PRACTICAL FILE

DIGITAL SIGNAL PROCESSING

ECECCI1S

SUBMITTED BY:

Keshav Yadav
ECE-1
2020UEC2520
2
 INDEX
Exp Title Date
No.

1. To calculate and analyze the transfer function of an LTI system

specified by difference equation.

2. To compute DTFT and plot frequency response. To verify circular

time shifting and reversal property.

3. To compute circular convolution and linear convolution using

circular convolution

4. To compute response of a system using overlap save and overlap

add method.

5. To compute 100-point DFT of a given sequence.

6. To determine DFT coefficient using Goertzel algorithm

7. To design an FIR filter for a given specification and plot its

frequency response.

8. To design a digital differentiator using Hamming window.

9. To design a Hilbert transformer using Hamming window.

10. To design a Butterworth filter for given specifications.

11. To truncate and round off a given number and represent it in

signed magnitude form.

12. To write MATLAB code for zero input limit cycle operation.
 EXPERIMENT-1

Aim:-
An LTI system is specified by the difference equation

y(n) =09y(n — 1) +z(n)

(i) Find H(z) and sketch its pole-zero plot

(ii) Plot |H(e’*)| and ZH(e’¥).
(iii) Calculate and plot the steady-state response y,(n) to z(n) = cos(0.057n)u(n)

MATLAB Code:-
x=[1];
y=[1,-0.9]7
subplot(4,1,1);
zplane (X,y) %plotted pole zero plot
h=freqz(x,y); %transfer function
subplot(4,1,2)
plot(abs (h))
title ('Magnitude response');
subplot(4,1,3)
plot(angle (h))
title ('Phase response');

n=1:100
z2=zeros (length(n));
z2(1)=0;
for 1i=2:length(n)
z2(1)=0.9%2z2 (i-1) +cos (0.05*pi*i);
end
subplot(4,1,4);
stem(n,z2);
title("Steady state response");
Output:-

1:5
-]
>
go @%
5,
g i
E
Real Part
o Magnitude response
sk ]
o n
0
o Phase response
05p ]
TO0

Steady state response )

T iy
n L h
40 50 0 70 80 a0 100

Result-

The transfer function of the given difference equation was found and its
frequency response and pole-zero plot was plotted.
 EXPERIMENT-2

(a) Let z(n) be a 4-point sequence:

1 0<n<3
z(n) =
0 otherwise

(i) Compute DTFT X (¢’“) and plot its magnitude and phase
(i) Compute 4-point. 8-point. and 16-point DFT of z(n).

(b) Consider a 9-point sequence (n) = {1,2,3,4,5,6,7,8,9}

1
(i) Determine and plot the following sequences: x((n —3)q). and z((n+ 3)q) and verify the circular
time-shifting property.
(1) Determine and plot z((—n)o). Verify the circular time-reversal property.

MATLAB Code:-

2a)clc;
close all;
clear all;

x= [1,1,1,11;

w=linspace (0,4*pi, 628);

X=zeros (length(w));
for i=1:length(w)
sum=0;
for n=1:4
sum=sum+ (exp (-j*w (1) *n))
end
X (1)=sum;
end
figure;
subplot(5,1,1)
plot(w,abs (X))
title ('Magnitude response of DTET');
subplot(5,1,2)
plot(w,angle (X))
title ("Phase response of DTFT");

wk=linspace (0,4*pi,8); %4 point DFT, double the range so

8 points given
X2=zeros (length (wk)) ;
for i=1:length(wk)
sum=0;
 for n=1:4
A=-j*wk
(1) *n;
Z=exp (BA);
y=x(n)*z;
sum=sum+y;
end
X2 (1)=sum;
end
subplot(5,1,3)
stem(wk, abs (X2));
title ("Magnitude response of 4 point DFT");

wkl=linspace(0,4*pi,16);
X3=zeros (length (wkl)) ;
for i=1l:length(wkl)
sum=0;
for n=1:4
A=-j*wkl(1)*n;
Z=exp (BA);
y=x(n)*z;
sum=sum+y;
end
X3 (1)=sum;
end
subplot (5,1,4)
stem(wkl,abs(X3));
title ("Magnitude response of 8 point DFT");

wk2=1linspace (0, 4*pi, 32);

