Digital Signal Processing Reference
In-Depth Information
The inverse transform is then given by:
where n
¼
0
;
1
;
...
;
N
1
X
N
1
x
ð
n
Þ¼
1
N
jnk
2p
N
X
ð
k
Þ
exp
ð
4
:
27
Þ
k
¼
0
allowing to get the time series x(n) from given frequency samples X(k).
A pair of above equations describes DFT. It allows to pass between time and
frequency domains and has many applications to analysis of signals and systems,
to analysis and synthesis of digital filters, and others, [
1
,
4
,
8
].
Example 4.10 Determine the DFT of the signals: (a) x(n) = 1 for 0 B n B N - 1,
(b) x
ð
n
Þ¼
cos n
ð
2p
=
N
Þ
ð
Þ
for the same range of samples as in example (a).
Solution
(a)
¼
X
ð
k
Þ¼
X
N
1
exp
jnk
2p
N
1
exp
ð
jk2p
Þ
1
exp
ð
jk2p
=
N
Þ
n
¼
0
and thus
X
ð
k
Þ¼
N
for k
¼
0
:
0
for other k
(b)
The considered signal can be decomposed as follows:
¼
0
:
5
þ
exp
jn
2p
N
¼
0
:
5
f
x
1
ð
n
Þþ
x
2
ð
n
Þg:
n
2p
N
jn
2p
N
x
ð
n
Þ¼
cos
exp
Now, calculating the DFT for both exponential components one obtains:
¼
1
exp
½
j
ð
k
1
Þ
2p
1
exp
j
ð
k
1
Þ
2
N
X
1
ð
k
Þ¼
DFT
f
x
1
ð
n
Þg ¼
X
N
1
exp jn
2p
N
ð
k
1
Þ
;
n
¼
0
which yields:
X
1
ð
k
Þ¼
N
for k
¼
1
where 0
k
N
1
0
for other k
and analogously:
X
2
ð
k
Þ¼
1
exp
½
j
ð
k
þ
1
Þ
2p
1
exp
j
ð
k
þ
1
Þ
2
N
:
Search WWH ::
Custom Search