DSP Lab Manual

VISVESVARAYA TECHNOLOGICAL UNIVERSITY

JNANASANGAMA BELAGAVI 590018

KARNATAKA STATE, INDIA

Page | 1

DIGITAL SIGNAL PROCESSING

LAB MANUAL
SUBJECT CODE: 15ECL57
V SEMESTER E&C ENGINEERING

Manual authored by Manual Approved by

ANIL B.V. Dr.Ravikumar M.S.

Associate Professor Professor and Head

DEPT. OF ELECTRONICS AND COMMUNICATION ENGINEERING

K.V.G. COLLEGE OF ENGINEERING SULLIA, D.K. - 574327
KARNATAKA STATE, INDIA

K.V.G. College of Engineering, Sullia D.K.

 DSP Lab Manual

DSP Lab
B.E., V Semester, EC/TC
[As per Choice Based Credit System (CBCS) scheme]
Page | 2
Subject Code:15ECL57 IA Marks 20

Number of Lecture Hours/Week

01Hr Tutorial (Instructions)+02 Hours Laboratory=03

Exam Marks 80 Exam Hours 03

CREDITS 02
Laboratory Experiments
Following Experiments to be done using MATLAB / SCILAB / OCTAVE
or Equivalent:

1. Verification of sampling theorem.

2. Linear and circular convolution of two given sequences, Commutative,
Distributive and Associative property of convolution.
3. Auto and cross correlation of two sequences and verification of their
properties
4. Solving a given difference equation.
5. Computation of N point DFT of a given sequence and to plot
magnitude and phase spectrum (using DFT equation and verify it by
built-in routine).
6. (i) Verification of DFT properties (Linearity and Parsevals theorem,
etc.)(ii) DFT computation of square pulse and sinc function etc.
7. Design and implementation of FIR filter to meet given specifications
(using different window techniques).
8. Design and implementation of IIR filter to meet given specifications.
Following Experiments to be done using DSP kit
9. Linear convolution of two sequences
10. Circular convolution of two sequences
11. N-point DFT of a given sequence
12. Impulse response of first order and second order system
13. Implementation of FIR filter
Conduct of Practical Examination:
1. All laboratory experiments are to be included for practical
examination.
2. Strictly follow the instructions as printed on the cover page of answer
script for breakup of marks.
3. Change of experiment is allowed only once and Marks allotted to the
procedure part to be made zero.

K.V.G. College of Engineering, Sullia D.K.

 DSP Lab Manual

Page | 3

I - CYCLE EXPERIMENTS

Expt. No. Experiment Name

1 Verification of Sampling Theorem

2 a) Linear convolutionof two sequences
i) Causal ii) Noncausal
b) Circular convolution of two sequences
Commutative, Distributive and Associative
properties of convolution.
3 a) Auto correlation of two sequences
b) Cross correlation of two sequences
Verification of the properties of Correlation.
4 a) Solving a difference equation without initial
conditions
b) Solving a difference equation with initial
conditions

K.V.G. College of Engineering, Sullia D.K.

 DSP Lab Manual

Page | 4

PART - A
MATLAB PROGRAMS

K.V.G. College of Engineering, Sullia D.K.

 DSP Lab Manual

EXPT No: 1 DATE:

VERIFICATION OF SAMPLING THEOREM

AIM: To execute MATLAB m-file program for the sampling of a composite
Page | 5 continuous time signal and to reconstruct it from the samples based on Nyquist
Sampling criteria.

THEORY:In signal processing, sampling is the mapping of a continuous-time

continuous-valued signal to a discrete-time continuous-valued signal. A common
example is the conversion of audio signal (a continuous signal) to a sequence of
samples (a discrete-time signal).

Figure 1.1 The sampled continuous time signal

Nyquist Sampling Theorem:

If the highest frequency contained in an analog signal xa(t) is Fmax = B Hz and the
signal is sampled at a rate Fs> 2 Fmax = 2B, then xa(t) can be exactly recovered from
its sample values using the interpolation function
()
() = (1.1)
()

Thus xa(t) may be expressed as

() =
= ()( /) (1.2)

where xa(/) = xa(nT) x[n] are the samples of xa(t). When the sampling of xa(t)
is performed at the minimum sampling rate Fs = 2B, the reconstruction formula

( )
becomes: () =
= ()

(1.3)
( )

The sampling rate FN = 2Fmax = 2BHz is called the Nyquist rate.

K.V.G. College of Engineering, Sullia D.K.