X4=zeros (length (wk2)) ;
for i=1:length(wk2)
sum=0;
for n=1:4
A=-j*wk2
(1) *n;
Z=exp (BA);
y=x(n)*z;
sum=sum+y;
end
X4 (1)=sum;
end
subplot(5,1,5)
stem(wk2,abs (X4));
title ("Magnitude response of 16 point DFT");
2b)
clc
clear all
a=[1,2,3,4,5,6,7,8,9];
%Verified circular shifting for first shift x<n-3>
outlifin:zeros(length(a));
for i=1l:length(a)
j=mod (1+3, length(a));
if (§<=0)
j=j+length(a )
outl_fin(j)= a(i);
else
outl fin(j)=a(i);
end
end
%Verifying the shift by plotting magnitude and phse
spectrum
A=fft (a,length(a));
outl finfft=fft (outl fin,length(a));
subplot(2,3,1)
stem(a)
subtitle ("Original sequence");
subplot(2,3,2)
stem(angle (2))
subtitle ("Phase plot of original"
subplot(2,3,3)
stem(abs (R))
subtitle ("Magnitude plot of original”
subplot (2, 3,4)
stem(outl_fin)
subtitle ("Shifted sequence");
subplot(2,3,5)
stem(angle (outl_finfft)
subtitle ("Phase plot of shifted")
subplot(2,3,6)
stem(abs (outl_finfft))
subtitle ("Magnitude plot of shifted")

% Finding x<n+3>9
outZifin:zeros(length(a));
for i=l:length(out2_fin)
j=1-3;
if (§<=0)
j=j+length(a);
out2_fin(j)=a(i);
else
 out2 fin(j)=a(i);
end
end

% x<-n>
out3_ fin=zeros(length(a))
for i=1l:length(a)
j=length(a)-i+1
out3 fin(j)=a(i)
end
figure
subplot(2,2,1)
stem(a)
subtitle ("Original sequence");
subplot(2,2,2)
stem(angle (2))
subtitle ("Phase of the sequence");
subplot(2,2,3)
stem(out3_fin)
subtitle ("reverse sequence");
£3=fft (out37fin, length(a))

subplot(2,2,4)
stem(angle (£3))
subtitle ("Phase of reversed sequence");

Output:-

Magnitude response of DTFT

4= T -
. g " s
e 7 N S
0
(] 2 4 6 10 12
Phase response of pTFT
2 I ~ ~ I
0f< L | 1 ‘ | i
2E ~ L
0 2
. Magnnude response of4 point Dn
3 & o
o 2 o 1 ! o 2
(] 2 T o 10 12
e Magnitude response ofB point DFT
2 o) . [o)
& o
p T g 5 8 8 [ T 2.9 6 8 @ T
(4 2 4 6 8 10 12
Magnitude response of 16 point DFT ‘

2223 y?».m(T)T.‘T[LAQ‘{MfQ\KSP A’]T

2 6
 Original sequence Phase of the sequence
10
¢ ®
- ol 2t o @9
@ [
5 (0] 0

¢ d

o2 0
f 5 10
). 0 5
&
)
o reverse sequence Phase of reversed sequence
P ® i [}
05 ‘
@ @
5 0 o
° j
P -0.5} o)

) At -
0 'l' N &
0 5 10 0 5

Result-
DTET for the given sequence was found and its frequency response was plotted.
Timeshifting and reversal property was verified.

