Digital Signal Processing Reference
In-Depth Information
Fast Hartley Transform
Whereas complex additions and multiplications are required for an FFT, the Hartley
transform [1-8] requires only real multiplications and additions. The FFT maps a
real function of time into a complex function of frequency, whereas the fast Hartley
transform (FHT) maps the same real-time function into a real function of frequency.
The FHT can be particularly useful in cases where the phase is not a concern.
The discrete Hartley transform (DHT) of a time sequence x ( n ) is defined as
N
-
Â
1
2
p
nk
N
Ê
Ë
ˆ
¯
() =
()
Hk
xn
cas
,
k
=
01
, ,...,
N
-
1
(F.1)
n
=
0
where
cas u
=
cos
u
+
sin
u
(F.2)
In a similar development to the FFT, (F.1) can be decomposed as
(
) -
N
21
N
-
1
2
p
nk
N
2
p
nk
N
Ê
Ë
ˆ
¯ +
Ê
Ë
ˆ
¯
Â
Â
() =
()
()
Hk
xn
cas
xn
cas
(F.3)
n
=
0
nN
=
2
Let n
=
n
+
N /2 in the second summation of (F.3),
Search WWH ::




Custom Search