 DSP Lab Manual

Consider a continuous-time signal, with amplitude A and frequency F Hz.

xa(t) = A cos(t) = A cos(2Ft) (1.4)

If Fs is the sampling frequency then Fs= 1/T samples per second. T is the sampling
Page | 6 interval in seconds.
xa[nT]=Acos[nT] (1.5)

= Acos[2FnT] (1.6)

=Acos[2[]n] (1.7)

= Acos[2fn] (1.8)

= Acos[n] [] (1.9)

Comparing eqn1.7 and eqn1.8 it can be seen that f = [ ], and fis called discrete time

frequency with units cycles/sample. 2f = is called discrete-time angular

frequency with units radian/sample; = T describes the relationship between the
D.T. angular frequency and C.T angular frequency(rad/sec)
ALGORIHM:
Begin
Read Amplitude A and Frequency F, t0,t, t1
define t from t0 to t1 in steps of t
x1= A1 cos( 2F1t ); x2 = A2 cos( 2F2t )
x = x1 + x2
choose Fs based on max( F1, F2 )
define n from n0 = (t0/Ts) to n1 = ( t1/Ts) in steps of 1
xn1 = A1cos[ 2F1nT ]
xn2 = A2cos[ 2F2nT ]
xn = xn1 + xn2
()
compute = ()

forn0to n1in steps of 1

2( )
2
= (/)( /) ) = (2)
2( )
2
end
End

K.V.G. College of Engineering, Sullia D.K.

 DSP Lab Manual

PROGRAM: samp_thrm.m

clc %Clear Command Window

clear all % Clear all the variables
close all % Close all Figure Windows
Page | 7
% Enter the Amplitude values

A1 = input(Amplitude of signal x1(t) = ');

A2 = input(Amplitude of signal x2(t) = ');

% Enter the frequency values

F1 = input(Frequency of signal x1(t) = ');

F2 = input(Frequency of signal x2(t) = ');

% Define Beginning and End Time

t0 = input(' Beginning Time = ');

t1 = input( End Time = ');

% Define the Time Resolution

Delta_t= input('Resolution =);

% Define x-axis as Time axis

t = t0: Delta_t : t1;

% Define Analog Signals

x1 = A1 * cos( 2 * pi * F1 * t );
x2 = A2 * cos( 2 * pi * F2 * t );

% Generate the single composite Analog Signal xa(t)

xa = x1 + x2;

% Sum of the Amplitudes for plotting of xa(t)

A = A1 + A2;

% To plot x1(t), x2(t) and xa(t) signals

figure( 1 ) %Clear Command Window

subplot( 3, 1, 1 )
plot( t, x1, 'r' )
axis( [ t0t1 -A1 A1 ] )
ylabel( 'x1(t)' )
title ( ' Continuous Time Signals ' )
subplot( 3, 1, 2 )
plot( t, x2, 'g' )
axis( [ t0 t1 -A2 A2 ] )

K.V.G. College of Engineering, Sullia D.K.

 DSP Lab Manual

ylabel( 'x2(t)' )

subplot( 3, 1, 3 )
plot( t, xa, 'b' )
axis( [ t0 t1 -A A] )
ylabel( 'x3(t)' )
Page | 8
% Choose the Sampling Frequency satisfying Nyquists Sampling Theorem

Fs = input('Enter the Sampling Frequency in Hz = ' );

% Calculate the Sampling Interval

Ts = 1/Fs;

% Calculate the number of Samples

n0 = t0/Ts
n1 = t1/Ts;
n = n0:n1;
nT = n * Ts;
xn1 = A1 * cos( 2 * pi * F1 * nT );
xn2 = A2 * cos( 2 * pi * F2 * nT );

xn = xn1 + xn2;

figure(2)
subplot( 3, 1, 1 )
stem( n, xn1, 'r' )
axis( [ t0/Ts, t1/Ts, -A1, A1 ] )
ylabel( 'x1[n]' )
title ( ' Discrete Time Signals ' )

subplot( 3, 1, 2 )
stem( n, xn2, 'g' )
axis( [ t0/Ts, t1/Ts, -A2, A2 ] )
ylabel( 'x2[n]' )

subplot( 3, 1, 3 )
stem( n, xn, 'b' )
axis( [ t0/Ts, t1/Ts, -A, A ] )
ylabel( 'x [n]' )

% Find the Maximum Frequency for reconstruction

Fmax = max( [ F1 F2 ] );

Theta = zeros( length(nT), length(t) );

tn = nT' * ones( 1, length( t ) );
for row = 1 : length(nT)
Theta( row, : ) = t( 1, : ) - tn( row, : );
end

K.V.G. College of Engineering, Sullia D.K.

 DSP Lab Manual

% Generation of Interpolation Function

Interpolate = sinc( ( 2 * pi * Fmax ) .* Theta );

xr = xn * Interpolate;
Page | 9 figure(3)
subplot( 2, 1, 1 )
plot( t, xa, 'b' )
axis( [ t0, t1, -A, A ] )
xlabel( ' t--> ' )
ylabel(' xa(t) ' )
title ( ' Original Continous Time Signals ' )

subplot( 2, 1, 2 )
plot( t, xr, 'b' )
xlabel( ' t--> ' )
ylabel(' xr(t) ' )
title ( ' Reconstructed Continous Time Signals ' )

>>
Amplitude of signal x1(t) = 1.2
Amplitude of signal x2(t) = 0.8
Frequency in Hz of signal x1(t) = 4
Frequency in Hz of signal x2(t) = 8
Beginning Time = -1
End Time = 1
Resolution =0.0001
Enter the Sampling Frequency in Hz = 64

Figure 1.2Screenshot of the output of Nyquist Sampling Theorem Verification program for
perfect reconstruction.

K.V.G. College of Engineering, Sullia D.K.

 DSP Lab Manual

EXPT No: 2 a) DATE:

LINEAR CONVOLUTION OF TWO SEQUENCES

AIM: To execute MATLAB m-file program for finding the response of a given Linear
Page | 10 Time Invariant system by linear convolution summation with:

(i) Causal sequence

(ii) Non Causal sequence and
(iii) To verify commutative, associative and distributive properties.

THEORY:

For discrete time systems with input signal x(n) and finite duration unit impulse
response h(n), linear convolution summation is defined by

() =
= [][ ] = = [][ ] (3.1)

Eqn 3.1 represents the response of a causal system. For a non-causal system the
response is given by the convolution summation defined by the eqn3.2

y(n) == [][ ] = =
[][ ] (3.2)

If Nx is the length of x[n] and Nh is the length of h[n] then the length of y[n] is Ny
given by
Ny= Nx + Nh - 1 (3.3)

Graphical Computation of Linear Convolution Summation:

Reflection of h(k) resulting in h(-k)

Shifting of h(-k) resulting in h(n-k)
Element wise multiplication of the sequences x(k) and h(n-k)
Summation of the product sequence x(k) h(n-k) results in the convolution
Sample value for y(n)

Example:

x(n) = {1, 2, 3, 1}Nx = 4

h(n) = {1, 1, 1}Nh = 3

Ny =Nx + Nh - 1= 4 + 3 1 = 6

K.V.G. College of Engineering, Sullia D.K.

 DSP Lab Manual

Page | 11

K.V.G. College of Engineering, Sullia D.K.

 DSP Lab Manual

Page | 12

y[n] = { 1, 3, 6, 6, 4, 1 }

ALGORIHM: (i) Convolution - Causal

Begin

Read xn and hn
Compute
yn = xk hn,k , n
length of xn, hn, yn
display xn, hn, yn
End

K.V.G. College of Engineering, Sullia D.K.

 DSP Lab Manual

PROGRAM:(i)conv_causal.m

clc
Page | 13 clear all
close all
% Enter the samples of Input Sequence x[n] and Unit Impulse Response h[n]
x = input('Enter the Discrete Input Signal Sequence x[n] = ');
h = input('Enter the Impulse Response Sequence h[n] = ');

Nx = length( x );
Nh = length( h );
nx = 0 : 1 : Nx-1;
nh = 0 : 1 : Nh-1;

% Convolving Input x[n] and Impulse Response h[n] %

y = conv( h, x );
Ny = length( y );
ny = 0 : 1 : Ny-1;

disp(' The Convolution of x[n] and h[n] is y[n] = ' )

disp( y )
disp(' The number of samples in y[n] is Ny = ' )
disp(Ny )

figure( 1 )
subplot( 3, 1, 1 )
stem(nx, x, 'r', 'filled' )
xlabel(' n--> ' )
ylabel(' x[n] ' )
title(' Input Sequence x[n] ' )

subplot( 3, 1, 2 )
stem(nh, h, 'r', 'filled' )
xlabel(' n--> ' )
ylabel(' h[n] ' )
title(' Impulse Response h[n] ' )

subplot( 3, 1, 3 )
stem(ny, y, 'r', 'filled' )
xlabel(' n--> ' )
ylabel(' y[n] ' )

K.V.G. College of Engineering, Sullia D.K.

 DSP Lab Manual

title(' Output Sequence y[n] ' )

AFTER THE EXECUTION:

Page | 14 >>
Enter the Discrete Input Signal Sequence x[n] = [ 1, 2, 3, 4 ]
Enter the Impulse Response Sequence h[n] = [ 3 1 2 1 ]
The Convolution of x[n] and h[n] is y[n] =
3 7 13 20 12 11 4

The number of samples in y[n] is Ny =

7
>>

Fig 3.1 Screenshot of output of the conv_causal.m program

ALGORIHM: (ii) Convolution - Non Causal

Begin

Read xn,nx, hn andnh

Compute
yn = xkhn,k , n
Min(ny)= Min(nx)+Min(nh)
Max(ny)= Max(nx)+Max(nh)
ny = Min(ny):Max(ny)
displayxn, hn, yn
End

K.V.G. College of Engineering, Sullia D.K.

 DSP Lab Manual

PROGRAM: (ii)conv_noncuasal.m
clc
clear all
close all
Page | 15
% Enter the values of x[n] and h[n]
x = input(' Enter the input signal x[n] = ' );
nx = input( ' Enter the index values of x[n] = ' );
h = input(' Enter the Unit Impulse Response h[n] = ' );
nh = input( ' Enter the index values of h[n] = ' );

% Convolution %
y = conv( h, x );
% Index Sequence of y[n] %
ny = ( min(nx)+min(nh) ) : 1 : ( max(nx)+max(nh) );
disp(' The Output Sequence = ' )
disp( y )
disp(ny )
figure(1)
subplot( 3, 1, 1 )
stem ( nx, x, 'filled' )
axis( [ min(nx) max(nx) min(x) max(x) ] )
xlabel(' n--> ')
ylabel(' x[n] ')
title(' Input Signal x[n] ' )

subplot( 3, 1, 2 )
stem ( nh, h, 'filled' )
axis( [ min(nh) max(nh) min(h) max(h) ] )
xlabel(' n--> ')
ylabel(' h[n] ')
title(' Input Signal h[n] ' )

subplot( 3, 1, 3 )
stem ( ny, y, 'filled' )
axis( [ min(ny) max(ny) min(y) max(y) ] )
xlabel(' n--> ')
ylabel(' y[n] ')
title(' Input Signal y[n] ' )

AFTER THE EXECUTION:

>>
Enter the input signal x[n] = [ 3, 4, 5, 2, 1 ]
Enter the index values of x[n] = -2:2
Enter the Unit Impulse Response h[n] = [ 1, 3, 2, 1]
Enter the index values of h[n] = -1:2
The Output Sequence =
3 13 23 28 21 12 4 1
-3 -2 -1 0 1 2 3 4

K.V.G. College of Engineering, Sullia D.K.

 DSP Lab Manual

>>

Page | 16

Fig 3.2 Screenshot of output of the conv_noncausal.m program

Commutative, Associative & Distributive properties of Linear

Convolution

clc
clear all
close all
x = [3, 1, 2, 1];
h1 = [ 1, 2, 2, 1 ];
h2 = [ 2, 4, 2, 4 ];

%Commutative Property

y1 = conv(x, h1)
y2 = conv(h1, x)

if y1 == y2
disp('Commutative Property of convolution is proved');
else
disp('error in the computation')
end

K.V.G. College of Engineering, Sullia D.K.

 DSP Lab Manual

% Associative Property

y3 = conv(x, conv(h1, h2))

y4 = conv(conv(x, h1), h2)
Page | 17
if y3 == y4
disp('Associative Property of convolution is proved');
else
disp('error in the computation')
end

% Distributive Property

y5 = conv(x, h1+h2)
y6 = conv(x, h1) + conv(x, h2)

if y5 == y6
disp('Distributive Property of convolution is proved');
else
disp('error in the computation')
end

>>
y1 =
3 7 10 10 7 4 1

y2 =
3 7 10 10 7 4 1

Commutative Property of convolution is proved

y3 =

6 26 54 86 102 96 72 40 18 4

y4 =

6 26 54 86 102 96 72 40 18 4

Associative Property of convolution is proved

y5 =

9 21 24 34 19 14 5

y6 =

9 21 24 34 19 14 5

K.V.G. College of Engineering, Sullia D.K.

 DSP Lab Manual

Distributive Property of convolution is proved

EXPT No: 2 b) DATE:

CIRCULAR CONVOLUTION OF TWO SEQUENCES

Page | 18
AIM: To execute an m-file program to compute the circular convolution of two given
finite-duration discrete-time sequences and to verify the commutative, associative
and distributive properties.

THEORY:

Circular convolution of two finite duration discrete time sequences x[n] and h[n] is
given by

() =
= [][( )] = = [] [( )] (4.1)

IfNx is the length of x[n] and Nh is the length of h[n] then the length of y[n] is Ny
given by Ny = Max(Nx,Nh ) (4.2)

The computation of y[n] is based on the equalization of the lengths of x[n] and h[n]
by appending appropriate zeros with the following logic:

ifNx>Nh

hn = [ hn, 0, 0, 0, , 0 ]
(Nx -Nh) zeros
else
xn = [ xn, 0, 0, 0, , 0]
(Nh- Nx ) zeros
end

The length of y[n] is Ny = Max(Nx,Nh). Circular convolution performed in this

manner has no practical significance. HoweverLinear convolution result can be
obtained from circular convolution using the following steps:

Append ( Nh - 1 ) zeros to x[n]andthus resulting length is (Nx+ Nh - 1 )

Append ( Nx - 1 ) zeros to h[n] andthus resulting length is (Nx+ Nh - 1 )
Perform circular convolution y[n] = x[n] (*) h[n]
Resulting y[n] is same as linear convolution between x[n] & h[n]

Linear convolution is the response of an LTI system.

Example:
Given x1(n) = {1, 1, 2, 1} and x2(n) = {1, 2, 3, 4}. Arrange x1(n) and x2(n) in circular
fashion as shown below

K.V.G. College of Engineering, Sullia D.K.

 DSP Lab Manual

Page | 19

To get x2(-m), rotate x2(m) by 4 samples in clockwise direction.

x3(0) = x1(m) x2(-m)
= x1(0) x2(0) + x1(1) x2(3) + x1(2) x2(2) + x1(3) x2(1)
= 1 + 4 + 6 +2 = 13

x3(0) = 13

Keep x1(m) constant and rotate x2(-m) once to compute further values.
To get x3(1) rotate x2(-m) by one sample in anti-clockwise direction x2(1-m)

K.V.G. College of Engineering, Sullia D.K.

