ECE 3331b

Introduction to Signal Processing

Lecture 5

Instructor: Dr. Ilia G. Polushin

Department of Electrical & Computer Engineering

Faculty of Engineering

The University of Western Ontario

Winter 2017
Schedule: Laboratories
Matlab tutorials: January 23rd January 27th

4 Lab sessions:

Lab 1: January 30th February 3rd

Lab 2: February 13th February 17th
Lab 3: March 13th March 17th
Lab 4: March 27th March 31st

Location: SEB 1004

Reports are due one week after the lab session. To be submitted to locker #
206, 2nd floor of TEB
Schedule: Laboratories

Lab Schedule by Section:

Section 003: Tuesday 9:30 a.m. 12:30 p.m. TAs: Farid, Anastasiia

Section 004: Thursday 4:30 p.m. 7:30 p.m. TAs: Soulaimane, Yousef
Section 005: Friday 12:30 p.m. 3:30 p.m. TAs: Samantha, Hessam
Section 006: Wednesday 2:30 p.m. 5:30 p.m. TAs: Samantha, Hessam

Section 007: Wednesday 11:30 a.m. 2:30 p.m. TAs: Farid, Anastasiia
Section 008: Monday 2:30 p.m. 5:30 p.m. TAs: Soulaimane, Yousef
Topic 2: Discrete-Time Signals and Systems

Discrete-Time Signals

Discrete-Time Systems

Analysis of Discrete-Time Linear Time-Invariant (LTI) Systems

Discrete-Time Systems Described by Dierence Equations

Implementation of Discrete-Time Systems

Discrete-Time Systems
A Discrete-Time System is a device or an algorithm that operates
on a DT signal (called the input or excitation) to produce another DT
signal (called the output or response).

Figure 2.2.1 from Proakis & Manolakis, 2007

General description:

y(n) := T [x(n)] ,

where x(n) is input signal, y(n) is output signal,

T is the transformation (operator, processing) that the system per-
forms on x(n) to produce y(n).
Discrete-Time Systems

Simple examples of DT systems:

Identity system:

y(n) = x(n).

Unit delay system:

y(n) = x(n 1).

Moving average filter:

1
y(n) = [x(n + 1) + x(n) + x(n 1)] .
3
Discrete-Time Systems
Example: Accumulator

n
X
y(n) = x(k) = x(n) + x(n 1) + x(n 2) + . . .
k= 1

Accumulator equation can be rewritten as follows:

n 1
X
y(n) = x(k) + x(n) = y(n 1) + x(n).
k= 1

In the above equations, y(n 1) is called the initial condition. It summarizes

the eect of all previous inputs.

The output y(n) is determined by current input x(n) and initial condition
y(n 1).

A system is said to be initially relaxed if the initial condition is zero (y(n

1) = 0).
Block Diagram Representation of DT Systems
Example of a block diagram representation:
Block Diagram Representation of DT Systems

Basic building blocks:

An adder:

A constant multiplier:
Block Diagram Representation of DT Systems
Basic building blocks:

A signal multiplier:

A unit delay element:

A unit advance element:

Block Diagram Representation of DT Systems

System:

y(n) = 0.25y(n 1) + 0.5x(n) + 0.5x(n 1)

Block diagram representation:

Classification of DT Systems

Static vs. Dynamic Systems

Time-Invariant vs. Time-Variant Systems

Linear vs. Nonlinear Systems

Causal vs. Noncausal Systems

Stable vs. Unstable Systems

Static vs. Dynamic Systems

A DT system is said to be static (or memoryless) if the output at any instant n depends
(at most) on the input at the current time n, but not on past or future values of the input.

In other words, in memoryless systems, y(n) may depend on x(n) but not on x(i) where
i 6= n.

Examples of static (memoryless) systems:

y(n) = ax(n),

y(n) = nx(n) + bx3 (n).

Static vs. Dynamic Systems

A DT system is called dynamic (or system with memory) if it is not

static.

In other words, a system is dynamic if y(n) depends on x(i) where i 6= n.

Examples of DT dynamic systems:

P
m
y(n) = x(n k), Accumulator with finite memory
k=0

P
1
y(n) = x(n k), Accumulator with infinite memory
k=0
Time-Invariant vs. Time-Variant Systems

Informal Definition: A system is called time-invariant if its input-output characteristics

do not change with time.
Suppose a response of a relaxed (i.e., with zero initial condition) system to input x(n) is
y(n).

x(n) y(n)
T
The system is time-invariant if for any k 2 Z, the response to input x(n k) is y(n k),

x(n k) y(n k)
T
The system is called time-variant or time-varying if it is not time-invariant; in other
words, if there exists at least one x(n) and at least one k 2 Z such that

y(n k) 6= T [x(n k)] .