1C
 - EXPERIMENT-3

Aim:-
Let zy(n) = {1.2.2,1} and z(n) = {1, ~1,~1,1}.
(a) Write a MATLAB program to compute the 4-point circular convolution x;(n) ® x,(n)
(b) Write a MATLAB program to compute their linear convolution using circular convolution

MATLAB Code:-

2a)
clc;
close all;
clear all;
% Circular convolution
x1=[1,2,2,1]
x2=[1,-1,-1,1]
N=4
y=zeros (N) ;
for n=0:N-1
for m=0:N-1
nl=mod (n-m, N)
y(n+l)=y (n+1) +(x1(m+1) *x2 (nl+1))
end
end
subplot(2,2,1) %2 rows 2 columns 1lst value
stem(x1)
subtitle ("Sequencel™);
subplot(2,2,2)
stem(x2)
subtitle ("Sequence 2");
subplot(2,2,3)
stem(y)
subtitle ("Circularly convolved sequence");

2b)
N=4
a=[12 2 1];
Nl=length(a);
b=[1 -1 -1 1];
N2=length(b);
Ny=N1+N2-1;
az=[a zeros(l,Ny-N1)];
bz=[b =zeros(l,Ny-N2)];
y=zeros(1,Ny);
%circular convolution

10
 for n=0:Ny-1
m=0:Ny-1
nl=mod(n-m,Ny)
bs=bz (nl1+1)
ab=az.*bs;
y(n+l)=sum(ab);
end
n=0:Ny-1
subplot(1,3,1)
stem(n,az, 'filled");
xlabel('n'");
title('xl(n)");

subplot(1,3,2)
stem(n,bz, "filled");
xlabel('n'")
title('x2(n)");

subplot(1,3,3);
stem(b) ;
xlabel('n");
'
title('y(n)");

Output:-

Sequence1 Sequence 2
2 & 1 q

15 05

1 0

0s 0.5

0 -1 € S
1 2 3 a4 1 2 3

, Circularly convolved sequence

il

11
 2—g—oXii) * LU 1 yin)
18 08 08
16 06 06
14 04 04
12 02 02
i) 0 oo 9 O
08 02 02
06 04 04
04 08 06
02 08 08

2 4o e o 4 o 2 3 4
n n n

Result-
Circular convolution and linear convolution using circular convolution were
found andplotted.

12
 EXPERIMENT-4

Aim:-
The impulse response of LTI system is i(n) = {1, 0, —1}. Write a MATLAB program to compute
the response of the system if the input is z(n) = (n+1). 0 < n <9,
(a) Use overlap-save method. Let the length of each section is N = 6
(b) Use overlap-add method. Let the length of each section is N =4

MATLAB Code:-

2a)
n=0:9;
N=6;
x=n+1;
Lenx= length(x);
h=[1 0 -1];
M=length (h);
M1=M-1;
L=N-M1;
hz=[h zeros(l,N-M)]; S%Appending N-M Zeros
nhz=0:1length(hz)-1;
xz=[zeros(1,M1l) x zeros(l,N-1)]; %Preappending M-1 Zeros
nxz=0:1length(xz)-1;
K=ceil ((Lenx+M1-1) /L) ;
y=zeros (K,N) ;
for i=0:K-1
X1=xz (1*L+1:1*L+N)
for j=0:N-1
m=0:N-1;
nl=mod (j-m,N);
hs=hz (nl+1);
xh=xi.*hs;
vy (i+1,J+1)=sum(xh);
end
end
y=y(:,M:N)'; %DiscaRding First M-1 Samples
y=[y(:)]1'; %Concatening The Output
ny=0:1length(y)-1;
subplot(1,3,1);
stem(nxz,xz,"'filled");
xlabel('n");
ylabel ("x(n)");
subtitle('x(n)");
title ("Using Overlap-save method");
subplot(1,3,2);

13
stem(nhz,hz,'filled");
xlabel('n'");
ylabel('h(n)")
subtitle ('h(n) )7
subplot(1,3,3)
stem(ny,y, 'filled');
xlabel('n'");
ylabel ("y(n)');
subtitle('y(n)"' )7

