The Discrete Fourier Series

Discrete Fourier Series

Given a periodic sequence ~

x[n] with period N so that
~
x[n] = ~
x[n + rN]
The Fourier series representation can be written as
1 ~
~
x[n] = X[k ]e j(2 / N)kn
N k
The Fourier series representation of continuous-time periodic
signals require infinite many complex exponentials
Not that for discrete-time periodic signals we have

e j(2 / N)(k + mN )n = e j(2 / N)kne j(2mn ) = e j(2 / N)kn

Due to the periodicity of the complex exponential we only
need N exponentials for discrete time Fourier series
1 N 1 ~
~
x[n] = X[k ]e j(2 / N)kn
N k =0

Discrete Fourier Series

~
Given a periodic sequence x[n] with period N so that

x[n] = x[n + rN ]
Any periodic signal can be represented by a weighted sum of
complex exponentials. How???

Discrete Fourier Series

Given a periodic sequence ~

x[n] with period N so that
~
x[n] = ~
x[n + rN]
The Fourier series representation can be written as
1 ~
~
x[n] = X[k ]e j(2 / N)kn
N k
The Fourier series representation of continuous-time periodic
signals require infinite many complex exponentials
Not that for discrete-time periodic signals we have

e j(2 / N)(k + mN )n = e j(2 / N)kne j(2mn ) = e j(2 / N)kn

Due to the periodicity of the complex exponential we only
need N exponentials for discrete time Fourier series
1 N 1 ~
~
x[n] = X[k ]e j(2 / N)kn
N k =0

Discrete Fourier Series Pair

A periodic sequence in terms of Fourier series coefficients

1 N1
j 2 /N kn
x[n] = X [ k ] e ( )
N k=0

The Fourier series coefficients can be obtained via

N1

j 2 /N kn
X [ k ] = x[n]e ( )
n=0

Discrete Fourier Series Pair

A periodic sequence in terms of Fourier series coefficients
1 N 1 ~
~
x[n] = X[k ]e j(2 / N)kn
N k =0
The Fourier series coefficients can be obtained via

~
X[k ] =

N 1

~
j(2 / N )kn
x
[
n
]
e

n=0

For convenience we sometimes use

Analysis equation

WN = e j(2 / N)
~
X[k ] =

Synthesis equation

N 1

~
x[n]WNkn

n=0

1 N 1 ~
~
x[n] = X[k ]WNkn
N k =0

Example 1
DFS of a periodic impulse train

1 n = rN
[
]

rN
=

r =
0 else
Since the period of the signal is N
~
x[n] =

~
X[k ] =

N 1

N 1

n=0

n=0

~x[n]e j(2 / N)kn =

[n]e

j(2 / N )kn

= e j(2 / N)k 0 = 1

We can represent the signal with the DFS coefficients as

~
x[n] =

1 N1 j(2 / N)kn
[n rN] = e

N k =0
r =

Example 2
DFS of an periodic rectangular pulse train

The DFS coefficients

j(2 / 10 )k5
4
~
1

e
j(4 k / 10 ) sin(k / 2)
X[k ] = e j(2 / 10 )kn =
=
e
sin(k / 10)
1 e j(2 / 10 )k
n=0

351M Digital Signal Processing

Properties of DFS
Linearity

~
x1 [n]
~
x [n]
2

X1 [k ]
~
DFS

X2 [k ]
~
~
DFS

aX1 [k ] + bX2 [k ]
DFS

a~
x1 [n] + b~
x2 [n]
Shift of a Sequence

~
~
x [n]
DFS

X[k ]
~
DFS
j2 km / N~
[
]
xnm

e
X[k ]
~
e j2 nm / N~
x [n] DFS

X[k m]
Duality

~
~
x [n] DFS

X[k ]
~
X[n] DFS

N~
x [ k ]

Symmetry Properties

Symmetry Properties Contd

Periodic Convolution
Take two periodic sequences
~
~
DFS
x1 [n]
X1 [k ]
~
~
x [n] DFS

X [k ]
2

Lets form the product

~
~ ~
X3 [k] = X1 [k]X2 [k]

The periodic sequence with given DFS can be written as

N 1
~
~
x [n] =
x [m]~
x [n m]
3

m=0

Periodic convolution is commutative

~
x3 [n] =

N 1

~x [m]~x [n m]

m=0

Periodic Convolution Contd

~
x3 [n] =

N 1

~
~
[
]
x
m
1 x2 [n m]

m=0

Substitute periodic convolution into the DFS equation

N 1 N 1
~

X3 [k ] = ~
x1[m]~
x2[n m]WNkn
n = 0 m = 0

Interchange summations

~
X3 [k ] =

N 1

N1 ~
~
kn
x
[
m
]

x
[
n

m
]
W

1
N
2

m=0
n= 0

The inner sum is the DFS of shifted sequence

~
kn
km~
x2[n m]WN = WN X2 [k ]

N 1

Substituting

~
X3 [k ] =

N 1

n=0

N1 ~
~
kn
x
[
m
]

x
[
n

m
]
W

1
N
2
=
m=0
n= 0

~ ~
~
km~
[
]
x
[
m
]
W
X
k
=
X1 [k ]X2 [k ]
1
N
2

N 1

m=0

Graphical Periodic Convolution

The Fourier Transform of Periodic Signals

Periodic sequences are not absolute or square summable
Hence they dont have a Fourier Transform

We can represent them as sums of complex exponentials: DFS

We can combine DFS and Fourier transform
Fourier transform of periodic sequences
Periodic impulse train with values proportional to DFS coefficients

~ j
Xe =

( ) 2N ~X[k ] 2Nk

k =

This is periodic with 2 since DFS is periodic

The inverse transform can be written as

1 2 ~ j jn
1 2 2 ~
2k jn
X e e d =
X[k ]
e d

2 0
2 0 k = N
N

( )

2 k
j
n
1 ~ 2
2k jn
1 N1 ~
N
X[k ]

e d = X[k ]e

0
N k =
N
N k =0

Example
Consider the periodic impulse train

~
p[n] =

[n rN]

r =

The DFS was calculated previously to be

~
P [k ] = 1 for all k
Therefore the Fourier transform is

~ j
Pe =

( ) 2N 2Nk

k =

Relation between Finite-length and Periodic Signals

Consider finite length signal x[n] spanning from 0 to N-1
Convolve with periodic impulse train

~
x[n] = x[n] ~
p[n] = x[n]

r =

r =

[n rN] = x[n rN]

The Fourier transform of the periodic sequence is

~ j
j ~ j
X e = X e P e = X e j

( )

~ j
Xe =

( )

( )( )

( ) 2N 2Nk

k =

2 k
2 j N
2k
X
e

N
k = N

This implies that

~
X[k ] =

j 2Nk
= X e j
X e

( )

2 k
N

DFS coefficients of a periodic signal can be thought as equally

spaced samples of the Fourier transform of one period

Example
Consider the following
sequence

1 0 n 4
x[n] =
else
0

Example
Consider the following
sequence

1 0 n 4
x[n] =
else
0
The Fourier transform

( )

X e j = e j2

sin(5 / 2)
sin( / 2)

The DFS coefficients

~
sin(k / 2)
X[k ] = e j(4 k / 10 )
sin(k / 10)

Example
Given the aperiodic signal

x[n] = n u[n], < 1

For which the periodic signal
is constructed as:

x[n] =

x[n + rN ]
r=

Find the Fourier transform

Find DFS i.e.,

X (e

[n]
of x
X[k]

) of x[n]