Lecture 6
Signals and Systems
Chapter 5
Prof. Young-Seok Choi
Dept. Electronics and Communications Eng.
Kwangwoon University
1 / 32
Table of Contents
1. review on CTFT
2. Discrete-time Fourier transform (DTFT)
3. examples of DTFT
4. properties of DTFT
2 / 32
Fourier Transform
3 / 32
#FS
s an
periodic x(t)
---
T & No-o
#FT
.
aperiodic e)
X (jc)
Review on Continuous-time Fourier Transform I
in deriving the continuous-time Fourier Transform, we basically have the
following three steps
step 1: pad the aperiodic signal x(t) to construct a periodic replicate
x̃(t)
4 / 32
Review on Continuous-time Fourier Transform II
step 2: Since x̃(t) is periodic, we find the Fourier series coefficient ak
and represent x̃(t) as
1
X
x̃(t) = ak ejk!0 t
k= 1
by defining Z 1
j!t
F{x(t)} = X(j!) = x(t)e dt
1
which is known as the continuous-time Fourier Transform (CTFT), we
showed
1
ak = X(jk!0 )
T
step 3: setting T ! 1, we showed x̃(t) ! x(t) and
Z 1
1 1
F {X(j!)} = x(t) = X(j!)ej!t d!
2⇡ 1
5 / 32
#FS with N
periodic ilins with Ness Ar :
periodic
2π
N-0 wot -o
DTFT
,
Deriving Discrete-time Fourier transform I
the same concept can be applied to discrete-time signals in deriving the
DTFT, we also have three key steps
step 1: Consider an aperiodic discrete-time signal x[n]. We pad x[n] to
construct a periodic signal x̃[n]
6 / 32
Deriving Discrete-time Fourier transform II
step 2: Since x̃[n] is periodic, by discrete-time Fourier series, we have
X
x̃[n] = ak ejk(2⇡/N )n
k=<N >
where ak can be computed as
1 X
jk(2⇡/N )n
ak = x̃[n]e
N
n=<N >
now, note that x̃[n] = x[n] in a period. therefore, it holds that
1 X 1 X
jk(2⇡/N )n jk(2⇡/N )n
ak = x̃[n]e = x[n]e
N N
n=<N > n=<N >
1
1 X jk(2⇡/N )n
= x[n]e
N n= 1
7 / 32
Deriving Discrete-time Fourier transform III
if we define
1
X
X(ej! ) = x[n]e j!n
n= 1
then
1
1 X jk(2⇡/N )n 1 1
ak = x[n]e = X(ejk!0 ) = X(ejk(2⇡/N ) )
N n= 1 N N
: Wo :
8 / 32
Deriving Discrete-time Fourier transform IV
step 3: by combining two equations, we have
X X 1
jk!0 n
x̃[n] = ak e = X(ejk!0 ) ejk!0 n
N
k=<N > k=<N >
1 X jk!0 jk!0 n 2⇡
= X(e )e !0 , (!0 = )
2⇡ N
k=<N >
as N ! 1, !0 ! 0 and x̃[n] ! x[n]
X Z
1 jk!0 jk!0 n 1
x̃[n] = X(e )e !0 ! X(ej! )ej!n d!
2⇡ 2⇡ 2⇡
k=<N >
therefore, Z
1
x[n] = X(ej! )ej!n d!
2⇡ 2⇡
9 / 32
Deriving Discrete-time Fourier transform V
10 / 32
Definition of DTFT
Theorem
The discrete-time Fourier transform X(ej! ) of an aperiodic signal x[n] is
given by
1
X
F{x[n]} = X(ej! ) = x[n]e j!n
(analysis equation)
n= 1
and the discrete-time inverse Fourier transform is given by
Z
1 j! 1
F {X(e )} = x[n] = X(ej! )ej!n d! (synthesis equation)
2⇡ 2⇡
11 / 32
* CTFT
x(t) X(jw)
DTFT
Hins = X(ein)
Periodicity of X(ej! ) I
note that the CTFT X(j!) is aperiodic in general, but the DTFT
X(ej! ) is always periodic.
to see this, let us consider the DTFT (we want to check whether
X(ej! ) = X(ej!+2⇡ ))
1
X
X(ej(!+2⇡) ) = x[n]e j(!+2⇡)n
n= 1
X1
j!n j2⇡ n
= x[n]e e = X(ej! )
n= 1
j2⇡ n
because e = 1n = 1 for all n ! therefore X(ej! ) is periodic with
period 2⇡
12 / 32
Periodicity of X(ej! ) II
13 / 32
Convergence issues with DTFT
We o >
-
1 instiles
1. analysis equation will converge either if x[n] is absolutely summable
1
X
|x[n]| < 1,
n= 1
or if x[n] has finite energy, that is
1
X
|x[n]|2 < 1,
n= 1
2. synthesis equation has no convergence issue since the integral is over
a finite interval of integration
14 / 32
Examples of DTFT I
1. x[n] = an u[n], |a| < 1
Xein)=in
sol)
= Jason
Then ,
Klein))
-
(15) we a
(x() /wer =
15 / 32
Xein) =
Facto =
Facoswojsinc) =
actisin
* It
Wev
acosasin
= encosc + ar
latjbl =b
Xein) =
41 -
k[(racosn) +
Tsimco]
=
sin-tan
O-tant 1- acos Cl
-
yetano
got
......
