Digital Signal Processing Reference
In-Depth Information
When
σ
0
=
0,
the real and imaginary parts of this complex exponential
sequence are
|
A
|
cos
(ω
0
n
+
φ)
and
|
A
|
sin
(ω
0
n
+
φ)
, respectively, and are real
sinusoidal sequences with an amplitude equal to
|
A
|
.When
σ
0
>
0, the two
sequences increase as
n
→∞
and decrease when
σ
0
<
0as
n
→∞
.When
e
σ
0
n
.
ω
0
=
φ
=
0, the sequence reduces to the real exponential sequence
|
A
|
1.3.7 Properties of cos(
ω
0
n
)
e
jω
0
n
When
A
j
sin
(ω
0
n)
.
This function has some interesting properties, when compared with the
continuous-time function
e
jω
0
t
and they are described below.
First we point out that
ω
0
in
x(n)
=
1, and
σ
0
=
φ
=
0, we get
x(n)
=
=
cos
(ω
0
n)
+
e
jω
0
n
=
1
/T
,where
f
s
is the sampling frequency in hertz and
T
is the sampling period
in seconds, specifically,
ω
0
=
2
πf
0
/f
s
=
is a frequency normalized by
f
s
ω
0
T
,where
ω
0
=2
πf
0
is the actual real
frequency in radians per second and
f
0
is the actual frequency in hertz. Therefore
the unit of the normalized frequency
ω
0
is radians. It is common practice in
the literature on discrete-time systems to choose
ω
as the normalized frequency
variable, and we follow that notation in the following chapters; here we denote
ω
0
as a constant in radians. We will discuss this normalized frequency again in
a later chapter.
=
e
jω
0
n
, two frequen-
cies separated by an integer multiple of 2
π
are indistinguishable from each
other. In other words, it is easily seen that
e
jω
0
n
Property 1.1
In the complex exponential function
x(n)
=
e
j(ω
0
n
+
2
πr)
. The real part and
=
e
jω
0
n
, which are sinusoidal functions,
also exhibit this property. As an example, we have plotted
x
1
(n)
the imaginary part of the function
x(n)
=
=
cos
(
0
.
3
πn)
and
x
2
(n)
=
cos
(
0
.
3
π
+
4
π)n
in Figure 1.8. In contrast, we know that
two
e
jω
2
t
or their real and imag-
inary parts are different if
ω
1
and
ω
2
are different. They are different even if they
are separated by integer multiples of 2
π
. From the property
e
jω
0
n
e
jω
1
t
and
x
2
(t)
continuous-time functions
x
1
(t)
=
=
e
j(ω
0
n
+
2
πr)
above, we arrive at another important result, namely, that the output of a discrete-
time system has the same value when these two functions are excited by the
complex exponential functions
e
jω
0
n
or
e
j(ω
0
n
+
2
πr)
. We will show in Chapter 3
that this is true for all frequencies separated by integer multiples of 2
π
,and
therefore the frequency response of a DT system is periodic in
ω
.
=
Property 1.2
Another important property of the sequence
e
jω
0
n
is that it is
periodic in
n
. A discrete-time function
x(n)
is defined to be periodic if there
exists an integer
N
such that
x(n
x(n)
,where
r
is any arbitrary integer
and
N
is the period of the periodic sequence. To find the value for
N
such that
e
jω
0
n
+
rN)
=
is periodic, we equate
e
jω
0
n
to
e
jω
0
(n
+
rN)
. Therefore
e
jω
0
n
e
jω
0
n
e
jω
0
rN
,
=
which condition is satisfied when
e
jω
0
rN
=
1, that is, when
ω
0
N
=
2
πK
,where
Search WWH ::
Custom Search