Linear vs. Nonlinear Systems
A system is called linear if it satisfies the superposition principle; otherwise, the system
is nonlinear.
Given two arbitrary input sequences x1 (n), x2 (n), suppose the responses of an initially relaxed
(i.e., with zero initial condition) system to these inputs are y1 (n) and y2 (n), respectively:

x1 y1
T

x2 y2
T

The system is linear if and only if its response to input sequence 1 x1 (n) + 2 x2 (n), where
1, 2 R is 1 y1 (n) + 2 y2 (n)

1 x1 (n) + 2 x2 (n) 1 y1 (n) + 2 y2 (n)

T
 Linear vs. Nonlinear Systems
The superposition principle:

y(n) = T [a1 x1 (n) + a2 x2 (n)] = a1 T [x1 (n)] + a2 T [x2 (n)]

The system T is linear if y(n) y 0 (n) for any x1 (n), x2 (n), and any a1 , a2 2 R.
Linear vs. Nonlinear Systems
Example 2.2.5 (from Proakis & Manolakis, 2007): Determine if the following systems are
linear or nonlinear.

(a.) y(n) = nx(n)

Solution: Denote y1 (n) = nx1 (n), y2 (n) = nx2 (n). Then the response of the system to
input x(n) = 1 x1 (n) + 2 x2 (n) is:

y(n) = nx(n) = n (1 x1 (n) + 2 x2 (n)) = 1 nx1 (n) + 2 nx2 (n) = 1 y(n) + 2 y(n).

The system is therefore linear.

(b.) y(n) = x(n2 )

Solution: Denote y1 (n) = x1 (n2 ), y2 (n) = x2 (n2 ). Then the response of the system to input
x(n) = 1 x1 (n) + 2 x2 (n) is:

y(n) = x(n2 ) = 1 x1 (n2 ) + 2 x2 (n2 ) = 1 y(n) + 2 y(n).

The system is linear.

Linear vs. Nonlinear Systems
Example 2.2.5 (from Proakis & Manolakis, 2007): Determine if the following systems are
linear or nonlinear.

(c.) y(n) = x2 (n)

Solution: Denote y1 (n) = x21 (n), y2 (n) = x22 (n). The response of the system to input
x(n) = 1 x1 (n) + 2 x2 (n) is:
2
y(n) = x2 (n) = (1 x1 (n) + 2 x2 (n)) = 12 x21 (n) + 22 x22 (n) + 21 2 x1 (n)x2 (n)

6= 1 y(n) + 2 y(n).

The system is nonlinear.

(e.) y(n) = ex(n)

Solution: Denote y1 (n) = ex1 (n) , y2 (n) = ex2 (n) . The response of the system to input
x(n) = 1 x1 (n) + 2 x2 (n) is:

y(n) = ex(n) = e1 x1 (n)+2 x2 (n) = e1 x1 (n) e2 x2 (n) 6= 1 y(n) + 2 y(n).

The system is nonlinear.

Causal vs. Noncausal Systems

Informal Definition: A system is called causal if its current output does not depend on
future values of input. Otherwise, the system is noncausal.

Definition: A system is said to be causal if the output of the system y(n) at any given
time instant n depends only on present and past inputs x(n), x(n 1), x(n 2), . . . , but
does not depend on future inputs x(n + 1), x(n + 2), etc.

Mathematically, a causal system has a form

y(n) = F [x(n), x(n 1), x(n 2), . . .]

Real-time signal processing is always causal. If a signal is recorded and processed o-line
(not in real time), a noncausal SP algorithms can be implemented.
Causal vs. Noncausal Systems

Example 2.2.6 (from Proakis & Manolakis, 2007): Determine if the following
systems are causal or noncausal

(a.) y(n) = x(n) x(n 1). Answer: causal

P
n
(b.) y(n) = x(k). Answer: causal
k= 1

(c.) y(n) = ax(n). Answer: causal

(d.) y(n) = x(n) + 3x(n + 4). Answer: noncausal

(e.) y(n) = x (n2 ). Answer: noncausal

(g.) y(n) = x( n). Answer: noncausal

Stable vs. Unstable Systems

Definition: A (relaxed) system is said to be bounded input - bounded output (BIBO)

stable if every bounded input produces a bounded output. More precisely, for each Mx < 1
there exists My < 1 such that

|x(n)| Mx for all n ) |y(n)| My for all n.

Many other definitions of stability are possible.

Any practically useful system must be stable (at least in some sense).
Interconnection of DT Systems
Two basic types:

Cascade interconnection

Parallel interconnection
Interconnection of DT Systems
Cascade interconnection:

y(n) := Tc [x(n)] := T2 [T1 [x(n)]] = T2 T1 [x(n)] .