2b)
clear all;
close all;
clc;
n=0:9;
N=4;
x=n+1;
Lenx=length (x);
h=[1 0 -1];
M=length (h);
M1=M-1;
L=N+M-1;
hz=[h zeros(l,L-M)]; %Appending N-M Zeros
nhz=0:1length(hz)-1;
K=ceil (Lenx/N) ;
xx=[x zeros(l,N*K-Lenx)]; %Preappending M-1 Zeros
nxx=0:length(xx)-1;
y=zeros (K,N) ;
for i=0:K-1
X1=xxX (1*N+1:N* (1+1));
xr=[x1 zeros(1l,M1)1;
for j=0:L-1
m=0:L-1;
nl=mod (j-m, L) ;
hs=hz (nl+1);
xh=xr.*hs;
vy (i+1,J+1)=sum(xh);
end
end
yy=[1; %Adding Last M-1 Samples
for i=1:K-1;
y(i,:)=[y(i,1:N) y(i,N+1:L)+y(i+1,1:M1)];
end
yy=ly(1,1:L) y(2,M:L) y(3,M:L-Ml)];
ny=0:1length (yy)-1;
subplot(1,3,1);
stem(nxx,xx, "'filled");
xlabel('n");
ylabel ("x(n)");
subtitle('x(n)");
title ("Using overlap add method");
subplot(1,3,2);
stem(nhz,hz,'filled");
xlabel('n");
ylabel ('h(n)");
subtitle('h(n)");
subplot(1,3,3);
stem(ny, yy, 'filled');
xlabel('n");
ylabel ('y(n)");
subtitle('y(n)");

Output-
Using Overlap-save method
© X » h(n) 5

9 o 0.8
0
8 * 06

7 L 04
2
6 0.2

s
% ZE= oo oo 9l
ES
4 02
K
3 04

2 08
&
1 T 08 °
oo po—— 1 10
0 10 20 0 5 0 10

15
 Using overlap add method
x(n) h(n) n
0 * * 2
9 L 0.8

0
8 * 0.6

7 L] 04
2
6 0.2

Ts
% E= oo o0 9% = 4
4 02
5
3 04
2 06
-8
Y 08 .
0 e - 10 -
0 5 0 0 5 5 0
n n n

Result-

Response of the LTI system was found using overlap save and overlap add method.

16
 EXPERIMENT-5

Aim:-
Consider the sequence
z(n) = cos(0.487n) + cos(0.52wn)

(1) Determine and plot the 100-point DFT of x(n). 0 < n < 10 (pad 90 zeros).
(ii) Determine and plot the 100-point DFT of z(n), 0 < n < 100.
Comment on the results obtained in above two parts.

MATLAB Code-

:199;
k=1:100;
x=cos (0.48*pi*n)+cos (0.52*pi*n);
xpad=[x(1:10),zeros(1,90)]; %$first input signal with
90padded zeros
N=100;
w=2*pi/N;
DFTpadded=zeros (1,100);
for 1=1:100 %kth sample of DFT is 1 here
for i=1:100
a=xpad (1) *exp (-j*w* (1-1)*1);
DFTpadded
(1) =DFTpadded (1) +a;
end
end

xoriginal=x(1:100);
DFToriginal=zeros(1,100);
for 1=1:100 %kth sample of DFT is 1 here
for i=1:100
a=xoriginal
(1) *exp (-J*w* (i-1)*1);
DFToriginal (1)=DFToriginal(1l)+a;
end
end

subplot(2,2,1);
stem(x);
title("Original signal ");
subplot(2,2,2);
stem(DFToriginal);
title ("DFT of original signal");
subplot(2,2,3);
stem(xpad) ;
title("Zero padded input signal");
17
subplot(2,2,4);
stem(DFTpadded) ;
title ("DFT of zero padded signal");

Output:-

Original signal

Zero padded input signal

()

Result-

100 point DFT using zero padding and without zero padding was found,both
wereperiodic with n=100.

18
 EXPERIMENT-6

Aim:-
Let 2(t) = sin(27697t) + sin(271209¢), sampled at the rate of f, = 8000 Hz to get a discrete-time
signal z(n). Compute the DFT coefficient X (k) at the frequency bin k = 18,20,22,24,31,34,38
using Goertzel algorithm and detect which key of touch-tone pad is pressed

MATLAB Code-
clear all;
close all;
clc;
N=205; %DFT length
n=0:N-1;
x=s1n (2*pi*n*697/8000)+sin (2*pi*n*1209/8000);
if length (x)<N
xz=[x zeros(l,N-length(x))];
else
XZ=X
end
xl=[xz 0]
k=[18 20 22 24 31 34 38];
for i=1:7
W(i) =exp(-j*2*pi*k(i)/N);
den (i, :)=[1 -2*cos(2*pi*k(i)/N) 11;
vk(i,:) =filter(l, den(i, :), x1);
Xk(1)=vk(1,N+1)-W(i)*vk (i, N);
end

stem(k,abs
(Xk), 'filled");
xlabel ('k");
ylabel(' [X (k) |");
title ('DFT at k=18,20,22,24,31,34,38");

19
Output-

120 DFT at k=18,20,22,24,31,34,38

60
IX()I

40

201 4
=3

=3

=)
N
M

NB
N
S

No

©
R

©
£

©
&

©
Result- &

DFT coefficient at various values of k were found using Goertzel algorithm.

20
 EXPERIMENT-7

Aim:-
Design an FIR filter to meet the following specifications: passband edge F, = 2 KHz. stopband edge
F, = 5 KHz. Passband attenuation A, = 2 dB. Stopband attenuation A, = 42 dB, and Sampling
frequency Fr = 20 KHz. Determine the impulse response and provide a plot of the frequency
response of the designed filter

MATLAB Code-
clear all;
close all;
clc;

Fp=2000; $%Fp=2KHz
Fs=5000; %Fs=5KHz
Ft=20000; %Ft=20KHz
wp= (2*pi*Fp)
/Ft;
ws=(2*pi*Fs)/Ft;
trwidth=ws-wp; 3%Transition Width

M=cell (6.2*pi/trwidth); %FilterLength

tau=(((M-1)/2))
wc= (wp+ws) /2; %cutoff frequency
n=0:M-1;
hd=(sin(wc* (n-tau)))./(pi* (n-tau)); %desired filter
d(tau+l1)=0.35; %found using l'hopital's rule manually
whan=hann (M) ;
whan=whan."'; sconverting row to column matrix as to be
multiplied later
h=hd.*whan; S%$practical filter impulse response after
multiplying with window
w=0:0.02:pi;
Hw=freqz (h,1,w);
MagHw=abs (Hw) ; $Magnitude Response
HwdB=20*1log (MagHw) ;

subplot(2,2,1);
stem(n,hd, "filled");
axis([-1 M -0.15 0.51);
xlabel('n");
ylabel ('hd(n)");
title('Ideal Impulse Response');

subplot(2,2,2);
stem(n,whan, 'filled'");
axis([-1 M -0.1 1.2]);

21
 xlabel('n");
ylabel ('w(n)");
title('Hann Window');

subplot(2,2,3);
stem(n,h, 'filled");
axis([-1 M -0.15 0.51);
xlabel('n'");
ylabel ("w(n)');
title('Practical Impulse Response');

subplot(2,2,4);
plot (w/pi, HwdB) ;
axis ([0 1 -100 10]);
xlabel ('Frequency in pi ');
ylabel ('dB");
title ('Magnitude Response');

Output-

Ideal Impulse Response Hann Window

04 1

Zo2 G \
2 308
0
hd o

0 5 10 15 20 0 5 10 15 20
n B n
Practical Impulse Response Magnitude Response
0
04

5To2 T T % &0
0 .n‘o‘ . ™ ‘.m

-100
0 5 10 15 20 0 05 1
n Frequency in pi

Result-
Impulse response of the FIR filter was found and its frequency responsewas
plotted.

22
 EXPERIMENT-8

Aim:-
Design a length-21 digital differentiator using a Hamming window.

