0 1 2

Addition
A X(n)õ(n+1) {0, =
2, 0, 0, 0}

n
-1 01 2
EAddition
x(n) (n) ={0,0. 1.0,0)

-2-1 01 2
EAdition
A x(n) 8{n-1) = {0, 0, 0, 2 ,0}

-2 -1 l01 2
OEAddition
x(n)&(n-2) = {0, 0, 0, 0,1)

n
2-1 lo 1 2

Original Sequence x(n) = {1,2, 1, 2, 1}

n
-2-1 01 2
Fig. 3.22. x (n) =Addition of all decomposed samples

LINEAR CONVOLUTION (CONVOLUTION SUM)

It is interesting Proof of system LTI is completely
tsng to note that the proof of linear convolution
chararacterized
or

by unit impulse response h (n).

(n) 1s zero ne
relaxed LSI (LTI) system. A relaxed system means if input
x

put v Slaer its output is denoted

byh h (n). He) is zero. Let us say unit impulse ö (n) is applied to this system then
He h (n) is called as
lere, impulse response o SYSte .(3.10)
0)8n) T y (n) =h n) be De
Since the then output
shOuld

delay input by k' samples,

elayed by same
h 1s ime invariant, if we

An im amount. -

time-invariant systems lies in t h e i r

of linear was
ency-d n s for the analysis and representation to a complex exponential
input
of a system response
and freq L ation. The response The relation
between impulse

d e n n i t i o n of the frequency
r e s p o n s e .

transform pair.
ponse was then interpreted as a Fourier
 136
-k)
k)Ty (n) = h (n
(i)&(n- both sides by x (k), we get,
multiplying
Now,
k) y (n) = x (k) h (n - k)
(k) 8(n
-

(ii) x
th
linear; we
can apply superposition
theorem
Since the system
both sides,
we
is

get,
Hence, taKtg
y (n) = 2 x (k) h (n -

k)
(iv) x(k) 8 (n -k) k=-0o

k=-
Thus, we have

y (n)=2
x (k) h (n -

k
= - 0o

called as c o n v o l u t i o n this case sum. It

This summation is ut y (n)
and h(n). This i1s called as
linear convolution ad)ist
nd it is
is
onvolution of x(n) denotej
x (k) h (n -

k)
Therefore, x (n)» h (n) =

k=-
In this case, we can say that the behaviour of Ltl system is completelv a
impulse response.
char
Key Note:Although the convolution-sum expression is analogous to thecomvo
continuous-time linear system theory, the convolution sum should not he
voluta
approximation to the convolutional integral. The convolution integral plavs 1
nainira
roleincontinuous-time linear system theory. Basically, the convolution sum,
theoretical importance, often serves as an explicit realization of a
discrete-time
Thus, it is important to gain some insight into the properties of the convolution
calculations.
s

3.12 COMPUTATION OF LINEAR CONVOLUTION

The different methods used for the
computation of linear convolution may be ted a
G) Graphical method
(i) Using mathematical
(ii) Tabulation method
equation of convolution
(iv) Multiplication method.
i) Graphical Method
The linear
convolution of two
sequence is given by,
y (n)= x(k) h(n k -

We will K=~oo
have to
For n = 0, we calculate output for
get, different values of n.

y
(0) x(k) h ( -k)
Note that k= -

For n= 1, h(-k)
we
indicates folding of h (k).
get,
 DISCRETE-TIME SYSTEMS 137

k) .(3.17)
y (1)=2 x(k)h (1-
Here, the term h (1 - k) can be written as h (-k + 1). Thus, equation (3.17) becomes,

y (1) = x (k) h (- k+ 1) ..(3.18)

k=-
Here, h -k+ 1) indicates shifting of folded signal h (-k). It indicates that h(-k) is delayed by
1' sample. Similarly, for other values of'n' output y (n) is calculated.
Thus, different operations involved in the calculation of linear convolution may be listed as
under:
) Folding operation: It indicates folding of sequejnce h (k).
i) Shifting operation: Itindicates time shifting of h (-k) e.g.,h-k+ 1).
(i) Multiplication: It indicates multiplication of x (k) and h (n - k).

(iv) Summation: It indicates addition of all product terms obtained because of multiplication
of x (k) and h (n - k).

