Saravanan Vijayakumaran
sarva@ee.iitb.ac.in
Department of Electrical Engineering
Indian Institute of Technology Bombay
July 25, 2012
1 / 19
Baseband Signals
S(f ) = 0, [f [ > W
−W W
f
[S(f )[
2 / 19
Passband Signals
S(f ) ,= 0, [f ±f
c
[ ≤ W, f
c
> W > 0.
−f
c
−W −f
c
−f
c
+ W f
c
−W f
c
f
c
+ W
f
[S
p
(f )[
3 / 19
Sampling Theorem
Theorem
If a signal s(t ) is bandlimited to B,
S(f ) = 0, [f [ > B
then a sufﬁcient condition for exact recontructability is a uniform
sampling rate f
s
where
f
s
> 2B.
Baseband Signals B = W
Passband Signals B = f
c
+ W (Can we do better?)
4 / 19
Fourier Transform for Real Signals
Im[s(t )] = 0 ⇒ S(f ) = S
∗
(−f ) (Conjugate Symmetry)
⇒ [S(f )[ = [S(−f )[, arg(S(f )) = −arg(S(−f ))
−W W f
1
Re(S(f ))
−W W f
1
1
Im(S(f ))
5 / 19
Fourier Transform of a Real Passband Signal
−f
c
f
c
f
Re(S
p
(f ))
−f
c
f
c
f
Im(S
p
(f ))
6 / 19
Positive Spectrum of a Real Passband Signal
−f
c
f
c
f
Re(S
+
p
(f ))
−f
c
f
c
f
Im(S
+
p
(f ))
S
+
p
(f ) = S
p
(f )u(f )
7 / 19
Complex Envelope of a Real Passband Signal
f
Re(S(f ))
f
Im(S(f ))
S(f ) =
√
2S
+
p
(f + f
c
) =
√
2S
p
(f + f
c
)u(f + f
c
)
8 / 19
Complex Envelope in Time Domain
Frequency Domain Representation
S(f ) =
√
2S
+
p
(f + f
c
) =
√
2S
p
(f + f
c
)u(f + f
c
)
Time Domain Representation of Positive Spectrum
S
+
p
(f ) = S
p
(f )u(f )
s
+
p
(t ) = s
p
(t ) T
−1
[u(f )]
Time Domain Representation of Frequency Domain Unit Step
u(t )
−
−
1
j 2πf
+
1
2
δ(f )
u(f )
−
−
1
−j 2πt
+
1
2
δ(−t )
=
j
2πt
+
1
2
δ(t )
9 / 19
Complex Envelope in Time Domain
Time Domain Representation of Positive Spectrum
s
+
p
(t ) = s
p
(t )
_
1
2
δ(t ) +
j
2πt
_
=
1
2
[s
p
(t ) + j
ˆ
s
p
(t )]
Time Domain Representation of Complex Envelope
√
2S
p
(f )u(f )
−
−
1
√
2
[s
p
(t ) + j
ˆ
s
p
(t )]
√
2S
p
(f + f
c
)u(f + f
c
)
−
−
1
√
2
[s
p
(t ) + j
ˆ
s
p
(t )] e
−j 2πf
c
t
S(f )
−
−
1
√
2
[s
p
(t ) + j
ˆ
s
p
(t )] e
−j 2πf
c
t
s(t ) =
1
√
2
[s
p
(t ) + j
ˆ
s
p
(t )] e
−j 2πf
c
t
10 / 19
Passband Signal in terms of Complex Envelope
Complex Envelope
s(t ) = s
c
(t ) + js
s
(t )
s
c
(t ) Inphase component
s
s
(t ) Quadrature component
Time Domain Relationship
s
p
(t ) = Re
_
√
2s(t )e
j 2πf
c
t
_
= Re
_
√
2¦s
c
(t ) + js
s
(t )¦e
j 2πf
c
t
_
=
√
2s
c
(t ) cos 2πf
c
t −
√
2s
s
(t ) sin 2πf
c
t
Frequency Domain Relationship
S
p
(f ) =
S(f −f
c
) + S
∗
(−f −f
c
)
√
2
11 / 19
Upconversion
s
p
(t ) =
√
2s
c
(t ) cos 2πf
c
t −
√
2s
s
(t ) sin 2πf
c
t
s
c
(t )
s
s
(t )
+
√
2 cos 2πf
c
t
−
√
2 sin2πf
c
t
s
p
(t )
12 / 19
Downconversion
√
2s
p
(t ) cos 2πf
c
t = 2s
c
(t ) cos
2
2πf
c
t −2s
s
(t ) sin 2πf
c
t cos 2πf
c
t
= s
c
(t ) + s
c
(t ) cos 4πf
c
t −s
s
(t ) sin 4πf
c
t
s
c
(t )
s
s
(t )
LPF
LPF
√
2 cos 2πf
c
t
−
√
2 sin2πf
c
t
s
p
(t )
13 / 19
Inner Product and Energy
Let s(t ) and r (t ) be signals.
Deﬁnition (Inner Product)
¸s, r ¸ =
_
∞
−∞
s(t )r
∗
(t ) dt
Deﬁnition (Energy)
E
s
= s
2
= ¸s, s¸ =
_
∞
−∞
[s(t )[
2
dt
14 / 19
I and Q Channels of a Passband Signal
s
p
(t ) =
√
2s
c
(t ) cos 2πf
c
t
. ¸¸ .
I Component
−
√
2s
s
(t ) sin 2πf
c
t
. ¸¸ .
Q Component
x
c
(t ) =
√
2s
c
(t ) cos 2πf
c
t
x
s
(t ) =
√
2s
s
(t ) sin 2πf
c
t
I and Q Channels of a Passband Signal are Orthogonal
¸x
c
, x
s
¸ = 0
15 / 19
Passband and Baseband Inner Products
¸u
p
, v
p
¸ = ¸u
c
, v
c
¸ +¸u
s
, v
s
¸ = Re(¸u, v¸)
Energy of Complex Envelope = Energy of Passband Signal
s
2
= s
p

2
16 / 19
Complex Baseband Equivalent of Passband Filtering
s
p
(t ) Passband signal
h
p
(t ) Impulse response of passband ﬁlter
y
p
(t ) Filter output
y
p
(t ) = s
p
(t ) h
p
(t )
Y
p
(f ) = S
p
(f )H
p
(f )
S
+
(f ) = S
p
(f )u(f )
H
+
(f ) = H
p
(f )u(f )
Y
+
(f ) = Y
p
(f )u(f )
Y
+
(f ) = S
+
(f )H
+
(f )
Y(f ) =
√
2Y
+
(f + f
c
) =
√
2S
+
(f + f
c
)H
+
(f + f
c
) =
1
√
2
S(f )H(f )
17 / 19
Complex Baseband Equivalent of Passband Filtering
y(t ) =
1
√
2
s(t ) h(t )
y
c
=
1
√
2
(s
c
h
c
−s
s
h
s
)
y
s
=
1
√
2
(s
s
h
c
+ s
c
h
s
)
18 / 19
Thanks for your attention
19 / 19