Digital Signal Processing Reference
In-Depth Information
PARSEVALs Theorem
For the derivation of the WIENER- KHINTSHINE theorem the following context is
required:
Assuming that the FT and IFT exist for the signals
³
³
FT
jt
ω
FT
jt
ω
x
( )
t
⎯⎯→
X
(
ω
)
=
xt
( )
e t
bzw.
xt
( )
⎯⎯→
X
(
ω
)
=
xt
( )
e t
1
1
1
2
2
2
−∞
−∞
1
1
IFT
³
jt
ω
IFT
³
jt
ω
X
()
ω
⎯⎯→
xt
()
=
X
()
ω
e d
ω
bzw.
X
()
ω
⎯⎯→
x t
()
=
X
()
ω
e dt
1
1
1
2
2
2
2
π
2
π
−∞
−∞
PARSEVAL´s theorem holds good
1
³
³
x
()
txtdt
()
=
X
( )
ωωω
X
( )
d
1
2
1
2
2
π
−∞
−∞
Derivation:
ª
º
1
³
³
³
jt
ω
x
()
txtdt
()
=
xt
()
X
( )
ωω
e d dt
«
»
1
2
1
2
2
π
¬
¼
−∞
−∞
−∞
ª
º
1
³
³
=
X
()
ω
xte
()
jt
ω
dtd
ω
«
»
2
1
2
π
¬
¼
−∞
−∞
1
³
=
XXd
()
ωωω
()
1
2
2
π
−∞
Derivation of the WIENER-KHINTSHINE- theorem
T
1
{
}
³
φ
=
xx
D
=
lim
x x
( )
τ
(
τ
td
)
τ
vow IFT
x x
( )
τ
(
τ
t
)
x x
1
1
1
1
1
1
2
T
11
T
→∞
T
11
{
}
j
ω
t
³
jt
ω
=
lim
XXe
(
ωωω
)
(
)
d
(time shift: IFT
x
(
τ
t
)
=
Xe
(
ω
)
)
1
1
1
1
22
T
π
T
→∞
−∞
1
1
2
³
jt
ω
=
lim
Xe
(
ω
)
d
ω
1
2
π
2
T
T
→∞
−∞
1
1
2
1 () = S()
³
jt
ω
=
S(
ω
)
ed
ω
, if lim
X
ω
ω
2
π
2
T
T
→∞
−∞
Derivation:
The average spectral noise density may be expressed as
an average power per frequency
dP
dP
S (f)=
bzw. S (
ω ωπ ω
)= 2
f
df
d
Search WWH ::




Custom Search