Discrete-Time Signals and Systems

(3)
 Discrete-Time Signals: Sequences
• Discrete-time signals are represented by sequence of numbers
– The nth number in the sequence is represented with x[n]
• Often times sequences are obtained by sampling of continuous-time signals
– In this case x[n] is value of the analog signal at xa(nT)
– Where T is the sampling period
10

-10
0 20 40 60 80 100 t (ms)
10

-10
0 10 20 30 40 50 n (samples)
Copyright (C) 2005 Güner Arslan 351M Digital Signal Processing 2
 Unit-sample sequence ( )
• The unit-sample sequence plays the same role for discrete-
time signals and systems that the unit impulse function (Dirac
delta function) does for continuous-time signals and systems.
• The unit-sample sequence is often referred to as a discrete-
time impulse, or simple as an impulse.

 [ n] = 
0, n ≠ 0,
[n] =
1, n = 0,

Copyright (C) 2005 Güner Arslan 351M Digital Signal Processing 3

 Basic Sequences and Operations

• Delaying (Shifting) a sequence

y[n] = x[n − no ]

• Unit sample (impulse) sequence


0 n  0
[n] = 
1 n = 0
x[n] =  x[k ] [n − k ]
k = −

• Unit step sequence

0 n  0 n 
u [n ] =  =   [k ] =   [n − k ]
1 n  0 k =− k =0

• Exponential sequences
x[n] = An

In this example 0 < α < 1 and A

Copyright (C) 2005 Güner Arslan is positive
351M Digital Signal Processing 4
 Sinusoidal Sequences
• Important class of sequences
xn = A cos (o n +  )
➢ Sinusoidal sequences

➢ An exponential sequence with complex α and A:  =  e j and A = A e j

o

xn = A n = A e j  e j n = A  e j ( n + )
n o
n o

xn = A  cos (o n +  ) + j A  sin(o n +  )

n n

• x[n] is a sum of weighted sinusoids

• Different from continuous-time, discrete-time sinusoids
– Have ambiguity of 2k in frequency
cos((o + 2k ) n +  ) = cos(o n +  )
– Are not necessary periodic with 2/o.
– In the discrete-time signal, a periodic sequence is one for which:
x[n] = x[n + N], for all n,
2k
cos (o n +  ) = cos (o n + o N +  ) only if N = is an integer
o
Example:
Copyright for ωo=3Güner
(C) 2005 π/4, the 351M
smallest value
Arslan of NDigital
with k Signal Processing
an integer 5
is N=8 (corresponding to k=3). For ωo=1,
there are no integer values of N or k for which N=2πk/ωo
 Discrete-Time Systems
• Discrete-Time Sequence is a mathematical operation that maps a given input
sequence x[n] into an output sequence y[n]

y[n] = T{x[n]} x[n ] T{.} y[n]

• Example Discrete-Time Systems

– Moving (Running) Average:
M2


1
y[n] = x[n − k ]
M 1 + M 2 + 1 k =− M
1

1
= {x[n + M 1 ] + x[n + M 1 − 1] + .......+ x[n] + x[n − 1] + .......+ x[n − M 2 ]}
M1 + M 2 + 1
– Maximum:
y[n] = maxx[n], x[n − 1], x[n − 2]

– Ideal Delay System: y[n] = x[n − nd ]

n is a fixed positive integer
Copyright (C)d 2005 Güner Arslan calledSignal
351M Digital the delay of the system
Processing 6
 Memoryless System
• Memoryless System
– A system is memoryless if the output y[n] at every value of n depends only on the
input x[n] at the same value of n

• Example Memoryless Systems

– Square:

y[n] = (x[n])2
– Sign:

y[n] = signx[n]
• Counterexample
– Ideal Delay System:

y[n] = x[n − nd ]
nd is a fixed positive integer called the delay of the system
Copyright (C) 2005 Güner Arslan 351M Digital Signal Processing 7
 Linear Systems
• Linear System: A system is linear if and only if

T {x1[n] + x2 [n]} = T x1[n]+ T x2 [n] (additivity)

and
T ax[n] = aT x[n] (homogeneity or scaling)
• Examples
– Ideal Delay System: y[n] = x[n − nd ]
T {x1[n] + x2 [n]} = x1[n − nd ] + x2 [n − nd ]
T {x2 [n]} + T x1[n] = x1[n − nd ] + x2 [n − nd ]
= y1[n] + y 2 [n]