Hy(e™) = {" —jwe

\ %
we—Jow O<w<m
—rT<w<0

MATLAB Code

clear all;
close all;
clc;
M=21; S$Hamming Window Length=21
tau=(M-1)/2;
n=0:M-1;
hd=((cos (pi* (n-tau))) ./ (n-tau))-((sin(pi*(n-
tau))) ./ (pi*(n-tau).~2)); %manually calculated impulse
response
d(tau+l)=0; gmanually calculated using 1l'hopital's
whamm=hamming (M) '; %converted to column matrix for
multiplication
h=hd. *whamm;
w=0:0.01:pi;
Hw=freqz (h,1,w);
Hrw=exp (-j* (pi/2-10*w))
. *Hw; %amplitude plotted using the
function H(e” (jw))=Hr(e"jw) *exp(J (pi/2-tau*w))

subplot(2,2,1);
stem(n,hd, "filled");
axis([-1 M -1.2 1.21);
xlabel('n'");
ylabel ('hd(n)");
title('Ideal Impulse Response');

subplot(2,2,2);
stem(n,whamm, 'filled"');
axis([-1 M -0.2 1.21);
xlabel('n");
ylabel ("w(n)');
title ('Hamming Window');

subplot(2,2,3);
stem(n,h, 'filled");
axis([-1 M -1.2 1.2]);
23
 xlabel('n");
ylabel ('h (n)");
title('Practical Impulse Response');

subplot(2,2,4);
plot (w,Hrw) ;
axis([0 pi 0 pil);
xlabel ('Frequency');
ylabel ('Amplitude');
title ('Amplitude Response');

Output-

Ideal Impulse Response Hamming Window

hd(n)

w(n)

e
o

o
o

. wTfl TTT?»
0 5 10 15 20 0 5 10 18 20
n n
Practical Impulse Response Amplitude Response

0.5
Amplitude
n

-0.5

0 1 2
Frequency

Result-

A differentiator was designed using Hamming window.

24
 EXPERIMENT-9

Aim:-
Design a length-25 digital Hilbert transformer using a Hanning window.

Hy(e) :{ Je "‘eo Q<w<m

S
T<w<0

MATLAB Code-

clear all;
close all;
clc;
M=25; %Hanning Window of Length=25
tau=(M-1)/2;
n=0:M-1;
hd=(2./(pi* (n-tau))).* (sin(pi* (n-tau)/2).72);
d(tau+1)=0;
whamm=hamming(M) ';
h=hd. *whamm;
w=-pi:0.01l:pi; Shere we delay positive and negative
frequency phases by pi/2 so range is +ve and -ve bothS
Hw=freqz (h,1,w);
Hrw=exp (-j* ((pl/2)-12*w)) .*Hw; %amplitude plotted using
the function H(e” (jw))=Hr (e”jw)*exp (J (pi/2-tau*w))
%Note: This is type 3 or 4 type filter used as this is
antisymmetric
subplot(2,2,1);
stem(n,hd, "filled");
axis([-1 M -0.8 11);
xlabel('n'");
ylabel ('hd(n)");
title('Ideal Impulse Response');

subplot(2,2,2);
stem(n,whamm, 'filled"');
axis([-1 M -0.2 1.21);
xlabel('n'");
ylabel ("w(n)');
title ('Hanning Window');

subplot(2,2,3);
stem(n,h, 'filled");
axis([-1 M -0.8 11);
xlabel('n'");

25
 ylabel ( hin)");
title('Practical Impulse Response');

subplot(2,2,4);
plot (w/pi, Hrw);
axis([-1 1 -1.2 1.2]);
xlabel ('Frequency in pi Units');
ylabel ( 'ABmplitude');
title ('Amplitude Response');

Output-

4 Ideal Impulse Response Hanning Window

05
hd(n)

<o
o

-0.5 -wTfl fi'h

5 w0 1 W 25
o

n=

N
&

n
4 Practical Impulse Response Amplitude Response

0.5
Amplitude

-05 0 05
n=}

il
o

Frequency in pi Units

Result-

Hilbert transformer was designed.

