LTI (LSI) systems

Bharath Shekar
IIT Bombay
 Material

• https://ocw.mit.edu/resources/res-6-008-digital-
signal-processing-spring-2011/video-lectures/
lecture-2-discrete-time-signals-and-systems-part-1/

• https://www.youtube.com/watch?v=TuCYGjp7WKU

• Digital signals and systems

• LTI systems
 Digital (or discrete) signals
Height of the matchsticks represent the amplitude of the signal
as a function of sample number (always an integer)

Oppenheim & Schafer, Chapter 2

 t : sampling interval
n : sample index, always an integer
Discrete signal X[n] = Xa (n t) Analog signal

Oppenheim & Schafer, Chapter 2

 A few basic
[n] sequences

u[n]

exp[ a n]

sin[n + a]

Oppenheim & Schafer, Chapter 2

 Representation of signals

Any sequence can be represented as sum of scaled, shifted

delta sequences

1
X
x[n] = xk [n k]
k= 1

Oppenheim & Schafer, Chapter 2

 Systems and signals

A system takes a sequence as input and produces another

sequence as output

y[n] = T {x[n]}

Oppenheim & Schafer, Chapter 2

Systems and signals: example

y[n] = T {x[n]}

y[n] = x[n] this system does nothing to input

y[n] = x[n nd ] what does this do?

Oppenheim & Schafer, Chapter 2

Systems and signals: example
y[n] = x[n nd ] what does this do?
 Causal system

output value at some instant n depends on input at

the same and/or previous instants n

y[n] = x[n nd ]
 Linear system

A system y[n] = T {x[n]} is said to be linear if:

y1 [n] = T {x1 [n]}, y2 [n] = T {x1 [n]}

Say,
x3 [n] = c1 x1 [n] + c2 x2 [n]

then the system satisfies:

T {x3 [n]} = T {c1 x1 [n] + c2 x2 [n]}

= c1 T {x1 [n]} + c2 T {x2 [n]}

i.e. y3 [n] = c1 y1 [n] + c2 y2 [n]

Exercise: are the following systems linear?
Use the definition from the previous slide..

y[n] = x[n] y[n] = x[a n]

y[n] = x[n nd ] y[n] = m x[n] + c

 Time-shift Invariant (TI) system

A system y[n] = T {x[n]} is said to be TI if

a delay in input causes the same delay in output

Say x1 [n] = x[n k]

Then, T{} is TI if:

y1 [n] = T {x1 [n]}
= T {x[n k]}
= y[n k]
Exercise: are the following systems TI?
Use the definition from the previous slide..

y[n] = x[n] y[n] = x[a n]

y[n] = x[n nd ] y[n] = m x[n] + c

 LTI system
A system that is both linear and time-shift
invariant

Say, x3 [n] = c1 x1 [n k1 ] + c2 x2 [n k2 ]

then, y3 [n] = c1 y1 [n k1 ] + c2 y2 [n k2 ]

Observe how definitions of linearity and time-shift invariance

have been incorporated.

Question: is the wave equation LTI?

 LTI system and impulse response
An LTI system can be completely characterised
by its impulse response

[n] LTI h[n]

Impulse
Impulse
response

h[n] = T { [n]}
 LTI system and impulse response
Given the impulse response of LTI system, we
can work out output for an arbitrary input, i.e.

Given:

[n] LTI h[n]

we can find:

x[n] LTI y[n]

 LTI system and impulse response
Hint: use representation of general sequence
x[n].. see slide 9

1
X
x[n] = xk [n k]
k= 1

short hand for x[k]

Exercise: derive an expression for y[n]

LTI system and impulse response

1
X
x[n] = xk [n k] Representation
k= 1
of any sequence

y[n] = T {x[n]}
n X 1 o
=T xk [n k]
k= 1
LTI system and impulse response

y[n] = T {x[n]}
n X 1 o
=T xk [n k]
k= 1
1
X
= T {xk [n k]} Linearity
k= 1
 LTI system and impulse response
y[n] = T {x[n]}
n X 1 o
=T xk [n k]
k= 1
1
X
= T {xk [n k]}
k= 1
X1
Linearity
= xk T { [n k]}
k= 1

Impulse response:
[n] LTI h[n]
LTI system and impulse response:
convolution
y[n] = T {x[n]}
n X 1 o
=T xk [n k]
k= 1
1
X
= T {xk [n k]}
k= 1
X1
Linearity
= xk T { [n k]}
k= 1
1
X
= xk h[n k] Time-shift
k= 1 invariance
 LTI systems are characterised by impulse response.
For an arbitrary input the output is the convolution of
the input with the impulse response

Given: [n] LTI h[n]

we can find:

x[n] LTI y[n]

1
X
y[n] = x[k]h[n k] convolution
k= 1
y =x⇤h convolution is denoted by *
Geophone is an LTI system with respect to wave
propagation

Recorded data =
instrument (impulse) response
⇤ ground motion
 Further reading

• Oppenheim & Schafer: “Discrete time signal

processing”, parts of chapter 2

• Gubbins: “Time series analysis and inverse theory

for geophysicists”

• Yilmaz: chapter 2

• Steven Smith: “Digital Signal Processing .. ”