- Definition 1: Distribution
maps a function to a
number
- Definition 2: Impulse Distribution
is given by
δst≡s0
δ
s
t
s
0
(1)
Often we write
δ
δ as
δt
δ
t
and use notation
∫−∞∞δtstd
t
=s0
t
δ
t
s
t
s
0
(2)
Special case:
∫−∞∞δtd
t
=1
t
δ
t
1
. Why?
To a broader family of distributions
δ
t
o
st≡s
t
o
δ
t
o
s
t
s
t
o
Alternate notation:
∫−∞∞δt−
t
o
stdt
=s
t
o
t
δ
t
t
o
s
t
s
t
o
We can think of
δt
δ
t
as the limit of tall narrow pulses of area 1.
δt=limit
ε
ε→0
p
ε
t
δ
t
ε
ε
0
p
ε
t
But of course this limit does not exist
at
(t=o)⇒δt
t
o
δ
t
is not a function.
Find the DTFT of
∀n:xn=1
n
x
n
1
since
xn
x
n
is not absolutely
summable, we might expect something weird for an answer.
Xeiω=∑
r
=−∞∞2πδω+2πr
X
ω
r
2
δ
ω
2
r
(3)
Formally use DTFT formula
xn≟∫−ππXeiωeiωn12πdω
=∫−ππ(2π)∑r
=−∞∞δω+2πreiωn12πdω
x
n
≟
ω
X
ω
ω
n
1
2
ω
2
r
δ
ω
2
r
ω
n
1
2
The integral picks up the impulse at
ω=0
ω
0
and evaluates.
=2π2πeiωn|
ω
=00=1=xn
∀n:n
ω
0
0
2
2
ω
n
1
x
n
n
This "comb" of impulses works in the formula.
-
xn
x
n
is pure "DC" therefore it is the purest
possible low frequency signal.
- But why an
infinite sum of
δ
δ distributions?
If
xn=ei
ω
o
n
∀n:n
x
n
ω
o
n
n
"pure frequency"
then
Xeiω=∑
r
=−∞∞δω−
ω
o
+2πr
X
ω
r
δ
ω
ω
o
2
r
Distribution theory thus gives a special
meaning to the statement
∑
n
=−∞∞ei
ω
o
ne(−i)ωn=∑
r
=−∞∞(2π)δω−
ω
o
+2πr
n
ω
o
n
ω
n
r
2
δ
ω
ω
o
2
r
(
ω
o
=0)→(∑
n
=−∞∞e(−i)ωn=∑
r
=−∞∞(2π)δω+2πr)
ω
o
0
n
ω
n
r
2
δ
ω
2
r
"sum of a sinusiod = an impulse combination" (Reference)