www.getmyuni.

com

UNIT 2
Time-domain representations for LTI systems – 1
2.1 Introduction:
The Linear time invariant (LTI) system:

Systems which satisfy the condition of linearity as well as time invariance are known as linear time
invariant systems. Throughout the rest of the course we shall be dealing with LTI systems. If the
output of the system is known for a particular input, it is possible to obtain the output for a number
of other inputs. We shall see through examples, the procedure to compute the output from a given
input-output relation, for LTI systems.

Example – I:

2.1.1 Convolution:

A continuous time system as shown below, accepts a continuous time signal x(t) and gives out
a transformed continuous time signal y(t).

Figure 1: The continuous time system

Some of the different methods of representing the continuous time system are:

i) Differential equation
ii) Block diagram
iii) Impulse response
iv) Frequency response
v) Laplace-transform
www.getmyuni.com
vi) Pole-zero plot

It is possible to switch from one form of representation to another, and each of the representations
is complete. Moreover, from each of the above representations, it is possible to obtain the system
properties using parameters as: stability, causality, linearity, invertibility etc. We now attempt to
develop the convolution integral.

2.2 Impulse Response

The impulse response of a continuous time system is defined as the output of the system when its
input is an unit impulse, δ (t) . Usually the impulse response is denoted by h(t) .

Figure 2: The impulse response of a continuous time system

2.3 Convolution Sum:

We now attempt to obtain the output of a digital system for an arbitrary input x[n], from
the knowledge of the system impulse response h[n].
www.getmyuni.com

Methods of evaluating the convolution sum:

Given the system impulse response h[n], and the input x[n], the system output y[n], is
given by the convolution sum:

Problem:
To obtain the digital system output y[n], given the system impulse response h[n], and the
system input x[n] as:
www.getmyuni.com
1. Evaluation as the weighted sum of individual responses
The convolution sum of equation (…), can be equivalently represented as:

y[n] ฀ ฀ .....฀ ฀
x [฀ 1]h[n ฀ 1]฀ ฀x [0]h[ n]฀ ฀x[1]h[n ฀ 1]฀ .......
www.getmyuni.com

Convolution as matrix multiplication:

Given
www.getmyuni.com

Evaluation using graphical representation:

Another method of computing the convolution is through the direct computation of each value of the
output y[n]. This method is based on evaluation of the convolution sum for a single value of n, and
varying n over all possible values.
www.getmyuni.com
www.getmyuni.com
www.getmyuni.com

Evaluation from direct convolution sum:

While small length, finite duration sequences can be convolved by any of the above three methods,
when the sequences to be convolved are of infinite length, the convolution is easier performed by
direct use of the „convolution sum‟ of equation (…).

Example: A system has impulse response h[n] ฀ ฀ exp(฀ 0.n8)u[n] . Obtain the unit step
response.
Solution:
www.getmyuni.com
www.getmyuni.com
www.getmyuni.com
Evaluation from Z-transforms:

Another method of computing the convolution of two sequences is through use of Z-transforms. This
method will be discussed later while doing Z-transforms. This approach converts convolution to
multiplication in the transformed domain.

Evaluation from Discrete Time Fourier transform (DTFT):

It is possible to compute the convolution of two sequences by transforming them to the frequency
domain through application of the Discrete Fourier Transform. This approach also converts the
convolution operator to multiplication. Since efficient algorithms for DFT computation exist, this
method is often used during software implementation of the convolution operator.

Evaluation from block diagram representation:

While small length, finite duration sequences can be convolved by any of the above three methods,
when the sequences to be convolved are of infinite length, the convolution is easier performed by
direct use of the „convolution sum‟ .
www.getmyuni.com

2.4 Convolution Integral:

We now attempt to obtain the output of a continuous time/Analog digital system for an arbitrary
input x(t), from the knowledge of the system impulse response h(t), and the properties of the impulse
response of an LTI system.

Methods of evaluating the convolution integral: (Same as Convolution sum)

Given the system impulse response h(t), and the input x(t), the system output y(t), is given by the
convolution integral:

Some of the different methods of evaluating the convolution integral are: Graphical representation,
Mathematical equation, Laplace-transforms, Fourier Transform, Differential equation, Block
diagram representation, and finally by going to the digital domain.
www.getmyuni.com

Recommended Questions
1. Show that if x(n) is input of a linear time invariant system having impulse response h(n), then
the output of the system due to x(n) is

y(n) 
x(k)h(n  k)
k  
2. Use the definition of convolution sum to prove the following properties
1. x(n) * [h(n)+g(n)]=x(n)*h(n)+x(n)*g(n) (Distributive Property)
2. x(n) * [h(n)*g(n)]=x(n)*h(n) *g(n) (Associative Property)
3. x(n) * h(n) =h(n) * x(n) (Commutative Property)
3. Prove that absolute summability of the impulse response is a necessary condition for stability
of a discrete time system.
4. Compute the convolution y(t)= x(t)*h(t) of the following pair of signals:

5. Compute the convolution sum y[n] =x[n]* h[n] of the following pairs of sequences:

6. Show that if y (t) =x(t)* h(t), then

7. Let y[n] = x[n]* h[n]. Then show that

8. Show that

for an arbitrary starting point no.