 DSP Lab Manual

x3(1) = x1(m) x2(1-m)

= x1(0) x2(1) + x1(1) x2(0) + x1(2) x2(3) + x1(3) x2(2)
= 2 + 1 + 8 + 3 = 14
x3(1) = 14

Page | 20 To get x3(2) rotate x2(1-m) by one sample in anti-clockwise direction x2(2-m)

x3(2) = x1(m) x2(2-m)

= x1(0) x2(2) + x1(1) x2(1) + x1(2) x2(0) + x1(3) x2(3)

= 3 + 2 + 2+ 4 = 11
x3(2) = 11

To get x3(3) rotate x2(2-m) by one sample in anti-clockwise direction

x2(3-m)

x3(3) = x1(m) x2(3-m)

= x1(0) x2(3) + x1(1) x2(2) + x1(2) x2(1) + x1(3) x2(0)
= 4 + 3 + 4 + 1 = 12
x3(3) = 12
The convoluted signal is,
x3(n) = {13, 14, 11, 12}

ALGORIHM: Circular Convolution

Begin

K.V.G. College of Engineering, Sullia D.K.

 DSP Lab Manual

Read xn and hn
Compute
length of xn, hn
if (Nx>Nh)
hn= [hn, 0,0,0]
Page | 21 else
xn= [xn, 0,0,0]
yn = xk h((n,k)) , n
displayxn, hn, yn
End

PROGRAM:circ_conv.m
clc
clear all
close all

x = input(' Enter the samples of x[n] = ' );

Nx = length( x );

h = input(' Enter the samples of h[n] = ' );

Nh = length( h );

if ( Nx>Nh )
h = [ h zeros( Nx - Nh ) ];
else
x = [ x zeros( Nh - Nx ) ];
end

N = max(Nx, Nh );

temp = [ h(1) fliplr( h( 2:N ) ) ];

y(1) = sum( x .* temp );

for k = 2 : N
temp = circshift( temp', k-1 )';
y( k ) = sum( x .* temp );
end

disp(' Circular convolution of x[n] and h[n] is y[n] ' )

disp( y )

X1 = fft(x);
H1 = fft(h);
Y1 = X1 .* H1;
y1 = ifft(Y1);

disp(' Circular convolution of x[n] and h[n] in frequency domain is y1[n] ' )
disp( y1 )

X2 = fft(x, Nx+Nh-1);
H2 = fft(h, Nx+Nh-1);

K.V.G. College of Engineering, Sullia D.K.

 DSP Lab Manual

Y2 = X2.*H2;
y2= ifft(Y2, Nx+Nh-1);

disp(' Linear convolution of x[n] and h[n] in frequency domain is y2[n] ' )
disp( y2 )
Page | 22
figure( 1 )
subplot( 3, 1, 1 )
stem( 0 : Nx-1, x, 'r', 'filled' )
title(' Discrete Time Sequence x[n] ' )
xlabel(' n--> ' )
ylabel(' x[n] ' )

subplot( 3, 1, 2 )
stem( 0 : Nh-1, h, 'r', 'filled' )
title(' Discrete Time Sequence h[n] ' )
xlabel(' n--> ' )
ylabel(' h[n] ' )

subplot( 3, 1, 3 )
stem( 0 : N-1, y, 'r', 'filled' )
title(' Discrete Time Sequence y[n] ' )
xlabel(' n--> ' )
ylabel(' y[n] ' )

AFTER THE EXECUTION:

>>
Enter the samples of x[n] = [ 1, 2, 3, 4]
Enter the samples of h[n] = [ 1 ,2, 2, 1 ]
Circular convolution of x[n] and h[n] is y[n]
17 15 13 15
Circular convolution of x[n] and h[n] in frequency domain is y1[n]
17 15 13 15
Linear convolution of x[n] and h[n] in frequency domain is y2[n]
1 4 9 15 16 11 4

K.V.G. College of Engineering, Sullia D.K.

 DSP Lab Manual

Page | 23

Fig 4.1 Screenshot of output of the circ_conv.m program

Commutative, Associative & Distributive properties of Convolution

clc
clear all
close all

x = [3, 1, 2, 1];
h1 = [ 1, 2, 2, 1 ];
h2 = [ 2, 4, 2, 4 ];

%Commutative Property

X1 = fft(x);
H1 = fft(h1);

Y1x = X1 .* H1;
Y1h = H1 .* X1;

y1x = ifft(Y1x)
y1h = ifft(Y1h)

if y1x == y1h
disp('Commutative Property proved');
else
disp('Error in compuatation');
end

K.V.G. College of Engineering, Sullia D.K.

 DSP Lab Manual

%Associative Property

H2 = fft(h2);
Yxh = X1 .* ( H1 .* H2 );
Yhx = (X1 .* H1) .* H2;
Page | 24 yxh = ifft( Yxh )
yhx = ifft(Yhx)

if yxh == yhx
disp('Associative Property proved');
else
disp('Error in compuatation');
end

%Distributive Property
Y11 = X1 .* ( H1 + H2);
Y12 = X1 .* H1 + X1 .* H2;
y11 = ifft(Y11)
y12 = ifft(Y12)
if y11 == y12
disp('Distributive Property proved');
else
disp('Error in compuatation');
end

>>y1x =
10 11 11 10

y1h =
10 11 11 10
Commutative Property proved
yxh =
126 126 126 126

yhx =
126 126 126 126

Associative Property proved

y11 =
28 35 29 34

y12 =

28 35 29 34

Distributive Property proved

>>

K.V.G. College of Engineering, Sullia D.K.