EXAMPLE 3.40. Obtain linear convolution of following sequences

x (n) = {1, 2, 1, 2} and h (n) = {1, 1, 1}

Solution: If the sequences are given in terms of 'n', then, we obtain x (k) and h (k) by replacing'n' by k'.
Therefore, x (k) = {1, 2, , 2}
and h (k) {1, 1, 1}
Arrow is not mentioned in any sequences. So, by default it is at first position.
or x (k) = {1, 2, 1, 2}

and h (k) = {1, 1, 1}

Now, we have equation of convolution as under:

y (n) = x (k) h (n -k)

k= -
...1)
Now, we have to decide range of 'n' and 'k'.
Range of 'n
Sasically, to calculate y (n), we should know from which value of 'n', we should start lower range of n) and
what is
ending value of 'n' (higher range of n). We will use following notations to decide range ofy (n):
y= Lowest range of y (n)
yh= Highest range of y (n)
x= Lowest range of x (k)
xh = Highest range of x (k)

h= Lowest range ofh (k)

h= Highest range of h (k)
Now, we use
following formulae to calculate range of y (n)
y x t h and y, = Xh +hh
From the
given sequences, we have,
0, x3
h= 0 and h,= 2
Thus, x+h = 0 +0 0
and =5
Yh +h, 3 + 2 =
 to y (5).
Thus range of y (n) is from y (0)
Range of'k same as
the sequence x (k). Thus
us, in this
always
summation sign will be case, the
The value of k in the becomes,
0 to 3. Hence, equation
(i)
range of k is from
k =

k)
y (n) =2x (k)
h (n
-

of n from 0 to 5,
different values
substituting
Now. we shall calculate output y (n) by
Calculation of y (0)
Putting n = 0 in equation (ii), we get,
3

k) x (k) h (-k)
(k) (0
- =
x
(0)=
k =0
k=0
method. Here, h (-k) indicates folded t
Now. we have to perform calculations using graphical
h (k)
(i) Sketch x(k)
Sketch of x(k) is as shown in figure 3.23.

X(k)= (1. 2. 1. 2)
hk) = {1, 1, 1}
h(-k)={1,1,1)

-2 -1 0
O 1 2 3 O 12
Fig. 3.23. Sequence x (k) Fig. 3.24. Sequence h (K) Fig. 3.25. Sequence h (-k)
i ) Sketch h (k)
Sketch of h(k) is shown in figure 3.24.
(iii) Foldh (k) to obtain h (-k)
Sketch of h (-k) is shown in figure 3.25.
iv) We obtain multiplication of x (k) and h (-k). This
shown in figure 3.26.
multiplication takes place on sample to sample Dasis

x(k) ={0, 0, 1,2,1,2}

-2-1 lo1 2 3
k

Multiplication
h-k)={1, 1,1,0,0,0}

-2-1 l0 1 2 3

x(k) h(-k) =
{0, 0, 1,0, 0, 0}

Fig. 3.26.
Product x(k) and h (
 DISCRETE-TIME SYSTEMS

T h e s u m m a t i o n

ofall product terms. 139

of multiplication, have only
Inthis
case,
a s result
we one
sample at
Ah(-k+1)= (0.1,1,1}
n =0.
y (0)= 1
Therefore, 11 1
Calculation ofy (1)
in equation (1), we get,
2 1 01 2
Puttingn=. Fig. 3.27.
3
Sequence h (-k+ 1)
y(n) =2x(k) h(-k) ...(iv) x(k) {0,=
0, 1,2, 1, 2}
k = 0

it
Here, h (1 -k) can be written as h (-k+ 1). Thus, equation (iv)

becomes,
3
-2-1 0 1 2 3
y (n) = x (k) h(- k + 1)
.(v
k = 0
Multiplication
) Now, h (-k + 1) indicates delay of h - k ) by '1' sample.
Hence, we shift the sequence in figure 3.25 towards right by 1'sample. h(-k+1) ={0, 1,1, 1,0,0}
This sequence is shown in figure 3.27.
Gi) Let us obtain multiplication of x (k) and h (-k + 1). It is
shown in figure 3.28. k
(i) Now, according to equation (v), we have to add all product -2-1 0 1 2 3
terms shown in figure 3.28 to obtain y(n).
Thus, y (1)= (1 x 1) + (2 x 1) = 1 +2
x(k) h(-k+1) = {0, 0, 1, 2, 0, 0}
or y (1) = 3

Calculation of y (2) 2
Putting n =2 in equation (1i), we get,
B 2-1 01 2 3
y (2) = x (k) h (2 - k)
..(vi) Fig. 3.28. Product ofx (k)
k = 0
and h (-k+ 1)
The term h (2-k) is same ash(-k+ 2).
3 X(k) = {1, 2, 1,2}

y (2) = x (k) h (-k+2) ...(vii)

k =0
i) Here, h -k+2) indicates delay of h (-k) by 2 samples. This
15
obtained by shifting sequence in figure 3.29 towards right by *2 2 3
samples. This sequence is shown in O 1
figure 3.29.
Multiplication
h(-k+2) ={1.1,1}
h(-k+2) ={1. 1, 1.0

k
0 1 2
Fig.3.29. Sequence h(-k+ 2) lo1 2 3
Ce obtain multiplication of x (k) and h (-k+2). It is shown
in figure 3.29