In general,

T2 T1 6= T1 T2

However, if both T1 , T2 are linear and time-invariant, then

T2 T1 = T1 T2 .
Interconnection of DT Systems
Parallel interconnection:

y(n) := y1 (n) + y2 (n) = T1 [x(n)] + T2 [x(n)] = (T1 + T2 ) [x(n)] = Tp [x(n)] ,

where

Tp [x(n)] = T1 + T2
Topic 2: Discrete-Time Signals and Systems

Discrete-Time Signals

Discrete-Time Systems

Analysis of Discrete-Time Linear Time-Invariant (LTI) Systems

Discrete-Time Systems Described by Dierence Equations

Implementation of Discrete-Time Systems

Linear vs. Nonlinear Systems
A system is called linear if it satisfies the superposition principle; otherwise, the system
is nonlinear.
Given two arbitrary input sequences x1 (n), x2 (n), suppose the responses of an initially relaxed
(i.e., with zero initial condition) system to these inputs are y1 (n) and y2 (n), respectively:

x1 y1
T

x2 y2
T

The system is linear if and only if its response to input sequence 1 x1 (n) + 2 x2 (n), where
1, 2 R is 1 y1 (n) + 2 y2 (n)

1 x1 (n) + 2 x2 (n) 1 y1 (n) + 2 y2 (n)

T
Time-Invariant vs. Time-Variant Systems
Informal Definition: A system is called time-invariant if its input-output characteristics
do not change with time.
Suppose a response of a relaxed (i.e., with zero initial condition) system to input x(n) is
y(n).

x(n) y(n)
T
The system is time-invariant if for any k Z, the response to input x(n k) is y(n k),

x(n k) y(n k)
T
Analysis of DT Linear Time-Invariant Systems
How to calculate a response of LTI system using superposition principle:

Decomposition of the input signal into a linear combination of elementary signals

X
x(n) = ck xk (n).
k

Calculation of the systems response to each elementary signal

xk yk
T 8k

Application of the superposition principle to find the total response

X X
x(n) = ck xk (n) y(n) = ck yk (n)
k k
T
Analysis of DT Linear Time-Invariant Systems
Step 1: Decomposition of the input signal into a sum of elementary signals:
X
x(n) = ck xk (n).
k

How to choose elementary signals xk (n)?

One convenient choice is

xk (n) := (n k),

where (n k) is a unit sample sequence delayed by k steps.

Why is the above choice convenient?

One reason is that it is straightforward to represent any DT signal x(n) in the form

X
x(n) = ck (n k).
k
Analysis of DT Linear Time-Invariant Systems
Let x(n), n 2 {. . . , 1, 0, 1, 2, . . .} be an arbitrary DT signal. Let x(k) is the value of x(n)
at k-th instant.

Figure 2.3.1 from Proakis &

Manolakis

The signal x(k) (n k) is equal to x(n) at k-th instant and equal to zero everywhere else.
Analysis of DT Linear Time-Invariant Systems

Let x(n), n Z := {. . . , 1, 0, 1, 2, . . .}, be an arbitrary DT signal. Let x(k) is the value of

x(n) at k-th instant.
The signal x(k) (n k) is equal to x(n) at k-th instant and equal to zero everywhere else.
The sum of x(k) (n k) for all possible delays < k < is equal to the original signal
x(n),

+1
X
x(n) = x(k) (n k), nZ
k= 1
Analysis of DT Linear Time-Invariant Systems

Example 2.3.1 (from Proakis & Manolakis, 2007): Consider a finite-duration se-
quence
x(n) = {2, 4, 0, 3}
"
Resolve the sequence x(n) into a sum of weighted impulse sequences.

Answer:
x(n) = 2 (n + 1) + 4 (n) + 3 (n 2).
Analysis of DT Linear Time-Invariant Systems
How to calculate a response of LTI system using superposition principle:

Decompose the input signal x(n) into a linear combination of (n k), k Z,

+1
X
x(n) = x(k) (n k)
k= 1

Calculate the systems response to elementary input (n),

h(n) := T [ (n)] ,
(n) h(n)
T
Due to time invariance, the response of the system to the delayed elementary input
(n k) is h(n k), for each k Z.

h(n k) := T [ (n k)] ,
(n k) h(n k)
T
Analysis of DT Linear Time-Invariant Systems
Applying the superposition principle, we have
+1
X +1
X
x(n) = x(k) (n k) y(n) = x(k)h(n k)
k= 1 k= 1
T

P
+1
The response of an initially relaxed system to input x(n) = x(k) (n k) is
k= 1

" +1
# +1 +1
X X X
y(n) = T [x(n)] = T x(k) (n k) = x(k)T [ (n k)] = x(k)h(n k),
k= 1 k= 1 k= 1

where h(n) = T [ (n)] is the response of a relaxed LTI system to the unit sample sequence.
The function
h(n) := T [ (n)]
is called the impulse response. It represents the response of the relaxed LTI system to the
unit sample sequence (n).
Analysis of DT Linear Time-Invariant Systems

Convolution sum:

+1
X
y(n) = x(k)h(n k) (CS)
k= 1

where h(n) = T [ (n)] is the response of a relaxed LTI system to the unit sample sequence.
The impulse response function is the response of a relaxed LTI system to the unit sample
sequence,
h(n) := T [ (n)]

Formula (CS) allows to compute a response of a relaxed LTI system to an arbitrary input
signal x(n) based on systems impulse response.