...
- -
Examples of DTFT II
(
1, |n| N1
1. x[n] =
0, |n| > N1
sol)·
XLein) -jww
NcomTWI s
m = n+N,
JWN
=
NEN
=
16 / 32
DTFT of periodic signals I
F
in continuous-time, ej!0 t ! 2⇡ (! !0 )
let us consider the signal x[n] = ej!0 n
DTFT must be periodic in ! with period 2⇡.
therefore
1
X
F{x[n]} = X(ej! ) = 2⇡ (! !0 2⇡l)
l= 1
CTSIw-wo)
etfIW-Wor(w -wo 4)
000
-
#su 28(w-wo-l 17 / 32
eiron28(0-wo-2π)
Akt iton2f(-ko-2πl)
Periodic
ikwor-
pin] =
k <)
=
Art X2(w-ko- Wo =
2π
D j
=
zarf(w--2πt)
k <N) l
= = a
-
if k =
2N
Xlein) =R -
DTFT of periodic signals II
Now consider a periodic sequence x[n] with period N
X
x[n] = ak ejk(2⇡/N )n
k=<N >
in this case,
1
X ✓ ◆
2⇡k
F{x[n]} = X(ej! ) = 2⇡ak !
N
k= 1
Now consider a periodic sequence x[n] with period N
X
x[n] = ak ejk(2⇡/N )n
k=<N >
=a0 + a1 ej(2⇡/N )n + a2 ej2(2⇡/N )n + · · · + aN 1e
j(N 1)(2⇡/N )n
18 / 32
DTFT of periodic signals III
S(w-e
a .
(w--
a
Na
eTn] Xjc)
19 / 32
DTFT of periodic signals
P1
1. x[n] = k= 1 [n kN ]
soll
Are
For OEM EN-1, TTn] =
SIN]
-jk( .
0
:. Ar
= Sales
= /
= - .
XIew) - Sw
Xein) = -
20 / 32
Properties of the DTFT I
the properties of Fourier Transform is very similar to those of Fourier
Series
I periodicity: X(ej! ) = X(ej(!+2⇡) )
F F
I linearlity: if x1 [n] ! X1 (ej! ) and x2 [n] ! X2 (ej! ), then
F
ax1 [n] + bx2 [n] ! aX1 (ej! ) + bX2 (ej! )
21 / 32
Properties of the DTFT II
F
I time shifting: x[n n0 ] !e j!n0
X(ej! )
F
I frequency shifting: ej!0 n x[n] ! X(ej(! !0 )
)
22 / 32
Properties of the DTFT III
F
I conjugation: x⇤ [n] ! X ⇤ (e j!
)
F
I time reversal: x[ n] ! X(e j!
)
23 / 32
Properties of the DTFT IV
I time expansion
(
x[ nk ], if n is a multiple of k
x(k) [n] =
0, if n is not a multiple of k
D8
↓~
(122(3)
,
263 =
=
1952
x253)
=
=
x22]
x2
24 / 32
Y(xpins] Xnlein) =
Turk
-
=
d) Eckse
r=
La =
x[r]
n= rk
eTR) X(etn)
①
2 xir
> =
I
V= -
Properties of the DTFT V
F
x(k) [n] ! X(ejk! )
periodic with
2π
e
periodic
with
k = 2
k a
periodiwcit
3
=
25 / 32
Properties of the DTFT VI
I di↵erentiation in Frequency
F dX(ej! )
nx[n] !j
d!
I Parseval’s relation
1
X Z
1
|x[n]|2 = |X(ej! )|2 d!
n= 1
2⇡ 2⇡
26 / 32
Properties of the DTFT VII
I convolution:
F
y[n] = h[n] ⇤ x[n] ! Y (ej! ) = H(ej! )X(ej! )
ex) Consider an LTI system with impulse response h[n] = [n n0 ].
The frequency response is
1
X
H(ej! ) = [n n0 ]e j!n
=e j!n0
n= 1
The DTFT of the output y[n] is
Y (ej! ) = e j!n0
X(ej! )
27 / 32
Properties of the DTFT VIII
I multiplication:
Z
F 1 1
x[n]y[n] ! X(ej✓ )Y (ej(! ✓)
)d✓ = X(j!) ⇤ Y (j!)
2⇡ 2⇡ 2⇡
28 / 32
Duality
I duality in DTFS: for two periodic sequences with period N
FS FS 1
x[n] ! ak , f [k] () f [n] ! x[ k]
N
I duality between DTFT and CTFS:
1
X 1
X
(DTFT) X(ej! ) = x[n]e j!n
= x[ n]ej!n
n= 1 n= 1
29 / 32
Summary I
30 / 32
Summary II
31 / 32
Recommended video
For further references on Fourier transform, please see the following
youtube video. I’m sure that it will help a lot.
https://www.youtube.com/watch?v=spUNpyF58BY&t=113s
32 / 32
