S-72.

245 Transmission Methods in

Telecommunication Systems (4 cr)
Noise in analog carrier wave (CW) modulation
systems
2 Helsinki University of Technology,Communications Laboratory, Timo O. Korhonen
Noise in analog CW modulation systems
Understanding noise
Lowpass presentation of bandpass noise and its conversion
to baseband noise
Noise statistics of quadrature presentation in rectangular and
polar coordinates
Modeling detectors for linear and exponential modulation
Analysis of post-detection SNR
Synchronous detector
PM-detector
FM-detector
3 Helsinki University of Technology,Communications Laboratory, Timo O. Korhonen
Noise in carrier wave modulation
systems:basic definitions
Objectives: Define bandpass noise and use it to analyze post
detection SNR of analog CW systems
Assume signal is ergodic, e.g., all ensemble averages E[] equal
the corresponding time averages <>. Then, for instance





where the time average is defined by
( ) E[ ( )] v t v t < >=
2 2
( ) E[ ( )] v t v t < >=
( ) ( ) E[ ( ) ( )] v t v t v t v t t t < >=
/ 2
/ 2
1
( ) lim ( )
T
T i i
T
v t v t dt
T

}
< >=
/ 2
/ 2
1
( ) ( )
T
T i i
v t v t dt
T

}
< >= (for a known period)
or
average value
average power
autocorrelation
4 Helsinki University of Technology,Communications Laboratory, Timo O. Korhonen
The system model
We consider normalized ergodic analog message whose
amplitude and power are normalized
2 2
( ) 1, ( ) , 1
x x
x t S x x t S s = =< > s
Channel loss
Modulated signal
Transmitted power
Received power
Received signal (not altered by H
R
)
Detector
2
/
R T c
S S L x = =
Post-detection filter
Pre-detection filter
Pre-detection noise (after H
R
)
5 Helsinki University of Technology,Communications Laboratory, Timo O. Korhonen
Detection models
Pre-detection signal v(t) is presented in quadrature-carrier form:



Detection models:
( ) ( ) cos[ ( )]
( ) cos( ) ( ) cos( )
v c v
i c q c
v t A t t t
v t t v t t
e |
e e
= +
=
( ) ( ) cos ( )
( ) ( )sin ( )
i v
q v
v t A t t
v t A t t
|
|
=

v
( ) Synchronous detector
( ) Envelope detector
( )
( ) Phase detector
( ) / Frequency detector
i
v v
v
v t
A t A
y t
t
d t dt
|
|

0
0
( ) cos[ 2 ( ) ], ) e t
A
}
= + >
t
t C C C
x t A t f x d t t
(Remember that FM was defined by:
6 Helsinki University of Technology,Communications Laboratory, Timo O. Korhonen
Pre-detection noise in bandpass channel
Signal and noise are statistically independent and therefore their
power can be added to form the total pre-detection power:

The pre-detection (bandpass) noise power is filtered from the
channel noise:


2 2 2
= + = +
c R R
v x n S N
2
0
( / 2) ( ) 2 ( / 2)
T
B
R R T
N H f df df B q q q

} }
= = =
from
channel
to detector
7 Helsinki University of Technology,Communications Laboratory, Timo O. Korhonen
Pre-detection SNR
Pre-detection signal-to-noise ratio for bandpass channels is
defined by

Note that above B
T
is the transmission bandwidth passing
channel noise power to the detector

For comparison, we can write the received signal-to-noise in
terms of baseband system (BW = W) SNR defined by

and therefore also

Note that always (limiting case is the SSB with B
T
= W)


(We will see, however, that post detection SNR can be much
larger than !)
/ /( )
R R R T
S N S B q =
2
( / 2) ( ) q q

}
= =
R T R
N B H f df
/( )
R
S W q =
/ /( ) ( / )( / ) /
R R R T R T T
S N S B S W W B W B q q = = =
T
B W > /
R R
S N s

8 Helsinki University of Technology,Communications Laboratory, Timo O. Korhonen

Bandpass noise:
We assume stationary, zero mean Gaussian noise process for
which

Bandpass noise in terms of lowpass equivalent signals


The in-phase and quadrature components
are independent and hence


Their average is zero and their average power is the
same:


2 2
0,
n R
n n N o = = =
( ) ( )cos( ) ( )sin( )
i c q c
n t n t t n t t e e =
2 2 2
i q R
n n n N = = =
( ) ( ) 0
i q
n t n t =
0
i q
n n = =
i
n
q
n
n
A
n
|
( ) ( ) cos[ ( )]
n c n
n t A t t t e | = +
9 Helsinki University of Technology,Communications Laboratory, Timo O. Korhonen
Bandpass noise has Rayleigh distributed
envelope and evenly distributed phase
I-Q components of the bandpass noise can be presented in
envelope - phase format:


