   0 ) )
Time Reversal
j
xx((nn))
 FF
XX((ee ))
j


F [ x(n)]   x (
n  
 n ) e  jn


  x (
n  
n ) e  j (  ) n

 X ( e  j )
Differentiation in Frequency

dd j
nx ( n ) 
nx(n)  j j FF
X ( e
X (e ) j)
dd


F [nx(n)]   nx (
n  
n ) e  jn

1  de  jn
 
 j n  
x ( n)
d
d  d
 j 
d n  
x ( n ) e  jn
 j
d
X ( e jn
)
The Convolution Theorem

yy((nn)) 
 xx((kk)h
)h((nnkk))

FF
 Y
Y ((ee
j
j)  X (e j
)  X ( e j) H (e j
) H ( e j)
)
kk

 
   j( n  k ) 
F [ y (n)]   y ( n )e  j n
  x(k )  h(n)e 
n   k    n   

    j n

   jn 
    x(k )h(n  k ) e   x ( k )e  jk
  h( n)e
 n  


n    k    k  


   jn   X (e j ) H (e j )
  x(k )  h(n  k )e 
k    n   
The Modulation or Window Theorem

11 
 Y (e )    X (e )W (e
j jj j (j 
( ) )) d
yy((nn)) xx((nn))ww((nn))
FF
Y ( e ) 
j X ( e )W ( e )d
22 
 1 
 w(n) x(n)e 
j j (   )
j
Y (e )   j n
 X ( e )W (e )d
n  
2  

1 
w(n)  X (e j )e jn d e  jn
 
 
2 n     
1    j (   ) n 
    w(n) X (e )e
j
d
2  n  
 

1   
 j (   ) n 
  X (e )  w(n)e
j
d
2  
 n   
 Parseval’s Theorem

11 

nn
xx((nn))yy**((nn))
22

XX((ee
j
j)Y **(e j
)Y ( e j) d
) d


Facts: xx((nn))
FF
 X (e j))
X ( e j

yy**((nn))
FF
 YY**((eejj))
11 
 Y (e )  2   X (e )W (e
jj jj j (j
( ) )) d
yy((nn))xx((nn))ww((nn))
FF
Y ( e )  X ( e ) W ( e )d
2  

1  
 x ( n ) y * ( n )e
n  
 j
 
2 n    
X ( e j
)Y ( e  j (   )
) d

Letting =0, then proven.

Parseval’s Theorem
Energy Preserving

11 

nn
| | x(n) |    | X (e ) | d
x ( n 22
) | 
22  
| X ( e j
j) |22d



 

 | x ( n) |
n  
2
  x ( n) x * ( n)
n  

1 

j j
 X ( e ) X *
( e ) d
2  

1 

j 2
 | X ( e ) | d
2  
 Example: Ideal Lowpass Filter
H ( e j ) 1 c

j j n
h( n)  H ( e ) e d
2 c  

1  c jn

2 c 
e d
 1 c jn
  c c  
2 jn c   
e d ( jn)
c
1
j 1 |  | c 
2 jn
e jn
H (e )    c

0 c     sin c n

n
Example: Ideal Lowpass Filter

sin
sin c nn The ideal lowpass fileter
hh((nn)) c nn00,,11,,22,,

nn Is noncausal.

0.6

0.4

0.2

-0.2
-60 -40 -20 0 20 40 60
 Example: Ideal Lowpass Filter

sin
sin c nn TheToideal lowpass fileter
approximate the id
hh((nn)) c nn00,,11,,22,,

nn Is eal
noncausal.
lowpass filter usin
g a window.
0.6

0.4 MM
sin
sin ccnn jjnn
H((ee ))0.2 
j
H j ee
nnMM nn
0

-0.2
-60 -40 -20 0 20 40 60
 Example: Ideal Lowpass Filter
2
M=3
1

-1
-4 -3 -2 -1 0 1 2 3 4
2
M=5
1
MM
sin
sin ccnn jjnn
H((ee ))  
j 0
H j ee
nnMM nn -1
-4 -3 -2 -1 0 1 2 3 4
2
M=19
1

-1
-4 -3 -2 -1 0 1 2 3 4
Discrete-Time Signals and Systems

Existence of Fourier
Transform
Key Issue Does X(ej) exist f
or all ?

Synthesis
11 
xx((nn))  XX((eejj))eejjnndd
22 
We need that |X(ej)
Analysis | <  for all 

XX((eejj)) 
 x ( n )e
x ( n ) e j
jnn

nn

Sufficient Condition for Convergence

 | x(n) | 
n  
j
| X (e ) |  for all 
 
j
| X (e ) |   x ( n )e
n  
 jn
  | x ( n)e
n  
 jn
|

  | x
n  
( n ) || e  j n
|

  | x ( n) |  
n  
More On Convergence
M
 Define X M ( e j )   x (
n M
n ) e  j n

 Uniform Convergence
lim | X (e j )  X M (e j ) | 0
M 

 Mean-Square Convergence

lim 
M   
| X (e j )  X M (e j ) |2  0
Discrete-Time Signals and Systems

Important Transform
Pairs
Fourier Transform Pairs
Sequence Fourier Transform
(n) 1

 ( n  nd ) e  j n d
1
n
a u ( n) (| a | 1)
1  ae  j

1
u (n)  j
  (  2k )
1  ae k  

1
(n  1)a u (n)
n
(1  ae  j ) 2
Fourier Transform Pairs

Sequence Fourier Transform

r n sin  p (n  1) 1
u (n) (| r | 1)
sin  p 1  2r cos  p e  j  r 2 e  2 j

sin c n j 1 |  | c
X (e )  
n 0 c |  | 

1 0  n  M sin[( M  1) / 2]  jM / 2
x ( n)   e
0 otherwise sin( / 2)
Fourier Transform Pairs

Sequence Fourier Transform


e j0 n  2(  
k  
0  2k)


cos(0 n  )   [e j (  0  2k )  e  j (  0  2k )]
k  