Scribd
Upload
51 views

Lab Report # 7

This lab report discusses multi-rate signal processing. Task 1 generates a 3 kHz sinusoid sampled at 8 kHz and takes its FFT. Task 2 decimates this signal by 2 and plots the result, noting the sampling frequency no longer meets Nyquist criteria. Task 3 generates two multi-tone signals, decimates them by 2, and explains which frequency components will alias due to insufficient sampling. The signals are plotted throughout to visualize the effects of decimation and aliasing.

Uploaded by

marryam nawaz
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
51 views

Lab Report # 7

This lab report discusses multi-rate signal processing. Task 1 generates a 3 kHz sinusoid sampled at 8 kHz and takes its FFT. Task 2 decimates this signal by 2 and plots the result, noting the sampling frequency no longer meets Nyquist criteria. Task 3 generates two multi-tone signals, decimates them by 2, and explains which frequency components will alias due to insufficient sampling. The signals are plotted throughout to visualize the effects of decimation and aliasing.

Uploaded by

marryam nawaz
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 5

COMSATS INSTITUTE OF INFORMATION

TECHNOLOGY ISLAMABAD

DIGITAL SIGNAL PROCESSING

LAB REPORT # 07
MULTI RATE SIGNAL PROSESSING

COURSE INSTRUCTOR:
SIR MUMTAZ ALI

SUBMITTED TO:
SIR MUBIN SABIR

SUBMITTED BY:
MUHAMMAD SHAHZAD AKHTAR
SP08-BET-062
TASK NO 1:
Use your sinegen function to generate a sinewave with f=3kHz and fs=8kHz.
Calculate its fft with zero frequency component in the middle.Plot it on a properly
scaled w-axis

(a)Code:
fm=3000;
fs=8000;
N=1000;
[y,t]=singen(fm,fs,N);
subplot(211)
plot(t,y), legend('Sinusoid')
y1=fft(y);
w=fs/2*(linspace(-1,1,length(y1)));
subplot(212)
plot(w,fftshift(abs(y1))), legend('fft')
xlim([-4000 4000])

function [y,n]=singen(fm,fs,N)
n=0:1:N-1;
y=sin(2*pi*(fm/fs)*n);

Results:

1
Sinusoid
0.5

-0.5

-1
0 100 200 300 400 500 600 700 800 900 1000

500
fft
400

300

200

100

0
-4000 -3000 -2000 -1000 0 1000 2000 3000 4000
TASK NO 2:
Decimate the signal in task 1 with ‘M = 2’ and plot
the resultant signal. Give your observations about
the plots in your reports
Code:
fm=3000;
fs=8000;
N=1000;
[y,t]=singen(fm,fs,N);
subplot(311)
plot(t,y), legend('Sinusoid')
y1=fft(y,512);
w=fs/2*(linspace(-1,1,512));
subplot(312)
plot(w,fftshift(abs(y1))), legend('fft')
R=2;
y2= decimate(y,R);
y3=fft(y2,512);
subplot(313)
w=fs/2*(linspace(-1,1,512));
plot(w,abs(fftshift(y3))),legend('decimated fft')

function [y,n]=singen(fm,fs,N)
n=0:N-1;
y=sin(2*pi*(fm/fs)*n);

Results:
Decimated graph is not correct because the sampling frequency after decimation not
fulfilling the nyqusite criteria.

1
Sinusoid
0

-1
0 100 200 300 400 500 600 700 800 900 1000
300
fft
200

100

0
-4000 -3000 -2000 -1000 0 1000 2000 3000 4000
1.5
decimated fft
1

0.5
-4000 -3000 -2000 -1000 0 1000 2000 3000 4000
TASK NO 3:
Repeat task 2 for following signals:
a) x1[n] = cos(2*pi*f0/fs*n)+cos(2*pi*f1/fs*n)
• f0 = 2kHz; f1 = 6kHz; fs = 10kHz;
b) x2[n] = cos(2*pi*f0/fs*n)+cos(2*pi*f1/fs*n)+ cos(2*pi*f1/fs*n)
• f0 = 2kHz; f1 = 4kHz; f2 = 3kHz; fs = 10kHz;
Explain which components will be aliased & why?

Code:
f0=2e3;
f1=6000;
fs=10e3;
N=1000;
n=0:N-1;
R=2;
subplot(611)
x1=cos(2*pi*f0/fs*n)+cos(2*pi*f1/fs*n);
plot(n,x1), legend('x1')
x2=fft(x1);
subplot(612)
w=fs/2*linspace(-1,1,length(x2));
plot(w,fftshift(abs(x2))),legend('x1(w)')

x3=decimate(x1,R);
x4=fft(x3);
subplot(613)
w=fs/2*linspace(-1,1,length(x4));
plot(w,fftshift(abs(x4))),legend('deci x1(w)')
f0=2e3;
f1=4e3;
f2=3e3;
fs=10e3;
subplot(614)
y=cos(2*pi*f0/fs*n)+cos(2*pi*f1/fs*n)+ cos(2*pi*f2/fs*n);
plot(n,y), legend('x2')
y1=fft(y);
subplot(615)
y2=decimate(y,R);
y3=fft(y2);
w=fs/2*linspace(-1,1,length(y1));
plot(w,fftshift(abs(y1))),legend('x2(w)')
xlim([-5000 5000])
subplot(616)
w=fs/2*linspace(-1,1,length(y3));
plot(w,fftshift(abs(y3))),legend('deci x2(w)')
xlim([-9000 9000])

Results:

2
0 x1
-2
0 100 200 300 400 500 600 700 800 900 1000
1000
500 x1(w)
0
-5000 -4000 -3000 -2000 -1000 0 1000 2000 3000 4000 5000
400
200 deci x1(w)
0
-5000 -4000 -3000 -2000 -1000 0 1000 2000 3000 4000 5000
5
0 x2
-5
0 100 200 300 400 500 600 700 800 900 1000
1000
500 x2(w)
0
-5000 -4000 -3000 -2000 -1000 0 1000 2000 3000 4000 5000
400
200 deci x2(w)
0
-8000 -6000 -4000 -2000 0 2000 4000 6000 8000

Comments for Part (a) :

As the frequencies are 2k,6k and fs=10k. So by applying Nyquiste criteria we get to
know that the aliasing will be happened when sampling it, that’s why the 6k frequency
signal is not sampled correctly.

When we decimate it by factor of 2.

Fs’=fs/2=5k is the new sampling frequency so after sampling the fm=2k will decimate
correctly and the other will not.

Comments for Part (b) :

As the frequencies are 2k,3k,4k and fs=10k. So its fulfill the Nyquiste criteria, So all the
frequencies will sampled correctly.

When we decimate it by factor of 2.

Fs’=fs/2=5k is the new sampling frequency so after sampling the fm=2k will decimate
correctly and the other will not.

You might also like