We add all product terms to obtain y 2, 1, 0}

x(k) h(-k+2) {1,
=

Thus, (2)
or yy (2)= (1
(2) = 4
x1) + (1
2) + (1 x x 1) =1+2+1
Caleulation of y (3)
utting n =
3 in
uation (ii), 01 2 3
we get,
h (-k+
y (3)= x (k) h (3 - k) Fig. 3.30. Product of x (k) and
..(viii)
= 0
 h (-k t 3).
The term h (3 -

k) is same as

+3)
h(-k
Thus v(3)=2 x(k)k 0
It in Bhown in figure 3:1
3) indicates delay of h (- k) by '*' samples.
() Here h (-k +

h(-k+3) {0. 1, 1, 1)
1 1

o ii
1 23
k

Fig. 3.31. Sequence h(-k

+ 3)
(ii) Let us obtain multiplication of x (k) and h (-k +3). This 8equence i8 shown in figuro & 0
2.

X(k) {1. 2,
=
1, 2)

1
k
01 2 3
Multiplication
h-k+3) =(0. 1, 1,1)

n
01 2 3

x{k) h(-k+3) = {0, 2, 1, 2}

k
01 2 3
Fig. 3.32. Product x (k) and h
ii) Let us add all (- k +
3)
product term to obtain y (3).
Therefore, y3) =(0 x 1) +(1 x 2) +(1 x
or, y3) = 5
1) + (2 x
1) +0 +2+1+2
Calculation of y(4)
Putting n 4 in equation (ii),
=
h(-k+4) ={0, 0, 1,1,1
we get,

1 1 1
yl4)=x (k) h(4 -k)
k =0
The termh (4- is
k) same as h
-k+ 4). ot233
Thus, 4
Sequence
h(-R**
Fig. 3.33.
y4) =x (k)
() Here, h-k+
4)
h(-k+4)
k 0
.xi)
) Let us indicates delay
obtain multiplication of xof(k)h(- k) by 4' samples. It is
andh-k+ shown in figure
4). This
sequence is shown
3.0
shown in
in figure 3.34.
E
 DISCRETE-TIME SYSTEMS
x(k)= {1., 2, 1,2, 0} 141

0 1 2 K
EMultiplication
h(-k+4) (0.0, 1,1, 1)

32 ''
01 2 3 4

x(k) h(-k+4) ={0. 0, 1, 2.0)

k
0 1 2 3 4
Fig. 3.34. Product of x (k) and h(-k+ 4)
(ii) Let us add all product terms to obtain y (4). h-k+5) ={0. 0,0, 1.1, 1}
Therefore, y +(1
(4)=x x0)+(2 x 0) 2
1) + (2 x 1) + (0 x 1) = 0 +0 +1+2+0
y (4) = 3
or
Calculation of y (5) l01 2 3 4 5
Putting n =5 in equation (i), we get,
Fig. 3.35. Sequence h (-k+ 5)

y (5)= x (k) h (5 - k) .(xii) X(k)= {1, 2, 1,2, 0, 0}

k = 0

The termh (5 -k) is same as h (-k+ 5).

2 2

Thus, y (5)=2 x (k) h (-k +5) ..(xii)

k = 0
0) Here, h (-k+5) indicates delay of h (-k) by 5' samples. It is shown lo 1 2 3 4 5
in figure 3.35. Multiplication

u)inLet us obtain multiplication ofx

shown
(k) and h (-k+5). This sequence is h(-k+5) {0, 0, 0, 1,
= 1, 1}
figure 3.36.
) Let us add all product terms to obtain y(5).
Thus, y5) (1 x0) + (2 x 0) + (1 0) + (2 x 1) + (0
= x x 1) + (0x 1)
I'he result shown in k
0,1.1.1

figure 3.37. of convolution y(n) represented graphically

1s
as

To12
y(n)= {1,3, 4, 5, 3, 2)
0, 0, 2,0, 0}
x(k) h(-k+5) {0.
=

ce
h-k+

23 4 5
3.36. Product of x (k) and
n Fig.
O 1 2 3 4 5 h(-k+ 5)
Fig.3.37. Sketch of y (n) {1, 3, 4, 5,
= 3
,2
24