Waveform Encoding - Digital Communication: Principles and System Modelling - page 48

• Noise transfer function (when X(S) = 0):

1
s +

Y ( s ) =− Y ( s ) × N ( s ) (2.51)

Y ( s )
N ( s ) =

1 s

= (2.52)

1
s

s + 1

1 +

Clearly, it signifies transfer function of HPF.

Transfer function shown in Fig. 2.34 illustrates the modulator's main action. As
the loop integrates the error between the sampled signal and the input signal, it low
pass filters the signal and high pass filters the noise. In other words, the signal is
left unchanged as long as its frequency content doesn't exceed the filter's cutoff fre-
quency, but the S-D loop pushes the noise into a higher frequency band. Grossly
over-sampling the input causes the quantization noise to spread over a wide band-
width and the noise density in the bandwidth of interest (baseband) to significantly
decrease.

Fig. 2.34 Noise performance
of signal delta modulation

Noise response

Signal response

Frequency

2.7 Linear Predictive Coder (LPC)

2.7.1 Concept

From the Sect. 2.3.3 (DPCM prediction), it is understood that, for prediction of the
next sample, only the delayed sample may not reduce the error of prediction to our
threshold level of satisfaction. For that, we have incorporated a series of delayed
samples with suitable scaling factors. The word 'suitable' is important in respect of
the design (Fig. 2.35 )
The accuracy of the prediction is dependent on the proper choice of the scaling
factors. Here the predicted sample is expressible as linear combination of the series
of samples as under: