Transcript Document
Discrete-Time Fourier Transform Properties
Quote of the Day
The profound study of nature is the most fertile
source of mathematical discoveries.
Joseph Fourier
Content and Figures are from Discrete-Time Signal Processing, 2e by Oppenheim, Shafer, and Buck, ©1999-2000 Prentice Hall
Inc.
Absolute and Square Summability
• Absolute summability is sufficient condition for DTFT
• Some sequences may not be absolute summable but only
square summable
xn
2
n
• To represent square summable sequences with DTFT
– We can relax the uniform convergence condition
– Convergence is in mean-squared sense
xne
Xe
j
xne
jn
Xe
lim
M
jn
n
n
j
Xe X e
j
M
j
2
0
– Error does not converge to zero for every value of
– The mean-squared value of the error over all does
Copyright (C) 2005 Güner Arslan
351M Digital Signal Processing
2
Example: Ideal Lowpass Filter
• The periodic DTFT of the ideal lowpass filter is
Hlp e
j
c
1
0 c
• The inverse can be written as
1
1 c jn
j
jn
hlp n
Hlp e e d
e d
c
2
2
sin cn
1
1
jn c
e c
e jcn e jcn
2jn
2jn
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
3
Example: Generalized DTFT
•
•
•
•
DTFT of xn 1
Not absolute summable
Not even square summable
But we define its DTFT as a pulse train
2 2r
Xe
j
r
• Let’s place into inverse DTFT equation
1
j
jn
xn
X
e
e
d
2
jn
1
2
2
r
e d
2 r
Copyright (C) 2005 Güner Arslan
e jnd e j0n 1
351M Digital Signal Processing
4
Symmetric Sequence and Functions
Conjugateantisymmetric
Conjugate-symmetric
xn xe n xo n
x e n
x o n
1
xn x* n
2
Xe ej X*e e j
Xo ej X*o e j
X e 12 Xe X e
X ej Xo ej Xe ej X e e j
Copyright (C) 2005 Güner Arslan
1
xn x* n
2
Function
xo n x*o n
xe n x*e n
Sequence
1
X e j X* e j
2
351M Digital Signal Processing
o
j
j
*
j
5
Symmetry 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
6
Example: Symmetry Properties
• DTFT of the real sequence x[n]=anu[n]
1
X e j
1 ae j
• Some properties are
X e j
XR e j
XI e j
X e j
X e j
Copyright (C) 2005 Güner Arslan
if a 1
1
*
j
X
e
1 ae j
1 a cos
j
X
e
R
1 a2 2a cos
a sin
j
X
e
I
1 a2 2a cos
1
X e j
1 a2 2a cos
a sin
tan1
X e j
1 a cos
351M Digital Signal Processing
7
Fourier Transform Theorems
Sequence
DTFT
x[n]
y[n]
X(ej)
Y(ej)
ax[n]+by[n]
aX(ej)+bY(ej)
x[n-nd]
e jnd X ej
x[-n]
X(e-j)
nx[n]
dX e j
j
d
X(ej)Y(ej)
x[n]y[n]
1
j
j
X
e
Y
e
d
2
x[n]y[n]
2
1
j
Parseval's Theorem: xn
X
e
d
2
n
1
*
j
*
j
Parseval's Theorem: xny n
X
e
Y
e
d
2
n
2
Copyright (C) 2005 Güner Arslan
351M Digital Signal Processing
8
Fourier Transform Pairs
Sequence
DTFT
[n-no]
e jno
2 2k
1
k
1
1 ae j
anu[n] |a|<1
1
2k
j
1e
k
1
c
X e j
0 c
sinM 1 / 2 jM / 2
e
sin / 2
u[n]
sincn
n
1
xn
0
0nM
otherwise
2
e jon
Copyright (C) 2005 Güner Arslan
k
e
cos(on+)
o
k
j
o
2k
2k e j o 2k
351M Digital Signal Processing
9