Page 294 - Satellite Communications, Fourth Edition

274 Chapter Nine

a fact which is not apparent from the equation for signal-to-noise ratio,
which uses average values of signal and noise power. For example, if a
test tone is used to measure the signal-to-noise ratio in a TV baseband
channel, the result will depend on the position of the test tone within
the baseband, a better result being obtained at lower test tone fre-
quencies. For FDM/FM telephony, the telephone channels at the low end
of the FDM baseband would have better signal-to-noise ratios than
those at the high end.
To equalize the performance over the baseband, a deemphasis network
is introduced after the demodulator to attenuate the high-frequency com-
ponents of noise. Over most of the baseband, the attenuation-frequency
curve of the deemphasis network is the inverse of the rising noise-
frequency characteristic shown in Fig. 9.11b (for practical reasons it is not
feasible to have exact compensation over the complete frequency range).
Thus, after deemphasis, the noise-frequency characteristic is flat, as
shown in Fig. 9.13d. Of course, the deemphasis network also will atten-
uate the signal, and to correct for this, a complementary preemphasis
characteristic is introduced prior to the modulator at the transmitter.
The overall effect is to leave the postdetection signal levels unchanged
while the high-frequency noise is attenuated. The preemphasis, deem-
phasis sequence is illustrated in Fig. 9.13.
The resulting improvement in the signal-to-noise ratio is referred to
variously as preemphasis improvement, deemphasis improvement, or
simply as emphasis improvement. It is usually denoted by P,or [P] deci-
bels, and gives the reduction in the total postdetection noise power.
Preemphasis curves for FDM/FM telephony are given in CCIR
Recommendation 275-2 (1978) and for TV/FM in CCIR Recommendation
405-1 (1982). CCIR values for [P] are 4 dB for the top channel in mul-
tichannel telephony, 13.1 dB for 525-line TV, and 13.0 dB for 625-line
TV. Taking into account the emphasis improvement, Eq. (9.12)
becomes

c S d c C d [G ] [P] (9.13)
P
N N

9.6.6 Noise weighting
Another factor that generally improves the postdetection signal-to-noise
ratio is referred to as noise weighting. This is the way in which the flat-
noise spectrum has to be modified to take into account the frequency
response of the output device and the subjective effect of noise as per-
ceived by the observer. For example, human hearing is less sensitive to
a given noise power density at low and high audio frequencies than at
the middle frequency range.