Scribd
Upload
2K views

Lloyd Max Quantizer (MATLAB Program)

This document is a MATLAB assignment submission that quantizes a signal into different numbers of bits and calculates the mean squared error (MSE) for each resolution over multiple iterations. It contains code that: 1) Sets up arrays and variables to store the quantized signal samples, quantization levels, and MSE values for iterations. 2) Arranges the signal samples into suitable subdivisions based on the number of bits. 3) Calculates the quantization levels and MSE for each iteration by taking averages of the signal samples in each subdivision. 4) Plots the MSE versus the number of iterations and stores the converged MSE values for each resolution.

Uploaded by

Siddhant Rohela
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
2K views

Lloyd Max Quantizer (MATLAB Program)

This document is a MATLAB assignment submission that quantizes a signal into different numbers of bits and calculates the mean squared error (MSE) for each resolution over multiple iterations. It contains code that: 1) Sets up arrays and variables to store the quantized signal samples, quantization levels, and MSE values for iterations. 2) Arranges the signal samples into suitable subdivisions based on the number of bits. 3) Calculates the quantization levels and MSE for each iteration by taking averages of the signal samples in each subdivision. 4) Plots the MSE versus the number of iterations and stores the converged MSE values for each resolution.

Uploaded by

Siddhant Rohela
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

MATLAB ASSIGNMENT 3

DATE OF SUBMISSION: 07-10-16


NAME: SIDDHANT ROHELA

ROLL NUMBER: 14690

n = 2;

%number of bits

g = 1;

%counter for storing the convergent values of MSE

sat = zeros(1, 6);

% array for storing the convergent values of MSE

while n <= 64

%main loop

COURSE: EE320

A = zeros(n, 10000);

%array assigned for each subdivision in -10 to 10

q = zeros(1, n);

%counter to tell where the last non zero entry is in every row of A

L = -10:(20/n):10;

%making the divisions (i.e. their borders)

v = zeros(1, n);

%initializing the values of the array of quantization levels

MSE = zeros(1, 100);

%initializing the values of the array of MSE

plotter = 1:100;

%X-axis for plotting MSE v/s iterations

%////////CODE FOR ARRANGING THE SAMPLES OF THE SIGNAL INTO SUITABLE SUBDIVISIONS////////

for major=1:100

%loop for running iterations; major is the the counter of iterations

q = zeros(1,n);

%resetting the counter 'q' for a new iteration

for i=1:10000

%reading the samples of signal

for p=1:n

%pth subdivision

if ((L(p)<=X(i))&&((X(i)<L(p+1))))

%checking whether the ith sample lies in a given

subdivision
q(p)=q(p)+1;
A(p, q(p))=X(i);

%if yes, then increase the counter 'q' by 1,


%and fill the corresponding element of A with that sample

else
end
end
end

%////////CODE FOR CALCULATING THE QUANTIZATION LEVELS AND THE MEAN SQUARED ERROR////////

MSE(major)=0;

%initializing the MSE for 'major'th iteration

for p=1:n

%pth subdivision

if (q(p)==0)

%if there is no sample of signal within a group then,

v(p)=(L(p)+L(p+1))/2; %assume the level for that group as the average of the L and U limits
else
tempsum=0;
for z=1:q(p)

%intitialize the counter for tempsum

tempsum = tempsum + A(p, z);


end
v(p) = tempsum/q(p);

%optimal pth level is the av. of all samples in the pth sub division

for z=1:q(p)
MSE(major) = MSE(major) + ((v(p)-A(p, z))*(v(p)-A(p, z)));

%building the MSE

end
end
end

MSE(major) = (MSE(major))/10000;

%final value of MSE for that iteration

for y=2:n

%corresponding to the calculated values of quantization levels,

L(y) = (v(y-1)+v(y))/2;

%calculate the new set of subdivisions (i.e. their borders)

end

end

sat(g) = MSE(100);

%filling in the converged value of MSE for every resolution

%////////CODE FOR PLOTTING MSE v/s NUMBER OF ITERATIONS////////

hold on
plot (plotter, MSE)
xlabel('Number of Iterations');
ylabel('Mean Squared Error');
title('MSE v/s Number of Iterations');

end

g = g+1;

%proceeding with the counter 'g'

n = (2*n);

%proceeding with the counter 'n'

You might also like