T ax[n] = aT x[n] = ax1[n − nd ]

= ay1[n]
Copyright (C) 2005 Güner Arslan 351M Digital Signal Processing 8
 Time-Invariant Systems
• Time-Invariant (shift-invariant) Systems
– A time shift at the input causes corresponding time-shift at output

y[n] = T{x[n]}  y[n − no ] = T x[n − no ]

• Example
– Square
Delay the input the output is y1 n = (x[n − no ])2
y[n] = (x[n])2
Delay the output gives yn - n o  = (x[n − no ])2
• Counterexample
– Compressor System
Delay the input the output is y1 n = x[ Mn − no ]
y[n] = x[Mn] Delay the output gives yn - n o  = xM (n − no )
yn - n o   y1 n

Copyright (C) 2005 Güner Arslan 351M Digital Signal Processing 9

 Causal System
• Causality
– A system is causal if it’s output is a function of only the current and
previous samples

• Examples
– Backward Difference
y[n] = x[n] − x[n − 1]

• Counterexample
– Forward Difference

y[n] = x[n + 1] + x[n]

Copyright (C) 2005 Güner Arslan 351M Digital Signal Processing 10

 Stable System
• Stability (in the sense of bounded-input bounded-output BIBO)
– A system is stable if and only if every bounded input produces a bounded
output
x[n]  Bx    y[n]  B y  
• Example
– Square y[n] = (x[n])2

if input is bounded by x[n]  Bx  

output is bounded by y[n]  Bx2  
• Counterexample
– Log
y[n] = log10 x[n] ( )
even if input is bounded by x[n]  Bx  
output not bounded for x n  = 0  y0 = log10 ( xn ) = −
Copyright (C) 2005 Güner Arslan 351M Digital Signal Processing 11
 Linear-Time Invariant System
• Special importance for their mathematical tractability
• Most signal processing applications involve LTI systems
• LTI system can be completely characterized by their impulse response

[n-k] T{.} hk [n]

• Represent any input

xn =  xk  n − k 
k = −
  
y  n  = T x [n ] = T   x  k    n − k  Linearity
k =− 
 

 x  k T   n − k  =  x k  h n Time − Inv
k =− k =−
k

• From time invariance we arrive at convolution


✓If h[n] is the response
to [n], then the response
Copyright (C) 2005 Güner Arslan
yn =  xk  hn − k  = xk  hk 
351M Digital Signal Processing
k = −
12
to [n - k] is h[n – k].
 Linear-Time Invariant System
impulse response function of the system hk[n]
• We can approximate any complex input discrete signal as if it were simply the sum
of a number of impulses that are scaled copies of one another and shifted in time.
• For shift-invariant linear systems, we can measure the system's response to an
impulse and we will know how to predict the response to any input (combinations of
scaled and shifted copies of the impulse response: [n - k]) through the principle of
superposition.
• To characterize shift-invariant linear systems, then, we need to measure only one
thing: the way the system responds to an impulse of a particular intensity. This
response is called the system impulse response function hk[n].
• It follows that a linear system can be completely characterized by its impulse
response.

Copyright (C) 2005 Güner Arslan 351M Digital Signal Processing 13

 LTI System Example
1 1

impulse 0.5
LTI 0.5

0 0
-5 0 5 -5 0 5
2 2

Scaled
impulse
1
LTI 1

0 0
-5 0 5 -5 0 5
2 2

Time-delayed
impulse
1
LTI 1

0 0
-5 0 5 -5 0 5
2 4

Superposition of
two impulses
1
LTI 2

0 0
-5 0 5 -5 0 5

Copyright (C) 2005 Güner Arslan 351M Digital Signal Processing 14

Dicrete convolution

➢The linear convolution

y(n) of two discrete input
sequences x(n) and h(n) is
defined as the summation
over k of x(k)*h(n-k).