 DSP Lab Manual

EXPT No: 3 a) DATE:

AUTOCORRELATIONOF A SEQUENCE AND

Page | 25
VERIFICATION OF ITS PROPERTIES
AIM: To execute an m-file program to find the autocorrelation of a given sequence
and to verify its properties.

THEORY:The autocorrelation of a sequence x[n] is given by,

[ ] =
= [][ ] = = [ + ] [] (5.1)

Where the integer parameter l is called the lag indicator. Eqn. 5.1 indicates that the
sequence x[n] is shifted by lsamples with respect to itself. The operation defined by
eqn. 5.1 is identical to convolution except for the time folding. It may be recalled that
in convolution one of the sequence is time reversed and then the operations of
shifting multiplications and additions are carried out. However in correlation except
for the folding he rest of the three operations are carried out.

PROPERTIES OF AUTO CORRELATION:

(i) [ ]attainsmaximum value when the lag parameter l = 0.

[] =
= [][ ] = = |[]| = (5.2)
Where denotes the energy of the signal x[n].

(ii) The auto correlation results in a symmetric sequence

i. e., [] = [ ] (5.3)

Auto correlation finds application in digital communication.

Autocorrelation is a mathematical tool for finding repeating patterns, such as the
presence of a periodic signal which has been buried under noise, or identifying the
missing fundamental frequency in a signal implied by its harmonic frequencies. It is
used frequently in signal processing for analysing functions or series of values, such
as time domain signals. Informally, it is the similarity between observations as a
function of the time separation between them.

Example: x[n] = { 1, 2, 3, 4 }

rxx[ 0] => 1 2 3 4 rxx[ 1] => 1 2 3 4

1 2 3 4 1 2 3 4
_________________________ _______________________
1 + 4 + 9 + 16 = 30 2 + 6 + 12 =20
________________________ ______________________

K.V.G. College of Engineering, Sullia D.K.

 DSP Lab Manual

rxx[ 2] => 1 2 3 4 rxx[ 3] =>1 2 3 4

1 2 3 4 1 2 3 4

________________________ _______________________
3 + 8 = 11 4
Page | 26
___________________________ _______________________
rxx[ -1] => 1 2 3 4 rxx[ -2] => 1 2 3 4

1 2 3 4 1 2 3 4

__________________________ ________________________

2 + 6 + 12 = 20 3 + 8 = 11

rxx[ -3] => 1 2 3 4

1 2 3 4

_________________________
4
_________________________

ALGORIHM: Autocorrelation

Begin

Read xn
Compute
length of xn
rxx = xn x (n,l) ,
Ex= xn, xn
E = rxx(0)
displayxn, rxx
End

PROGRAM:Auto_Corr.m

clc
clear all
close all

x = input(' Enter the samples of the sequence x[n] = ' );

% Compute Auto-Correlation

rxx = xcorr( x );

K.V.G. College of Engineering, Sullia D.K.

 DSP Lab Manual

disp(' The Autocorrelation Sequence rxx[ l ] = ' )

disp(rxx )

N = length( x );
nx = N - 1;
Page | 27 l = -nx : 1 : nx;
L = length( l );
lx = L - 1;

figure( 1 )
subplot ( 2, 1, 1 )
stem ( 0 : nx , x , 'r', 'filled' )
xlabel(' n-- > ' )
ylabel(' x[n] ' )
title ( ' Input D.T. Sequeence x[n] ' );

subplot( 2, 1, 2 )
stem( 0 : lx, rxx, 'r', 'filled' )
xlabel(' l --> ' )
ylabel(' rxx[l] ' )
title(' Autocorrelation Sequence ' )

% Autocorrelation using Convolution Function

rxx_c = conv( x, fliplr( x ) );

disp(' The Autocorrelation computed using Convolution ' )

disp(rxx_c )

disp(' The Autocorrelation computed using Convolution ' )

disp(rxx_c )

% Autocorrelation with Zero Lag Quantifies Energy in x[n]

Ex = sum( x .^ 2 );
disp(' The Energy of the Signal ' )
disp( Ex );
disp(' The Energy of the Signal from autocorrelation' )
disp( rxx( round( L/2 ) ) )

AFTER THE EXECUTION:

>>
Enter the samples of the sequence x[n] = [ 1, 3, 2, 4 ]
The Autocorrelation Sequence rxx[ l ] =
4.0000 14.0000 17.0000 30.0000 17.0000 14.0000 4.0000

The Autocorrelation computed using Convolution

4 14 17 30 17 14 4

The Energy of the Signal

30
The Energy of the Signal from autocorrelation
30
>>

K.V.G. College of Engineering, Sullia D.K.