The PDF of envelope is Rayleigh
distributed defined by



Therefore mean and variance for the
bandpass noise are (integrate from above, how?)
2 2 2
, arctan
q
n i q n
i
n
A n n
n
| = + =
2 2
( ) exp ( )
2
n
n n
A n n
R R
A A
p A u A
N N
| |
=
|
\ .
2 2
2
2 2 2 2
2
/ 2, 2
/ 2 ( ) / 2
i q
n R n R
R n i q
n n n
A N A N
N A n n n
t
= =
= =
= = + =
i
n
q
n
n
A
n
|
Two independent
r.v.:s - sum of their
variances equals variance
of the envelope
10 Helsinki University of Technology,Communications Laboratory, Timo O. Korhonen
Post detection noise in synchronous detection
Signal component of synchronous detector:







Noise component of synchronous detector:



Detector extracts i-components and removes double frequency
components
( ) ( ) exp( )
( ) ( ) cos( ) ( )sin( )
( )
( ) cos( )
2
( ) ( )
cos(2 ) sin(2 )
2 2
DSB c
DSB c c c c
c
DSB c
c c
c c
v t x t j t
v t A x t t jA x t t
A x t
v t t
A x t A x t
t j t
e
e e
e
e e
=
= +
=
+ +
| |
( )cos( ) cos( ) ( )cos( ) ( )sin( )
cos(2 ) / 2 ( )si ( ) 1 n(2 ) / 2
c c i c q c
c i q c
n t t t n t t n t t
t n t t t n
e e e e
e e
= + (

= + +
( ) ( ) ( )
D c i
y t A x t n t = +
cos( )cos( ) 1/ 2 cos(2 ) / 2
cos( )sin( ) sin(2 ) / 2
= +
=
x x x
x x x
detected message
received DSB signal
11 Helsinki University of Technology,Communications Laboratory, Timo O. Korhonen
Post-detection SNR for DSB
Obtain signal and noise power after detection from:


where average noise and signal power are


Received average signal power is



and therefore SNR after DSB detector is
( ) ( ) ( )
D c i
y t A x t n t = +
2 2 2 2
( ) , ( ) =< > =< >=
D i D c c x
N n t S A x t A S
2 2 2 2 2
0
1/ 2
( )cos ( ) / 2 2 / e =< >= =
x
R c c x x R c
S
S A x t t A S S S A
2
2
/ /
D D c D
c
x
S S N A N
A
=
=
2
2
R
c
S
A
2 (DSB)
2 2
(DSB)
T
R R R
D B W D T
S S S
W N B N

q q
=
= = = =
12 Helsinki University of Technology,Communications Laboratory, Timo O. Korhonen
Comparing SNR for DSB and AM
It can be show, that for AM the post detection SNR is


Comparison of this to the SNR of DSB can done by noting that
in practice


Hence AM performs usually much worse than DSB
It can be shown that for SSB performance is the same as for
DSB, e.g.



/
1
x
D D
AM
x
S
S N
S
=
+
=
~
x
0.5 tone modulation
S 0.1 speech signal
x
S
,
/
R
D D
USB DSB
S
S N
W

q
= =
13 Helsinki University of Technology,Communications Laboratory, Timo O. Korhonen
Exponential modulation and channel noise
Both PM and FM have constant envelopes so the received
power is constant



Received SNR is yielding for wideband FM




where for wideband modulation
2 2
( ) cos[ ( )]
/ 2
c c c
R c c
x t A t t
S x A
e | = +
= =
2 2
2 2
R c c
R R T
A
N N B
S A
q
= =
1,
2( 2) 2
2
m
T m
f W
B f W
DW
|
| |
>> =
~ + ~
=
2 2
R R
R T
R
S S
N B DW D
S
q q
= = =
14 Helsinki University of Technology,Communications Laboratory, Timo O. Korhonen
Detection of exponential modulation
assuming small noise power
( )sin[ ( )]
( ) arctan
( ) cos[ ( )]
n n
v
c n n
A t t
t
A A t t
|
|
|
=
+
carrier+noise
noise
( ) cos[ ( )] ( )cos[ ( )]
c c n c n
Signal Noise
v t A t t A t t t e | e | = + + +
( )sin[ ( )]
( )
n n
v
c
A t t
t
A
|
| ~
small compared to A
c

Detected noise component
15 Helsinki University of Technology,Communications Laboratory, Timo O. Korhonen
Post-detection noise spectra for PM




The channel noise is bandpass noise filtered at the transmission
bandwidth and therefore the respective post-detection noise power
spectral density G
PM
(f) and the total noise power N
D
are
( )sin ( )
( )
( )
2
n n
v
c
q
R
A t t
t
A
n t
S
|
| ~
=
2
2 2
/ 2
sin
R c
q n n
i q R LP
S A
n A
n n N B
|
q