Copyright (C) 2005 Güner Arslan 351M Digital Signal Processing 15

 Properties of LTI Systems
• Convolution is commutative
 
xk  hk  =  xk hn − k  =  hk xn − k  = hk  xk 
k = − k = −

x[n] h[n ] y[n] h[n] x[n] y[n]

• Convolution is distributive
xk  (h1k  + h2 k ) = xk  h1k  + xk  h2 k 

h1[n]
x [n] + y[n] x[n] h1[n]+ h2[n] y[n]
h2[n]

Copyright (C) 2005 Güner Arslan 351M Digital Signal Processing 16

 Properties of LTI Systems
• Cascade connection of LTI systems

x[n] h1[n] h2[n] y[n]

x[n ] h2[n] h1[n] y[n]

x [n] h1[n]h2[n] y[n]

Copyright (C) 2005 Güner Arslan 351M Digital Signal Processing 17

 Stable and Causal LTI Systems
• An LTI system is (BIBO) stable if and only if
– Impulse response is absolute summable


 hk   
k = −
– Let’s write the output of the system as
 
yn =  hk xn − k    hk  xn − k 
k = − k = −

– If the input is bounded

x[n]  Bx
– Then the output is bounded by

yn  Bx  hk 
k = −
– The output is bounded if the absolute sum is finite
• An LTI system is causal if and only if
hk  = 0 for k  0
Copyright (C) 2005 Güner Arslan 351M Digital Signal Processing 18
 Stable and Causal LTI Systems

Causality Condition : Neither necessary nor sufficient

condition for all systems,
h[n] = 0 for n  0 But necessary and sufficient for LTI system

 
y[n] =  x[k ]h[n − k ] =  x[n − k ]h[k ]
k = − k = −
−1 
y[n] =  x[n − k ]h[k ] +  x[n − k ]h[k ]
k = − k =0 But x[n-k] for k>=0 shows
 The future values of x[n].
= 0 +  x[n − k ]h[ k ] So y[n] depends only on the
k =0
Future values of x[n].

so y[ n] =  x[n − k ]h[k ]
k =0

Copyright (C) 2005 Güner Arslan 351M Digital Signal Processing 19

 Linear Constant-Coefficient Difference Equations
• An important class of LTI systems of the form
N M

a
k =0
k yn − k  =  b xn − k 
k =0
k

• The output is not uniquely specified for a given input

– The initial conditions (auxiliary conditions) are required
– Linearity, time invariance, and causality depend on the initial conditions
– If initial conditions are assumed to be zero system is linear, time invariant,
and causal
• Example
– Moving Average
y[n] = x[n] + x[n − 1] + x[n − 2] + x[n − 3]
– Difference Equation Representation
0 3

a
k =0
ky n − k  =  bk xn − k 
k =0
where ak = bk = 1

Copyright (C) 2005 Güner Arslan 351M Digital Signal Processing 20

 Linear Constant-Coefficient Difference Equations
• Linear Constant-Coefficient Difference Equations

input : x[n] output : y[n]

N M

a
k =0
k y[n − k ] =  b x[n − k ]
k =0
k

 x[k ]
Example)
y[n] = : Accumulator
k = −
n −1
y[n − 1] =  x[k ]
k = −
n −1
y[n] = x[n] +  x[k ] = x[n] + y[n − 1]
k = −

y[n] − y[n − 1] = x[n] x[n] is the difference of y[n]

Copyright (C) 2005 Güner Arslan 351M Digital Signal Processing 21
 Linear Constant-Coefficient Difference Equations

22 Digital Signal Processing

Copyright (C) 2005 Güner Arslan 351M Digital Signal Processing 22
 Frequency-Domain Representation of Discrete-Time
Signals and Systems

• Sinusoidal and complex exponential sequences play a particularly

important role in representing discrete-time signals. This is because
complex exponential sequences are eigenfuntions of linear time-
invariant systems, and the response to a sinusoidal input is sinusoidal
with the same frequency as the input and with amplitude and phase
determined by the system. This fundamental property of linear time-
invariant systems makes representations of signals in terms of sinusoids
or complex exponentials (i.e., Fourier representations) very useful in
linear system theory.

Copyright (C) 2005 Güner Arslan 351M Digital Signal Processing 23

 Frequency-Domain Representation of Discrete-Time
Signals and Systems

• Frequency-Domain Representation:
let theinput x[n] = e jn for −   n   ( complex exponentia
l) l
    
y[n] =  h[k ]x[n − k ] =  h[k ]e j ( n − k )
= e jn 
  h[k ]e − jk 

k = − k = −  k = − 
Define

( ) =  h[k ]e 
H e j −j k
e jn : eigenfunct ion of the system
k = − H (e j ) : eigenvalue
 y[n] = H (e  ) e  = H (e  ) x[n]
j j n j

( )
H e j : frequency response of the system, complex number
( ) ( ) ( )
H e j = H e j + jH e j = H e j e jH (e )
R I ( ) j

