10.

The Discrete-Time Fourier

Transform (DTFT)

10.1. Definition of the discrete-time Fourier transform

The Fourier representation of signals plays an important role in both
continuous and discrete signal processing. In this section we consider discrete
signals and develop a Fourier transform for these signals called the discrete-time
Fourier transform, abbreviated DTFT.
The discrete-time Fourier transform of a discrete sequence x (m ) is defined as
follows:

( ) ∑ x(m) e
∞
~ ~
X e jω = − jmω
(10.1)
m = −∞

where ω~ is called the normalized frequency.

( ) ~
The notation X e jω is justified by the observation that the frequency
dependency is in exponential form.
In order for the DTFT of a sequence to exist, the summation in (10.1) must
converge. It will hold if x (m ) is absolutely summable, that is

∑ x(m) < ∞ .
m = −∞
(10.2)

Note that the DTFT of a discrete-time sequence is a function of a continuous

~.
variable ω
222

Since

( ) ∑ x(m) e
∞ ∞
X e j(ω+ 2 π ) = − jm (ω
∑ x(m) e
~ ~ +2π) ~ − jm 2 π
− jmω
= e =
m = −∞ m = −∞

( )
∞

∑ x(m) e
~
− jmω ~
= = X e jω
m = −∞

then the DTFT is periodic in ω~ with a period of 2π .

( )
~
Since X e jω is periodic in ω~ with the period equal to 2π , we can express it
~ . Therefore
by an exponential Fourier series in variable ω

( ) ∑ ~c e
∞ ∞

∑ ~c
~ ~ ~
X e jω = m
jmω
= −m e
− jmω
(10.3)
m = −∞ m = −∞

where

∫ ( )
1 ~ ~ ~
c~− m = X e jω e jmω dω . (10.4)
2π − π

Comparison of equations (10.1) and (10.3) shows that the discrete signal x (m )
~
( )
corresponding to the spectrum X e jω is x (m ) = ~
c . Therefore, from equation −m
(10.4) the inverse discrete-time Fourier transform is

π
x(m ) = ∫ ( )
1 ~ ~ ~
X e j ω e j m ω dω . (10.5)
2π − π

To shed more light on the subject we will derive the DTFT using an
alternative approach.
Let us consider the Fourier transform of a continuous-time signal x(t )

∞
X ( jω) = ∫ x(t ) e
− jωt
dt .
−∞

We approximate the integral as follows:

 223

∞
X ( jω) ≅ ∑ x(mT ) e
m = −∞
s
− jωmTs
Ts (10.6)

and replace the product ωTs by the normalized frequency

~ = ωT = ω .
ω s
fs

If we ignore the scale factor Ts and replace x (mT s ) by x (m ) , the resulting

( ) ~
summation, denoted by X e jω , is the DTFT

( ) ∑ x(m) e
∞
~ ~m
X e jω = − jω
. (10.7)
m = −∞

Example 10.1

Let us consider the signal

x (m ) = a mu (m ) a <1.

The DTFT of this signal is

( ) ∑a ∑ (ae )
∞ ∞
~ ~ ~ m
X e jω = m − jmω
e = − jω
.
m=0 m=0

Since the expression on the right hand side is the geometric series, we obtain

( )
X e jω =
~ 1
~
1 − ae − jω

provided a < 1 .

Example 10.2

Let x (m ) be the unit sample δ(m ) . The DTFT of this signal is

( ) ∑ δ(m)e
∞
~ ~m
X e jω = − jω
= 1.
m = −∞
224

10.2. Some properties of the DTFT

In this section we formulate some properties of the discrete time Fourier
transform.

Periodicity

This property has already been considered and it can be written as follows

( )
X e j(ω + 2 π ) = X e jω .
~
( ) ~
(10.8)

Linearity

The DTFT is a linear operator, i.e. the discrete-time Fourier transform of a signal

x (m ) = a1 x1 (m ) + a2 x2 (m )

is

( )~
( ) ~
X e jω = a1 X 1 e jω + a2 X 2 e jω ( ) ~

( )
where X k e jω is the DTFT of xk (m )
~
(k = 1, 2) .
Shifting

Let us consider a shifted signal

x̂ (m ) = x (m − m0 ) .

The DTFT of this signal is (see (10.7))

( ) ∑ x(m − m ) e
∞ ∞

∑ x(m − m ) e − j( m − m 0 )ω
~ ~ ~ − jm ω~
X̂ e jω = 0
− j mω
= 0 e 0
.
m = −∞ m = −∞

Let k = m − m0 , then

( ) ( )
∞

∑ x(k ) e
~ ~ ~ ~ ~
X̂ e jω = e − jm0 ω − jkω
= e − jm0 ω X e jω .
k = −∞

Thus, we conclude that shifting in time results in the multiplication of the DTFT
~
by a complex exponential e − jm0 ω .
 225

Example 10.3

Let us consider the shifted unit sample x (m ) = δ(m − m0 ) . Using the shifting
property and knowing that the DTFT of δ(m ) is 1, we obtain

( ) ~
X e jω = e − jm0ω .
~

Frequency shifting

Let us consider a signal x (m ) multiplied by e jmω 0

x̂(m ) = x(m ) e jmω0 .