 DSP Lab Manual

Page | 28

Fig 5.1 Screenshot of output of the Auto_Corr.m program

EXPT No: 3 b) DATE:

CROSS CORRELATION OF TWO GIVEN SEQUENCES

AND VERIFICATION OF ITS PROPERTIES
AIM: To execute an m-file program to find the cross-correlation of two given
sequences and to verify its properties.

THEORY:Correlation is a mathematical operation that closely resemble

convolution. The objective in computing the correlation between two different
signals is to measure the degree to which the signals are similar and thus to extract
some useful information.
The cross-correlation of two real-valued, finite energy, sequences x[n] and y[n] is
given by,

[ ] =
= [][ ] (6.1)

[ ] = = [] [ ] (6.2)
= , , ,

Where the integer parameter l is called the lag indicator.

Properties of cross-correlation:
) [ ] (6.3)

K.V.G. College of Engineering, Sullia D.K.

 DSP Lab Manual

) [] = [ ] (6.4)
Where Ex and Ey are energies of the signals x[n] and y[n] respectively. The
application of correlation finds in RADAR, SONAR and so on.

Page | 29 EXAMPLE:

Compute the cross correlation of the two sequences:x[n] = { 1, 3, 2, 4 } and

y[n] = { 1, 2, 4, 3 }.
Solution:

ALGORIHM: Cross-correlation

Begin

Read xn, yn
Compute
length of xn , yn
rxy = xn y (n,l) ,
ryx = ynx (n,l) ,
displayxn,yn,rxy, ryx
End

PROGRAM:Cross_Corr.m

clc
clear all
close all

x = input(' Enter the samples of the sequence x[n] = ' );

y = input(' Enter the samples of the sequence y[n] = ' );

rxy = xcorr( x, y );
ryx = xcorr( y, x );

K.V.G. College of Engineering, Sullia D.K.

 DSP Lab Manual

% Verify the failure of commutative property with rxy and ryx %

disp(' The Cross-Correlation between x[n] and y[n] ' )

disp(rxy )

disp(' The Cross-Correlation between y[n] and x[n] ' )

Page | 30 disp(ryx )

figure( 1 )
subplot( 2, 1, 1 )
stem ( x, 'filled')
xlabel(' n--> ' )
ylabel(' x[ n ] ' )
title(' First Sequence x[n] ' )

subplot( 2, 1, 2 )
stem ( y, 'filled')
xlabel(' n--> ' )
ylabel(' y[ n ] ' )
title(' Second Sequence y[n] ' )

figure(2 )
subplot( 2, 1, 1 )
stem ( rxy, 'filled')
xlabel(' l--> ' )
ylabel(' rxy[ l ] ' )
title(' The Cross-Correlation Sequence rxy[ l ] ' )

subplot( 2, 1, 2 )
stem ( ryx, 'filled' )
xlabel(' l--> ' )
ylabel(' ryx[ l ] ' )
title(' The Cross-Correlation Sequence ryx[ l ] ' )

AFTER THE EXECUTION:

>>

Enter the samples of the sequence x[n] = [ 1, 3, 2, 4 ]

Enter the samples of the sequence y[n] = [ 1, 2, 4, 3 ]
The Cross-Correlation between x[n] and y[n]
3.0000 13.0000 20.0000 27.0000 23.0000 10.0000 4.0000

The Cross-Correlation between y[n] and x[n]

4.0000 10.0000 23.0000 27.0000 20.0000 13.0000 3.0000

>>

K.V.G. College of Engineering, Sullia D.K.

 DSP Lab Manual

Page | 31

Fig 6.1 Screenshot of the output of Cross_Corr.m program

K.V.G. College of Engineering, Sullia D.K.

 DSP Lab Manual

EXPT No: 4 DATE:

LINEAR DIFFERENCE EQUATION

Page | 32 AIM: To execute an m-file program to find the solution of a given difference
equation(i) without initial conditions and (ii) with initial conditions.

THEORY:Thegeneral form linear constant coefficient difference equation is given

by: [] =
= [ ] + = [ ] (7.1)

Applying z-transform to eqn. (7.1) [] =

= [ ]

+

= []

(7.2)

Rearranging eqn(7.2)
()
=

[] = = (7.3)
() +
=

H(z) is called transfer function of the system.

Linear Constant Coefficient Difference Equations(LCCDEs) model LTI systems.
Equation (7.1) governs a recursive or Infinite Impulse Response(IIR) filter. If
except for , & =1, then
[] =
= (7.4)

Equation(7.4) is identical to convolution summation with {bk} being the filter

coefficients of a Finite Impulse Response (FIR) Filter. Equation 7.1 models a
recursive IIR system.

EXAMPLE:

ALGORIHM: LINEAR DIFFERENCE EQUATION WITHOUT INITIAL

CONDITIONS(INITIALLY RELAXED SYSTEM)

Begin

Read ak, bk
Compute , , andhnfromrational H(z) using partial fraction expansion.

= + + +
+ +

hn= IZT{ Hz}
define input signal to be unit step input
xn = un
solv =
= + =
displayr,p,k, h, yn
End

K.V.G. College of Engineering, Sullia D.K.