Copyright (C) 2005 Güner Arslan 351M Digital Signal Processing 24

 Frequency-Domain Representation of Discrete-Time
Signals and Systems
Example)
x[n] = e jn
y [ n ] = x [ n − nd ] : ideal delay
y [ n ] = e j( n − n d
)
= e − jn e jn (when e jn is inputed)
d

( )
 H e j = e − jn d

h [ n ] =  [ n − nd ]

( )  [ n − n ] e = e
 H e j = d
− jn − j n d

n = −

H (e ) = cos (n ), H (e ) = − sin(n )

R
j
d I
j
d

H (e ) = 1 H (e ) = −n
j j
d
Copyright (C) 2005 Güner Arslan 351M Digital Signal Processing 25
 Linear Constant-Coefficient Difference Equations

➢Broad class of signals can be represented as a linear

combination of complex exponentials.

General input :

x[ n] = k
 k e j n k

y[ n] =  (
 k H e j e j  n
k
) k

Copyright (C) 2005 Güner Arslan 351M Digital Signal Processing 26

 Linear Constant-Coefficient Difference Equations
A j j 0 n A − j − j 0 n
x[ n] = A cos( 0 n +  ) = e e + e e
2 2
y[ n] = e H (e )e + e H (e − j 0 )e − j 0 n
A j j 0 j 0 n A − j
2 2
If h[ n] is real  H (e − j 0 ) = H * (e j 0 )

y[ n] =
A
2
 
H (e j 0 )e j e j 0 n + H (e − j 0 )e − j e − j 0 n

= A H (e j 0 ) cos( 0 n +  +  ) where  = H (e j 0 )
For ideal delay H (e j 0 ) = 1 ,  = H (e j 0 ) = − 0 nd
y[ n] = A cos 0 ( n − nd ) +  

Copyright (C) 2005 Güner Arslan 351M Digital Signal Processing 27

 Discrete-Time Fourier Transform
• Many sequences can be expressed as a weighted sum of complex
exponentials as

( ) =  xne

j − jn
Xe (forward transform)
Forward Fourier Transform
n = −
• Where the weighting is determined as

xn =
1 
2 − 
X e j
e jn
d ( ) (inverse
Inverse transform)
Fourier Transform

– X e ( )
j
is the Fourier spectrum of the sequence x[n]
– The phase wraps at 2 hence is not uniquely specified
– The frequency response of a LTI system is the DTFT of the impulse
response
( ) =  hke 1 
( )

He j − jk
and hn =  H e j
e jn
d
k = − 2 − 

Copyright (C) 2005 Güner Arslan 351M Digital Signal Processing 28

 Absolute and Square Summability

• For a given sequence if the infinite sum convergence, the DTFT

exist
( )
X e j   for all 

( ) =  xne
  
Xe j

n = −
− jn
  xn e
n = −
− jn
=  xn  
n = −

• All stable systems are absolute summable and have finite and
continues frequency response

Copyright (C) 2005 Güner Arslan 351M Digital Signal Processing 29

 Absolute and Square Summability
• Absolute summability is sufficient condition for DTFT
• Some sequences may not be absolute summable but only square (spectrum
energy) summable 

 xn
2

n = −

• Such sequences can be represented by Fourier transform if

( ) =  x n e
M

( ) =  xne
 j − j n
j − jn
Xe XM e
n =− M
n = −

lim  X(e ) − X (e )
2
j j
M =0
M→ 
−

• In other words, the error X e −X M e ( )

j
( ) j
may not approach zero at
each value of  as M →  but the total energy in the error does.
Copyright (C) 2005 Güner Arslan 351M Digital Signal Processing 30
 Convolution Theorem
Convolution vs. Multiplication

x1[n]  x2 [n]  X 1[ ]. X 2 [ ]

DFT

x1[n]. x2 [n]  X 1[ ]  X 2 [ ]

DFT

Copyright (C) 2005 Güner Arslan 351M Digital Signal Processing 31

 Proof of Convolution Property

y(t ) =  x( )h(t −  )d
−
• Taking Fourier transforms gives:


