Digital Signal Processing Reference
In-Depth Information
BER
crossover value:
-
3 dB
10
-1
10
-1
10
−2
-∞dB
−5
−3
−1
E
b
/N
tot
10
−3
0dB
3dB
white noise only:
ISSR has no effect.
10
−4
6dB
9dB
10
−5
sweep over different
N
c
/N
0
ratios:
-∞, 0, 3, 6 and 9dB.
−4
−2
0
2
4
E
b
/N
tot
[dB]
Figure 3.9.
Simulated performance of the issr decoder for different colored
noise (N
c
) versus white noise (N
0
) ratios. If only white noise is
present in the signal, issr offers no advantages.
N
c
/N
0
as the colored noise versus white noise ratio. The overall signal quality
E
b
/N
tot
is then expressed as (3.3):
E
b
N
tot
=
E
b
/
N
0
(3.3)
1
+
N
c
/
N
0
The ber characteristics for various N
c
/N
0
values are plotted in Figure 3.9. As
might be expected, the performance of the issr decoder drops when the noise
floor of the channel is being increased. But what is of more importance is the
location of the crossover point between the curves of issr decoding and hard-
decision demapping (topmost curve in Figure 3.9). Most error coding schemes
exhibit some sort of
coding threshold
below which the performance of the de-
coder actually becomes worse than the error rate of the uncoded system. It
turns out that, for all simulated ratios of N
c
/N
0
, the crossover point of the issr
decoder is located at E
b
/
N
tot
3 dB. Interestingly enough, this is exactly
the point where the snr of the qpsk signal intersects the 0 dB boundary.
2
Below this value, the amount of noise energy exceeds the energy of useful in-
formation as a result of which the issr decoder tries to correct the input signal
=−
2
snr
=
E
b
/
N
0
·
η
max
,where
η
max
=
2 is the maximum achievable bandwidth efficiency for qpsk modu-
lation using a raised cosine-rolloff filter with rolloff factor r
=
0
.
0 ([Cou97] p. 351).