The DTFT of this signal is

( ) ∑ x(m) e ( )
∞ ∞

∑ x(m) e − jm ( ω
~ −ω )
= X e j(ω − ω0 ) .
~ jmω 0 ~ ~
X̂ e jω = e − jmω = 0

m = −∞ m = −∞

Thus, multiplying a sequence by a complex exponential e jmω 0 results in shifting

in frequency of the DTFT.

Convolution theorem

The convolution of signals x (m ) and y (m ) is given by

∞
x(m ) ∗ y (m ) = ∑ x(k ) y (m − k ) .
k = −∞

The DTFT of the convolution is

∞
⎛ ∞ ⎞ − jmω~ ∞
⎛ ∞
~⎞
∑ ∑ ⎜
⎜
m = −∞ ⎝ k = −∞
x (k ) y (m − k ) ⎟e
⎟
⎠
= ∑ ⎜ x (k )
⎜
k = −∞ ⎝
∑
m = −∞
y (m − k ) e − jmω ⎟⎟ =
⎠
∞ ⎛
⎛ ~⎞ ~⎞
∞ ∞ ∞

∑ ∑
⎜ x (k ) ( ) − j( m − k )ω
∑ ⎜ x (k ) e − jkω~
∑ ( ) ⎟=
~ − jkω
= y m − k e e ⎟= y p e − jpω
⎜ ⎟ ⎜ ⎟
k = −∞ ⎝ m = −∞ ⎠ k = −∞ ⎝ p = −∞ ⎠
= X (e )Y (e )
~
jω ~
jω
226

where p = m − k . Thus, the DTFT of a convolution of signals x (m ) and y (m ) is

the product of the DTFTs of x (m ) and y (m ) .

Parseval’s theorem

Let us consider a discrete signal x (m ) . Parseval’s theorem states that

∫ ( )
∞

∑ x (m ) =
1 ~ 2
X e jω ~.
2
dω (10.9)
m = −∞ 2π − π

In Section 12 it will be shown that this equation gives signal energy in the time
domain and in the frequency domain.

10.3. Comparing of the DTFT to the DFT

Recall that the DFT of a sequence {xm = x (m )} where m = 0,1, " , N − 1 is

N −1 N −1 2π
−j
∑x ∑x
mn
− mn
Xn = mw = me
N n = 0,1, " ," N − 1 . (10.10)
m=0 m=0

On the other hand, the DTFT of the same sequence is

( ) ∑x e
N −1
~ ~
X e jω = m
− j mω
. (10.11)
m=0

Comparing (10.10) to (10.11) we find

X n = X e jω( ) ~
~ = 2 πn
ω
n = 0,1, " ," N − 1 . (10.12)
N

Equation (10.12) states that the coefficients of the DFT are samples of the
continuous spectrum given by the DTFT at ω ~ = 2π n .
N
Note that the DFT coefficients correspond to N samples of the X ( z ) taken at N
equally spaced points around the unit circle
 227

2π
j n
z=e N n = 0,1, " , N − 1 .

10.4. Generalized DTFT

Some discrete-time signals do not have a DTFT but they have a generalized
DTFT as explained below.
Let the DTFT of a signal x (m ) be X e jω = δ(ω
~
( )
~ ) . To find this signal, we use
the inverse DTFT:
π
x(m ) = δ(ω
~ ) e jmω~ dω
1 ~= 1 .
∫
2π − π 2π

This result states that the constant signal x(m) =

1
has the DTFT equal to δ(ω
~) .
2π
Hence, the constant signal x (m ) = 1 has the DTFT equal to 2πδ(ω
~ ) , or

x(m ) = 1 ↔ ( )
X e jω = 2πδ(ω
~ ~) . (10.13)

Note that the signal x (m ) = 1 does not have the DTFT in the ordinary sense
because the series
∞

∑e
~
− jm ω

m = −∞

is not convergent. Therefore we say that 2πδ(ω

~ ) is a generalized DTFT of the
signal x (m ) = 1 .
Now we consider a discrete signal x (m ) having the DTFT
( )
X e jω = δ(ω
~ ~ + ω ) . Then
0

π
x(m ) = δ(ω
~ + ω ) e jmω~ dω
1 ~ = 1 e − jmω0
∫
2π − π
0
2π

or

x(m ) = e − jmω0 ↔ ( )
X e jω = 2πδ(ω
~ ~ +ω )
0 (10.14)

holds.
228

Likewise, we find

x(m ) = e jmω0 ↔ ( )
X e jω = 2πδ(ω
~ ~ − ω ).
0 (10.15)