= = =

2
( )
( )
2
2
q
PM
T
R
R T
n t
f
G f
B
S
f
S B
q
| |
| |
= H
|
|
\ .
\ .
| |
= H
|
\ .
Note that after
detection signal
bandwidth is W and
thus a post detection
filter is required to remove
out-of-band channel noise
( ) /
W
W D PM R
N G f df W S q

}
= =
LP
B
16 Helsinki University of Technology,Communications Laboratory, Timo O. Korhonen
Post-detection SNR for FM
Recall the definition of FM-signal

Frequency discriminator (detector) differentiates the
instantaneous phase to cancel out the inherent integration in
phase. Now




Inspection in frequency domain (In order to find the respective
PSDs) yields after detector

and the signal PSD is
( ) 2 ( ) ( ) /
v
t x t d t dt e t | = =
( )
( ) 1 ( ) ( ) 1 ( )
( )
2 2 2
2
q
v S N S
R
Signal Noise
d n t
d t d t d t d t
x t
dt dt dt dt
S dt
| | | |
t t t


(


= = + = +
` `
)
)
( ) 2 ( )
v
X f j f f t = u
2

( ) ( ) ( 2
FM
G f X f t = =
2
2 2
) ( ) / (2 )
v
f f t u
2
2
( )
v
f f = u
| |
0
( ) cos ( ) , ( ) 2 ( )
t
c C C v v
t
x t A t t t f x d e | | t
A
= + =
}
( )
( 2 ) ( )
n
n
n
d x t
j f X f
dt
t
( )
( )sin ( )
( )
2
q
n n
v
c
R
n t
A t t
t
A
S
|
| ~ =
17 Helsinki University of Technology,Communications Laboratory, Timo O. Korhonen
Post-detection SNR in FM (cont.)
Therefore, the post-detection noise PSD can be written as






and now the PSD for FM post detection noise is





and the respective total noise power is
2
( )
2
FM
R T
f f
G f
S B
q | |
= H
|
\ .
( )
3
( ) / 3
W
W D FM R
N G f df W S q

}
= =
2
2
( ) ( )
FM v
G f f f = u
( )
( )
2
q
v
R
N f
f
S
u =
with
18 Helsinki University of Technology,Communications Laboratory, Timo O. Korhonen
For PM we have


For FM we have





Under wideband condition and
Destination S/N for PM and FM
2
2 2 2
/ , where
/
x
D D x x
R
S
S N S S
W S
|
| | t
q
A
A A
= = s
( )
2
2
3
2
/
/ 3
3 3
x
D D
R
R
x x
D
f S
S N
W S
f S
S D S
W W

q
A
A
=
| |
= =
|
\ .
2
3
/
4
T
D D x
B
S N S
W

| |
=
|
\ .
1 D>>
Note that S
D
/N
D
can
be increased just by
increasing deviation!
2 3
( )
2 3
q q

}
= =
W
D FM
W
R R
f W
N df
S S
( )
2
q q

}
= =
W
W D PM
R R
W
N df
S S
1,
2( 2) 2
2 /(2 )
m
T m
f W
T
B f W
DW D B W
|
| |
>> =
~ + ~
= =
19 Helsinki University of Technology,Communications Laboratory, Timo O. Korhonen
20 Helsinki University of Technology,Communications Laboratory, Timo O. Korhonen
FM post-detection S/N with deemphases
Deemphases filter (that is a lowpass filter connected after
detector) can suppress noise further. FM post-detection noise
PSD and total noise power without deemphases:


With deemphases filter (for simplification assume W/B
de
>>1):



where
2
( )
2
q | |
= H
|
\ .
FM
R T
f f
G f
S B
3
( )
3
W
W D FM
R
W
N G f df
S
q

}
= =
3
2
2
/
( ) ( ) arctan /
q
q

<<
}
(
(
= = ~
(
(
de
W
de
W D FM de de R
R de de
W B
B W W
N G f H f df B W S
S B B
2
2
1
( )
1 ( / )
=
+
de
de
H f
f B
2 2
2
2
/
/

q q
A A A
| | | |
= = =
| |
\ . \ .
x x R
D D x
de R de de
f S S S f f
S N S
B W S W B B
21 Helsinki University of Technology,Communications Laboratory, Timo O. Korhonen
Example
FM radio

Without deemphases


With deemphases


Therefore if DSB or SSB system could be exchanged to FM
system 640 fold transmission power savings could be achieved.
Note, however that the required transmission bandwidth is now
about 220 kHz /15 kHz = 15 times larger! Also, a problem is the
FM threshold effect that we discuss next.
75 kHz, 15 kHz, 5, 1/ 2, 2.1kHz
A
= = = = =
x de
f W D S B
2
2
1
2
/ 3
(3 5 ) 38

=
= =
D D x
S N D S
2
/ ( / ) 640
A
= ~
D D de x
S N f B S

q
=
R
S
W
22 Helsinki University of Technology,Communications Laboratory, Timo O. Korhonen
Comparison of carrier wave
modulation systems