PROPERTIES OF CONVOLUTION

Commutative Property(P-1)

( ) ∗ ℎ( ) ≡ ( ) ≡ ℎ( ) ∗ ( )

Literal meaning of commutative property is that, a group of quantities connected by operator gives
same result whatever the order of the quantities is involved. Which means, in convolution we can
change the order of two liner system and there will not be any change in the result.
We will get same result If we keep ( ) fixed and move ℎ( ) or vice-versa.

Associative Property(P-2)
Associative properties apply to three or more quantities. If you are convoluting, it does not matter
where you put parenthesis (bracket).

From input point of view

[ ( )∗ ( )] ∗ ( )= ( )
( ) ∗[ ( )∗ ( )] = ( )
From system point of view
Input to the system ( ) is convoluted with impulse response ℎ ( ) and we get output ( ). This
output is again convoluted with impulse response ℎ ( ) and say we get output / ( )

( ) ℎ ( ) ( ) ℎ ( ) /
( )

/( ) = [ ( ) ∗ ℎ ( )] ∗ ℎ ( )

The above system will be equivalent if we convolute ℎ ( ) and ℎ ( ) first and then again convolute
the resultant with ( )

( ) ℎ ( )∗ℎ ( ) /
( )

/( ) = [ℎ ( ) ∗ ℎ ( )] ∗ ( )
Distributive Property(P-3)
Convoluting the sum of two signals will give same result if we convolute them separately and then
add them
( ) ∗ [ℎ ( ) + ℎ ( )] ≡ [ ( ) ∗ ℎ ( )] + [ ( ) ∗ ℎ ( )]

ℎ ( )

∑ ( ) ≡ ( ) ℎ ( ) + ℎ( ) ( )
( )
ℎ ( )

Convolution Property of Delta Function(P-4)

If you convolute any signal with delta function you get back the same signal. Because if you
convolute you get overlap. If you integrate all overlap you get the original signal back. But how ti
comes? To understand graphically see the page five. An example has been illustrated without
calculation. However, there are two forms of this property.
Form-1 When weight/strength/area is 1
( )∗ ( − )= ( − )
When = 0 then ( ) ∗ ( )= ( )

Form-2 When weight/strength/area is more than 1

( )∗ ( − )= ( − )
When = 0 then ( )∗ ( )= (
Example Based on Property -4
Perform convolution ( ) = ( ) ∗ ℎ( ) where ( ) = ( ) and ℎ( ) = ( − 2)
Soln:
As per property 4 we can simply write
( ) ∗ ℎ ( ) = ( ) ∗ ( − 2) = ( − 2)
So property helps us to solve the problem quickly.
Derivative Property (P-5)
Finding the derivative of the output of an LTI system. There are two options. Both options will give
same result.

( ) = ( ) ∗ ℎ( ) = ( )= [ ( ) ∗ ℎ( )]

Method-1

= [ ( )] ∗ ℎ( )
Method-2

= ( )∗ ℎ( )

Example based on property 5

( )= ( )∗ ( )
Soln using method -1

( )= ( )∗ ( )= [ ( ) ∗ ( )] = ( ) ∗ ( )

= ( ) ∗ ( ) = ( ) (if you convolute any function/signal with impulse function then you get back
same signal, see the property 4)

Soln using method-2

( )= ( )∗ ( )= [ ( ) ∗ ( )] = ( )∗ ( )= ( )∗ ( )= ( )

(Those who have still confusion about convolution ( ) ∗ ( ) for them see the graphical
convolution below)

U(t) U(t)

U( ) U( )

U( ) U(− )

U( ) U( − )

t
Case-1 when t < o
U( ) U( − ) There is no overlap. Thus y(t)=0

t Case-2 when t > 0

U( ) U(t- )
The overlap amount is 0 to t. Value of u( ) in overlap=1, and
value of u(t- ) in overlap=1. Thus the convolution integral

= ( ). ( − ) = 1 = = ( )

t
As per definition this is ramp signal
, 0
( )=
0, ℎ

Convolution of a signal with delta function (Just for understanding)

δ( − 4) ℎ( ) =

4 ∞

δ( − 4) ℎ(− )

4 ∞

Case-1 when t<4, no overlap

δ( − 4) ℎ( − )

t 4 ∞
Case-2 when t=4,

δ( − 4) ℎ( − )

4 ∞
t=4
Case-3 when t=5

δ( − 4) ℎ( − )

4 t=5 ∞
Case-4 when t=6

δ( − 4) ℎ( − )

Case-5 when t=7 4 ∞

t=6

δ( − 4) ℎ( − )

Case-6 when t=8 4 t=7 ∞

δ( − 4) ℎ( − )

Case-7 when t=9 4 t=8

δ( − 4) ℎ( − )

4 ∞t=9
Integration and Differentiation (P-6)

Convolution of a signal with unit step signal is equal to integration of that signal

Mathematically, ( ) = ( ) ∗ ( ) = ( )

Proof of property 6
To prove property 6, we will employ different strategy.
If you perform the differentiation and then integration you get the same signal.