 DSP Lab Manual

PROGRAM: LCCDE1.m

% With ZERO initial conditions OR system initially relaxed.

Page | 33
clc
clear all
close all

b = input(' Enter the Numerator Coefficients = ' );

a = input(' Enter the Denomenator Coefficients = ' );

[ r, p, k ] = residuez( b, a );

disp(' Residues = ' )

disp( r )
disp(' Poles = ' )
disp( p )
disp(' Constants = ' )
disp( k )
disp(' ' )

[ h, n ] = impz( b, a );
% Impulse Response of the System

disp(' Impulse Response of the System h[n] = ' )

disp( h' )
disp(' ' )

figure( 1 )
subplot( 2, 1, 1 )
stem( n, h )
xlabel(' n--> ' )
ylabel(' h[n] ' )
title(' Impulse Response of the System ' )

% Exciting the System with Unit Step Signal

x = ones( 1, 12 );
y = filter( b, a, x );
disp(' The Response of the System y[n] for Step input =' )
disp( y )
disp(' ' )
% Alternatively
disp(' Alternatively the Response of the System y[n] for Step input =' )
y1 = filter( h, 1, x );
disp( y1 )

subplot( 2, 1, 2 )
stem( 0:length( y ) - 1, y )
xlabel(' n--> ' )
ylabel(' y[n] ' )
title(' Response of the System ' )

K.V.G. College of Engineering, Sullia D.K.

 DSP Lab Manual

AFTER THE EXECUTION:

>>

Page | 34
Enter the Numerator Coefficients = [ 1, 1/4, 1/8 ]
Enter the Denomenator Coefficients = [ 1, 1/3, 1/6 ]

Residues =
0.1250 + 0.0559i
0.1250 - 0.0559i

Poles =
-0.1667 + 0.3727i
-0.1667 - 0.3727i

Constants =
0.7500

Impulse Response of the System h[n] =

Columns 1 through 7

1.0000 -0.0833 -0.0139 0.0185 -0.0039 -0.0018 0.0012

Columns 8 through 11

-0.0001 -0.0002 0.0001 0.0000

The Response of the System y[n] for Step input =

Columns 1 through 7

1.0000 0.9167 0.9028 0.9213 0.9174 0.9156 0.9169

Columns 8 through 12

0.9168 0.9166 0.9167 0.9167 0.9167

Alternatively the Response of the System y[n] for Step input =

Columns 1 through 7

1.0000 0.9167 0.9028 0.9213 0.9174 0.9156 0.9169

Columns 8 through 12

0.9168 0.9166 0.9167 0.9167 0.9167

K.V.G. College of Engineering, Sullia D.K.

 DSP Lab Manual

Page | 35

Fig 7.1 Screenshot of output of the LCCDE1.m program

ALGORIHM: LINEAR DIFFERENCE EQUATIONWITH INITIAL

CONDITIONS

Begin

Read ak, bk
Compute
hn= IZT{ Hz}
define input signal to be unit step input
xn = un
solve =
= + = for yzi and yzs
= yzi+yzs
display h, yn
End

PROGRAM: LCCDE2.m

clc
clear all
close all

b = input(' Enter the Numerator Constant Coefficients = ' );

a = input(' Enter the Denomenator Constant Coefficients = ' );
ic = input( ' Enter the initial conditions y[-1], y[-2] : ' );

K.V.G. College of Engineering, Sullia D.K.

 DSP Lab Manual

% Impulse Response

[ h n ] = impz( b, a );

disp(' Impulse Response ' )

disp( h' )
Page | 36
% Unit Step Signal as Input

x = ones( 1, 12 );

% Zero Input Response

zi = filtic( b, a, ic );

% Total Response = Zero Input Response + Zero State Response

y = filter( b, a, x, zi );

disp(' Response of the System for Step Signal ' )

disp( y )

figure( 1 )
subplot( 2, 1, 1 )
stem( n, h )
xlabel(' n--> ' )
ylabel(' h[n] ' )
title(' Impulse Responnse ' )

N = length( y );
ny = 0:N - 1;
subplot( 2, 1, 2 )
stem(ny, y )
xlabel(' n--> ' )
ylabel(' y[n] ' )
title(' Response of the System ' )

AFTER THE EXECUTION:

>>
Enter the Numerator Constant Coefficients = [ 1, 1/3, 1/6 ]
Enter the Denomenator Constant Coefficients = [ 1, 1/4, 1/8 ]
Enter the initial conditions y[-1], y[-2] : [ 1, 0 ]
Impulse Response
Columns 1 through 7

1.0000 0.0833 0.0208 -0.0156 0.0013 0.0016 -0.0006

Columns 8 through 9

-0.0001 0.0001

K.V.G. College of Engineering, Sullia D.K.

 DSP Lab Manual

Response of the System for Step Signal

Columns 1 through 7

0.7500 1.0208 1.1510 1.0846 1.0850 1.0932 1.0911

Page | 37
Columns 8 through 10

1.0906 1.0910 1.0909

>>

Fig 7.2 Screenshot of output of the LCCDE2.m program

K.V.G. College of Engineering, Sullia D.K.