Y ( j ) =    x( )h(t −  )d e − jt dt


−
 − 
• Interchanging the order of integration, we have

Y ( j ) =  x( )  h(t −  )e − jt dt d
 

−
 − 
• By the time shift property, the bracketed term is e-jwtH(jw), so



− j 
Y ( j ) = x ( ) e H ( j )d
−

= H ( j )  x( )e − j  d
−

= H ( j ) X ( j )
Copyright (C) 2005 Güner Arslan 351M Digital Signal Processing 32
 x(t) Nyquist–Shannon X( f )

Sampling Theorem
FT
f
t - fmax fma
band-limited function
x
k =  
 f   k 1
p(t) = III   =  [ t − kT ]   ( f − kf )
t 1
p(f) = III   =  f −  =
 f
s
k = − 
T   T T T k = −
k = −

FT
t f
T
-fs fs
x(t). p(t) X ( f )  p( f )
1
FT f max  fs
2

t f
- fmax fma
-fs fs
x
x(t). p(t) X ( f )  p( f ) 1
f max = fs
2
or
FT
f s = 2 f max
t f
Copyright (C) 2005 Güner Arslan 351M Digital Signal Processing - fmax fma 33
-fs x fs
 Example: Ideal Lowpass Filter

• The periodic DTFT of the ideal lowpass filter is


1   c
Hlp e( )j
=
0 c    

• The inverse can be written as

hlp  n  =
1 
2  − 
( )
j j n
H lp e e d  =
1 c j n
2 c  − 
e d

sin c n
=
1
2 jn
j  n c
e  =
−c
1
2 jn
( )
e j c n − e − j c n =
n

• Not causal, Not absolute summable but it has a

DTFT, The DTFT converges in the mean-squared sense
• Role of Gibbs phenomenon
Copyright (C) 2005 Güner Arslan 351M Digital Signal Processing 34
 Example: Ideal Lowpass Filter
Example:
The impulse response is not causal, Not absolutely summable, but squarely
summable. Since sequence values approach zero as n → infinity, but only as 1/n

H M (e ) =  sin n e
M
j c − jn
Let
n=− M n
As M increases, the
oscillatory behavior at ω = ωc
(often referred to as the Gibbs
phenomenon) is more rapid (It
is also known as ripple),
which causes ringing artifacts.

35
Copyright (C) 2005 Güner Arslan 351M Digital Signal Processing 35
 Is absolute summabality a necessary condition?
➢Example: Consider a sequence x[n] whose Fourier transform is the periodic
impulse train

( )  2(  − 
X e j = 0 + 2r )
r = −
Assume that −   0  .


1
x[ n ] = 2(  − 0 )e jn d
2 −

= e j n 0
(Not absoultely summable nor square summable! ! )

➢The concept of Fourier transform representation can be extended to the class of

sequences that can be expressed as a sum of discrete frequency components such as:

If x[n] =  ak e j k n
k

X (e j
) =  2a  ( −  k k + 2r )
r = − k
Copyright (C) 2005 Güner Arslan 351M Digital Signal Processing 36
 Example: Generalized Discrete-time Fourier
transform (DTFT)
• DTFT of x[n] = 1
• Not absolute summable, Not even square summable But we define its DTFT
as a pulse train 
X (e j ) =  2 ( + 2r )
r = −

• Let’s place into inverse DTFT equation

x n  =
2
1 
− 
X ( )
e j
e j n
d

1     j n
=
2 −   
 r =−
2 ( + 2 r )  e d 


=   ( ) e j  n d  = e j 0 n = 1
−

Copyright (C) 2005 Güner Arslan 351M Digital Signal Processing 37

 Symmetric Sequence and Functions

Conjugate-symmetric Conjugate-antisymmetric

Sequence xe n = xe* − n xo n = − xo* − n

xn = xe n + xo n xe n =

1
2
(
xn + x* − n ) xo n =
1
2
(
xn − x* − n )
Function ( )
X e e j = X e* e − j( ) ( ) ( )
X o e j = − X o* e − j

( ) ( )
X e j = X o e j + X e e j ( ) ( )  ( )
Xe e j 1
( )
= X e j + X * e − j
2
( ) 12 X (e  )− X (e  )
X o e j = j * −j