Since

cos(ω0 m + α ) =
1 j( ω 0 m + α ) 1 − j( ω 0 m + α )
e + e =
2 2
1 1
= e jα e jω0 m + e − jα e − jω0 m
2 2

then using (10.14) and (10.15) we determine the DTFT of the signal
x (m ) = cos (ω0 m + α ) as follows:

( )~
X e jω =
2
e 2πδ(ω
1 jα ~ − ω ) + 1 e − jα 2πδ(ω
0
2
~ +ω )=
0

(
= π e jα δ ( ω
~ − ω ) + e − jα δ ( ω
0
~ +ω )
0 )
or

( ) (
x(m ) = cos (ω0 m + α ) ↔ X e jω = π e jα δ(ω
~ ~ − ω ) + e − jα δ(ω
0
~ +ω ) .
0 ) (10.16)

A similar approach leads to the DTFT of the signal x(m ) = sin (ω0 m + α )

x(m ) = sin (ω0 m + α ) =

1 j(ω0 m + α ) 1 − j(ω0 m + α )
e − e =
2j 2j
1 jα jω 0 m 1 − jα − jω 0 m
= e e − e e .
2j 2j

To determine the DTFT of x (m ) , we apply (10.14) and (10.15)

( )~
X e jω =
2j
e 2πδ(ω
1 jα ~ − ω ) − 1 e − jα 2πδ(ω
0
2j
~ + ω )=
0

(
= jπ e − jα δ(ω
~ + ω ) − e jα δ(ω
0
~ −ω )
0 )
or

( ) (
x(m ) = sin (ω0 m + α ) ↔ X e jω = jπ e − jα δ(ω
~ ~ + ω ) − e jα δ(ω
0
~ −ω ) .
0 ) (10.17)
 229

In the special case when α = 0 we have:

x(m ) = cos ω0 m ↔ ( )
X e jω = π(δ(ω
~ ~ + ω ))
~ − ω ) + δ(ω
0 0 (10.18)

x(m ) = sin ω0 m ↔ ( )
X e jω = jπ(δ(ω
~ ~ − ω )) .
~ + ω ) − δ( ω
0 0 (10.19)

10.5. Frequency response of LTI discrete systems

Let an LTI discrete system be represented by its unit sample response h(m )
(see Fig.10.1). The response of the system due to the input x (m ) is given by
convolution

∞
y (m ) = h(m ) ∗ x(m ) = ∑ h(k ) x(m − k ) .
k = −∞

x(m) y(m)
h(m)

Fig.10.1. An LTI discrete system

Convolution theorem states that

( ) ~
( ) ( )
~
Y e jω = H e jω X e jω
~
(10.20)

( ) ( )
where Y e jω is the DTFT of the output y (m ) , X e jω is the DTFT of the input
~ ~

( )
x (m ) and H e jω is called the frequency response function of the discrete
~

system.
( )
~ ~
Note that H e jω is, in general, a complex-valued function of the frequency ω
and can be written in the polar representation

H e jω = H e jω e jφ H (e
( )~
( )~ ~
jω
). (10.21)
230

Thus, the frequency response function of the discrete system is the DTFT of the
unit sample response h(m ) and is a continuous function of ω ~ . Having the

magnitude H e jω ( ) ~
( )
~
and the phase φ H e jω we find:

( ) ~
Y e jω = H e jω ( ) X (e )
~ ~
jω
(10.22)

( ) ~
( ) ~
φY e jω = φ H e jω + φ X e jω . ( ) ~
(10.23)

Example 10.4

Let us consider the input signal

x (m ) = Acos (ω0 m + α )

To find the output response of a system specified by a frequency response

( )
function H e jω , we apply (10.20). The DTFT of the signal x (m ) is given by
~

(10.16) repeated below

( )
~
(
X e jω = πA e jα δ(ω
~ − ω ) + e − j α δ( ω
0
~ +ω ) .
0 )
Hence, we obtain

( )
~ ~
( )
Y e jω = πAH e jω e jα δ (ω ~ − ω ) + πAH e jω~ e − jα δ (ω
0 ( )~ +ω )=
0
jα
( jω 0 ~
= πA e H (e ) δ (ω − ω0 ) + e H (e − jα − jω 0 ~
) δ (ω + ω0 ) = )
= πA H e ( ( )
jω 0
δ (ω − ω0 )e
~ j( φ H + α )
+ He jω 0
δ (ω( )
~ + ω ) e − j( φ H + α ) .
0 )
To find y (m ) , we use the inverse DTFT

∫( ( ) )
π
Aπ
y (m ) = H e jω0 (δ(ω
~ − ω ) e j(φ H + α ) + δ(ω
~ + ω ) e − j(φ H + α ) ) e jmω~ dω
~=
2π − π
0 0

=
A
2
( ) (e (
H e jω 0 )
+e (
j ω0 m + φ H + α − j ω0 m + φ H + α )
)=
= A H (e ) cos (ω m + φ + α ) .
jω 0
0 H