( ) = ( )

So let us differentiate first,

( )= ( )∗ ( )

( )= ( )∗ ( )

( )
= ( )∗ ( )

( )∗ ( ) [but from property 4 we know that ( ) ∗ ( ) = ( ) ]

( )
=

( )
∴ = ( )= ( )

Make your concept clear

Let us consider two signals

( )= and ℎ( ) = ( )
( ) ( )
1 1 1

So, we can see that ( ) and are same. Unit step function is a special case of switched
exponential with = 0. For further reading see the additional topic at the end of this file. Subsequent
problems have the use of switched exponential signal ( ) . Use of ( ) helps us to
understand the overlapping limit clearly.
Exercise 2.5(a), P-98, Srinath and Soliman
Use convolution integral where
( )= ( ) and ℎ( ) = ( )
( ) ( )
( − )
Soln:
( ) = ( ) ∗ ℎ( )
t
= ( ). ℎ( − )
−( − ) ( − )

= ( ). ℎ ( − )
t t
When t < 0, there is no overlap between ( ). ( − )
= ( ). ( ) ( − )
( ). ( − )

= ( ). ( − )
t
When t < 0, there is overlap between ( ). ( − )
from o to t and overlap amount is 1
= ( ) ( − )
( ). ( − )

= ( ) ( − )
t
Thus the limit of integration is changed from -∞ to + ∞ to 0 to t
= ×1

= [ ] = ( − )= ( − 1) ( )
Exercise 2.5(b), P-98, Srinath and Soliman
Use convolution integral where
( )= ( ) and ℎ( ) = ( )
( ) ( ) ( − )

Soln:
( ) = ( ) ∗ ℎ( )
t

= ( ). ℎ( − )
−( − ) ( − )

= ( ). ℎ ( − )
t t
When t < 0, there is no overlap between ( ). ( − )
= ( ). ( − )
( ). ( − )

= ×1
t
When t < 0, there is overlap between ( ). ( − )
= ( ) from o to t and overlap amount is 1

( ). ( − )
=− +

=− −
t
∴ ( ) = [− − ] Thus the limit of integration is changed from -∞ to + ∞ to 0 to t
= [− − ] − [−0 × − ]
= [− − + 1]. ( )
Exercise 2.5(c), P-98, Srinath and Soliman
Use convolution integral where

( )= ( ) + ( ) and ℎ( ) = ( )
Soln:
( ) = ( ) ∗ ℎ( )

= ( ). ℎ( − ) = ( ). ℎ( − )

= [ ( ) + ( )]. ( − )

= [ ( ). ( )] = ( ) ( − )

= ×1 = 1 = ×1 = 1

= −[ ] + [ ] = −( − ) + ( − 0)

= (− + 1 + ) ( ) = [1 + − ] ( )
Exercise 2.5(d), P-98, Srinath and Soliman
Use convolution integral where

( ) = ( ) and ℎ( ) = ( )+ ( )
Soln:
( ) = ( ) ∗ ℎ( ) = ℎ( ) ∗ ( ) [ ]

= ℎ( ). ( − ) = ℎ ( ). ( − )

= [ ( ) + ( )]. ( − )

= [ ( ). ( − ) + ( − ). ( )

= [ × 1] = ( − ) ( − 0)

1 1 1
=− [ ] + ( )= − + ( )+ ( )
2 2 2
1 1 1
= − + ( )+ ( )=− ( − 1) ( ) + ( )
2 2 2
Example 2.3.5 P-57 Srinath & Soliman
Find the impulse response ℎ( ) where
ℎ ( )= ( )
ℎ ( ) ℎ ( )
ℎ ( )=2 ( )
∑ ( )
ℎ ( )= ( )
( )
ℎ ( )=4 ( ) ℎ ( ) ℎ ( )

Solution: From the system diagram given above we can write

ℎ( ) = ℎ ( ) ∗ ℎ ( ) + ℎ ( ) ∗ ℎ ( )

Now ℎ ( ) ∗ ℎ ( ) = ℎ ( )ℎ ( − )

= ( )×2 ( ) ( − ) =2 ( ). ( − )

=2 ×1 =2

=2 [− ] = −2 ( − ) ( ) = 2( − ) ( )
Now ℎ ( ) ∗ ℎ ( ) = [ ( )] ∗ 4 ( )
We know that ( ) ∗ ( )= (
∴[ ( )] ∗ 4 ( ) = 4 ( )
∴ ℎ ( ) = 2( − ) ( )+4 ( )
Gg
Home Work
Exercise 2.14
Find the impulse response ℎ( ) where
ℎ ( )= ( ) ℎ ( )
ℎ ( )= ( )
ℎ ( )= ( )
ℎ ( )= ( )
( ) ℎ ( ) ℎ ( ) ℎ ( ) ( )
ℎ ( )= ( ) ∑

ℎ ( )

Hints: ℎ( ) = ℎ ( ) ∗ [ℎ ( ) + ℎ ( ) ∗ ℎ ( ) + ℎ ( )]
 Additional Reading Topic
Switched exponential signal

Switched signals are signals that change abruptly at certain times. When there is only one such change, the time
variable is often defined so that change happens at = 0. The simplest switched signal is unit step function, defined
as
1, 0
( )=
0, 0 (1)

( )

The unit step function undergoes an abrupt change from a constant value of zero to constant value of unity one at
= 0. Switched inputs are frequently used to model sources that changes from a state corresponding to a logical
value of “zero” to logical “one” to model disturbance that are suddenly applied or to model circuit behavior
immediately after a circuit is turned on.

A switched exponential signal is a signal of the following form:

, 0
( )=
0, 0 (2)
This function is illustrated below for a real and positive value of a. This function can also be written as ( ) =
( )

( )

Unit step function is a special case of switched exponential with a=0. Note that equation 1 and equation 2 are
interchangeable.