Beruflich Dokumente
Kultur Dokumente
EE 8443
3512 – PatternContinuous
– Signals: Recognition
and Discrete
• Objectives:
Derivation of the DTFT
Transforms of Common Signals
Properties of the DTFT
Lowpass and Highpass Filters
Discrete-Time Fourier Series
• Assume x[n] is a discrete-time periodic signal. We want to represent it as a
weighted-sum of complex exponentials:
xn ck e jk ( 2 / T ) n ck e jk0n , 0 2 / T
k N k N
Note that the notation <N> refers to performing the summation over an N
samples which constitute exactly one period.
• We can derive an expression for the coefficients by using a property of
orthogonal functions (which applies to the complex exponential above):
N , k 0, N , 2 N , ...
e jk ( 2 / N ) n
n N 0, otherwise
jr ( 2 / N ) n
• Multiplying both sides by e and summing over N terms:
xne
n N
jr ( 2 / N ) n
c e
n N k N
k
jk ( 2 / N ) n
e jr ( 2 / N ) n c e
n N k N
k
j ( k r )( 2 / N ) n
• We can show that the second sum on the right equals N if k = r and 0 if k r:
ck
1
xn e jk ( 2 / N ) n
N n N
xn c e k
jk ( 2 / T ) n
c e k
jk0 n
, 0 2 / T (synthesis )
k N k N
xne xne
1 jk ( 2 / N ) n 1 jk0 n
ck (analysis)
N n N N n N
• Define: X e
1
j
N
x[n]e
n
j n
. Note this is periodic in with period 2.
• This implies: c n
N
1
X e jk0 . Note that these are evenly spaced samples of our
j
definition for X e .
EE 3512: Lecture 16, Slide 3
The Discrete-Time Fourier Transform
• Derive an inverse transform:
2
X e e X e e
N N N
x n ( ) X e jk0 e jk0 n
~ 1 1 jk0 jk0 n 1 jk0 n
( ) jk 0
0
k N N 2 k N N 2 k N
2
• As N , ~ x [n] x[n], 0 0, 0 d
N
• This results in our Discrete-Time Fourier Transform:
X e e d
1
x[n] j j n
(synthesis equation)
2 2
x[n]e
j jn
X e (analysis equation)
n
Notes:
• The DTFT and inverse DTFT are not symmetric. One is integration over a
finite interval (2π), and the other is summation over infinite terms.
• The signal, x[n] is aperiodic, and hence, the transform is a continuous
function of frequency. (Recall, periodic signals have a line spectrum.)
• The DTFT is periodic with period 2. Later we will exploit this property to
develop a faster way to compute this transform.
EE 3512: Lecture 16, Slide 4
Example: Unit Pulse and Unit Step
• Unit Pulse:
x[n] [n]
x[n]e
X e j
n
jn
[n]e
n
jn
1
x[n]e
X e j
n
jn
[n n
n
0 ]e jn e jn0
X e e
j
1 jn0
and X e j e jn0 n0
Time delay produces a phase shift linearly proportional to . Note that these
functions are also periodic over the range [-,].
• Unit Step: x[n] u[n]
Since this is not a time-limited function, it has
no DTFT in the ordinary sense. However, it can
X e j
1
j
( )
be shown that the inverse of this function is 1 e
a unit step:
EE 3512: Lecture 16, Slide 5
Example: Exponential Decay
• Consider an exponentially decaying signal:
x[n] a n u[n] a 1
x[n]e
X e j jn
a e n jn
(ae j ) n
n n 0 n 0
1
1 ae j
X e j
1
1 ae j
1
(1 a cos( )) ja sin( )
1
(1 a cos( )) 2 (a sin( )) 2
Note :
(1 a cos( )) 2 (a sin( )) 2 1 2a cos( ) a 2 cos 2 ( ) a 2 sin 2 ( ) 1 2a cos( ) a 2
Hence,
X e j
1
1 2a cos( ) a 2
EE 3512: Lecture 16, Slide 6
The Spectrum of an Exponentially Decaying Signal
X e j
0
1 Lowpass Filter:
1 2a cos(0) a 2
1
(1 2a a 2 )
1
(1 a) 2
1
1 a
X e j
1
1 2a cos( ) a 2 Highpass Filter:
1
(1 2a a 2 )
1
(1 a ) 2
1
1 a
0, n N1
h[n] 1, N1 n N1
0, n N1
sin( ( N1 1 / 2))
X e e
N1 N1
j
n N1
jn
( e j ) n
n N1 sin( / 2)
sin c n
c
1
h[n]
2
c
(1)e jn d
n
• Differentiation in
Frequency:
nx[n] j
d
d
[ X e j ]
1
2
j
d
2
• Parseval’s Relation: x[ n ] X e
n 2 2
• Convolution: y[n] x[n] * h[n] Y (e j ) H (e j ) X (e j )