Digital Signal Processing Reference
In-Depth Information
50
0
-50
-100
-150
0
500
1000
1500
2000
2500
3000
3500
4000
Frequency (Hz)
0
-2000
-4000
-6000
-8000
0
500
1000
1500
2000
2500
3000
3500
4000
Frequency (Hz)
FIGURE 12.33
Frequency responses of the designed filter.
We implement the first-order SDMsystemusing the same continuous signal in Equation (12.22) . The
continuous signal is originally sampledat 1kHz andeach sample is encodedusing3bits. TheSDMsystem
uses an oversampling rate of 8 kHz ( L ¼ 8) and each sample is quantized using a 3-bit code. The anti-
aliasing lowpass filter is designedwith a cutoff frequency of U ¼ 2 pf max T ¼ 2 p 500 = 8 ; 000 ¼ p= 8
radians. Figure 12.33 shows the frequency responses of the designed filter while Figure 12.34
compares the time and frequency domain signals, where xðtÞ designates the continuous version, x q ðnÞ
denotes the quantized version using a regular sampling rate ( L ¼ 1), and y q ðnÞ is the enhanced version
using L ¼ 8. The detailed amplitude comparisons are given in Figure 12.35 . The measured SNRs are
14.3 dB in the regular sampling system and 33.83 dB in the oversampling SDM system. We observe
a significant SNR improvement of 19.5 dB. The detailed implementation using MATLAB is given in
Program 12.8.
Program 12.8. First-order SDM oversampling implementation.
clear all; close all;clc
ntotal
¼
512; % Number of samples
n
¼
0:ntotal;
L
¼
8; % Oversampling factor
nL
¼
0:ntotal*L;numb
¼
3;A
¼
2^(numb-1)-1; % Peak value
f1
A*0.2;% Frequencies and amplitudes
fmax ¼ 500;fs ¼ 1000; T ¼ 1/fs; % Sampling rate and sampling period
fsL ¼ L*fs;TL ¼ 1/fsL; % Oversampling rate and oversampling period
¼
150;C1
¼
0.5*A;f2
¼
175;C2
¼
A*0.3;f3
¼
200;C3
¼
 
Search WWH ::




Custom Search