Copyright (C) 2005 Güner Arslan 351M Digital Signal Processing 38

 Properties of DTFT
Sequence x[n] Discrete-Time Fourier Transform X(ej)
x* [n] X*(e-j)
x*[-n] X*(ej)
Re{x[n]} Xe(ej) (conjugate-symmetric part)
jIm{x[n]} Xo(ej) (conjugate-antisymmetric part)
xe[n] XR(ej)= Re{X(ej)}
xo[n] jXI(ej)= jIm{X(ej)}
Any real x[n] X(ej)=X*(e-j) (conjugate symmetric)
Any real x[n] XR(ej)=XR(e-j) (real part is even)
Any real x[n] XI(ej)= -XI(e-j) (imaginary part is odd)
Any real x[n] |X(ej)|=|X(e-j)| (magnitude is even)
Any real x[n] X(ej)= -X(e-j) (phase is odd)
xe[n] XR(ej)
xo[n] jXI(ej)
Copyright (C) 2005 Güner Arslan 351M Digital Signal Processing 39
Example: Illustration of Symmetry Properties
• DTFT of the real sequence x[n]=anu[n]
1
X ( e j ) = − j
if | a | 1
1 − ae
• Some properties are:

( )
X e j =
1
1 − ae − j 
= X *
e(− j
) ⇒ Conjugate-symmetric
1 − a cos 
( )
X R e j =
1 + a 2 − 2a cos 
= X R (e − j
) ⇒ real part is even
−a sin 
( )
X I e j =
1 + a 2 − 2a cos 
= − X I (e − j
) ⇒ imaginary part is odd

( )
X e j =
1
(
= X e − j ) ⇒ magnitude is even
1 + a 2 − 2a cos 
 −a sin  
( )
X e j  = tan −1 
 1 − a cos  
 = − X e(− j
) ⇒ phase is odd
Copyright (C) 2005 Güner Arslan 351M Digital Signal Processing 40
 Fourier Transform Theorems
x[n] X(ej)
y[n] Y(ej)
ax[n]+by[n] aX(ej)+bY(ej)
x[n-nd] d
( )
e − jn X e j

x[-n] X(e-j)
nx[n]
j
( )
dX e j
d

x[n]y[n] X(ej)Y(ej)

x[n]y[n] 1

2 − 
X e j
( )(
Y e j(  −  )
)
d


1
( )


 xn
2 2

2 −
j
Parseval's Theorem: = Xe d
n = −

1
( ) ( )

Parseval's Theorem:  xny n =  X e *
j
Y *
e j
d
n = − 2 − 
Copyright (C) 2005 Güner Arslan 351M Digital Signal Processing 41
 Fourier Transform Pairs

Sequence DTFT
[n-no] e − jn
o

1 
u[n] − j
+  ( + 2k )
1−e k = −

sin(cn) 
1   c
( )
X e j = 
c    
n 0

1 0  n  M sin (M + 1) / 2 − jM / 2
xn =  e
0 otherwise sin( / 2)


e jon  2( − 
k = −
o + 2k )

cos(on+) [e  ( − 
k = −
j
o + 2k ) + e − j  ( + o + 2k )]

Copyright (C) 2005 Güner Arslan 351M Digital Signal Processing 42

Example:
x[n] = a nu[n − 5]
1
 a u[n]  n

1 − ae − j
− j 5
e
 a n−5u[n − 5] 
1 − ae − j
5 − j 5
a e
 a u[n − 5] 
n

1 − ae − j

x[n-nd] ⇔ e − jn X e jd

( )
Copyright (C) 2005 Güner Arslan 351M Digital Signal Processing 43
Example:

Determining an inverse Fourier transform

j 1
X (e )=
(1 − ae − j  )(1 − be − j  )

j a /(a − b ) b /(a − b ) Using partial fraction

X (e )= − j
−
1 − be − j 
expansion.
1 − ae

 a b 
x [n ] = ( )a − (
n
)b n  u [n ]
 a −b a −b 

Copyright (C) 2005 Güner Arslan 351M Digital Signal Processing 44

Example:

Determining the Impulse response from the frequency

response

Copyright (C) 2005 Güner Arslan 351M Digital Signal Processing 45

Example:
Determining the Impulse response for a Difference Equation
1 1
y [n ] − y [n − 1] = x [n ] − x [n − 1]
2 4

To find the impulse response h[n], we set x [n ] =  [n ]

1 1
h [n ] − h [n − 1] =  [n ] −  [n − 1]
2 4
Applying the DTFT to both sides of equation. We obtain
j 1 − j j 1 − j
H (e ) − e H (e ) = 1 − e
2 4
1 − j
1− e
j 4
H (e ) =
1
1− e − j
2
Copyright (C) 2005 Güner Arslan 351M Digital Signal Processing 46