26
 EXPERIMENT-10

Aim:-
Write a MATLAB program to design the digital Butterworth filter that meets the following speci-
fications:
(i) 1-dB ripple in the passband 0 < |w| < 0.37
(ii) At least 40 dB attenuation in the stopband Sr<jw| <
Use bilinear transformation method. Assume 7" = 1.

MATLAB Code-
T=1;
wp=0.3*pi;
ws=0.8%pi;
Ap=1;
As=40;
Ohmp= (2/T) *tan(wp/2) ;
Ohms=(2/T) *tan(ws/2) ;
X=(10"~(0.1*Ap)-1)/ (10~ (0.1*As)-1);
N=ceil (0.5* (1ogl0(X)/ (logl0 (Ohmp/Ohms)))) ;
Ohmc=0Ohmp/ ( (10~ (0.1*Ap) -1) ~(1/(2*N)));
[b,al=butter(N,Ohmc, 'low','s"); %specified the
order,cutoff and low pass type butterworth
Hs=tf (b, a)
[numdig, dendig]=bilinear(b,a,1/T);
Hz=tf (numdig, dendig, T);
Hz
w=0:0.01:pi;
Hw=freqz (numdig, dendig, w) ;
subplot(1,2,1);
plot(w,abs (Hw)) ;
xlabel ("Frequency");
ylabel ("Magnitude") ;
subplot(1,2,2)
plot(w,20*10ogl0 (abs (Hw))) ;
xlabel ("Frequency");
ylabel ("Magnitued in dB");

27
Output-

//

-
Magnitued in dB
Magnitude

—
o
>

3
xR
o
=

-160
0.2
-180

0 -200
0 1 2 3 0 1 2 3
Frequency Frequency

Result-
Desired Butterworth filter was designed and its frequency response was plotted.

28
 EXPERIMENT-11

Aim:-
Consider a number r = ,—f = 0.375. Write a MATLAB program to find Q[z] by applying (i)truncation.
and (ii) rounding in signed magnitude representation. Assume (B + 1) = 3 bit register (including
sign bit).

MATLAB Code-
x=3/8;
B=2;
xl=abs (x);
Qx=0;
for k=1:B
Qxbeq
(k) =fix (x1*2) ;
Qx=fix (x1*2)/ (2°k) +Qx;
x1=(x1*2)-fix (x1*2);
end
sg=sign(x);
if sg>0
Qxbeq=[0 Qxbeq];
else
Qxbeg=[1 Qxbeq];
end
disp('Truncated Number');
disp(Qx);
disp('Binary equivalent of truncated number is:');
disp (Qxbeq) ;

Output-

>> expll
Truncated Number
0.2500

Binary equivalent of truncated number is:

0 0 L 1

fe>> |

Result-

The given number was truncated and rounded .

29
 EXPERIMENT-12

Aim:-
Write a MATLAB program for the zero-input limit cycle operation. Assume input z(n) = Z(n),
the impulse response of the system h(n) = ( %]“ u(n). (B + 1) = 5. and rounding is used for
quantization.

MATLAB Code-
a=-1/2;B=4;
x=[7/8 zeros(1,20)];ycap=0;
for n=1:21
ay=abs (a*ycap)
ay=ay+(1/2) *27 (-B) 7
Qy=07
for k= :B
Qy=fix(ay*2)/(27k)+Qy;
ay=(ay*2) -fix(ay*2) ;
end Qy=sign (a*ycap) *Qy;
y(n)=Qy+x(n); ycap=y(n);

end
k=0:20
stem(k,y, 'filled');
ylabel ("Amplitude');
xlabel('Time index n');

Output-

I
os

Pl o
L
¥
-
Ey
-
v
®
B
-
§
*
@
-
G
EY
Gy
EY
g

o5 o 2 = - s 10 12 1 1 18 20
Time index n

Result-
The amplitude and magnitude response were plotted for zero input limit cycle operation.

30