Tex as A& M University

ECEN 314– Spring 2022

Signals and Systems

Module 3

Signals and Systems in the Time Domain

Learning Objectives

 Be familiar with the form of some common signals.

 Understand what a delta function is and some of its basic properties.
 Be able to form basic operations on signals, most importantly, flipping,
shifting and scaling.
 Be able to decompose a signal into even and odd parts.
 Know the 6 basic properties of systems and be able to tell if a given
system possesses those properties.
 Understand what the convolutional integral is and how it relates the
input and output of an LTI system.
 Some Common Signals
A=2, f=1Hz, = /3

Sine Wave (sinusoid) 2

𝑥𝑥 𝑡𝑡 = 𝐴𝐴cos(𝜔𝜔𝜔𝜔 + 𝜃𝜃) 1

𝐴𝐴 = Amplitude
𝜔𝜔 = frequency (in rad./sec)
0

x(t)
𝜔𝜔 = 2𝜋𝜋𝜋𝜋, 𝑓𝑓 = frequency (in Hz) -1

𝜃𝜃 = phase (in radians) -2

0 0.5 1 1.5 2 2.5 3 3.5 4

t (sec)

Square-Wave
Described by:
T=1sec --> f=1Hz, Duty Cycle=33.9%, HighLevel=3, LowLevel=-1, Offset=0.5sec

 𝑇𝑇 = period
3

1
 𝑓𝑓 = = frequency (in Hz)
2

𝑇𝑇
 Duty cycle = % of time in high 1

state 0

 High level
 Low level
-1

 Offset 0 0.5 1 1.5 2 2.5 3 3.5 4

 Some Common Signals
Unit Step Function 𝑡𝑡
1, 𝑡𝑡 > 0, 𝑢𝑢(𝑡𝑡) � 𝑢𝑢 𝜏𝜏 𝑑𝑑𝑑𝑑
𝑢𝑢 𝑡𝑡 = �
0, 𝑡𝑡 < 0. −∞

Unit Ramp Function 𝑡𝑡 𝑡𝑡

𝑡𝑡
𝑡𝑡, 𝑡𝑡 ≥ 0,
� 𝑢𝑢 𝜏𝜏 𝑑𝑑𝜏𝜏 = �
−∞
0, 𝑡𝑡 < 0.
sgn(𝑡𝑡)

Signum (sign) Function

1, 𝑡𝑡 > 0, 𝑡𝑡
sgn 𝑡𝑡 = �
−1, 𝑡𝑡 < 0.

Rectangle Function
1, 𝑡𝑡 < 1/2,
rect 𝑡𝑡 = � rect(𝑡𝑡) tri(𝑡𝑡)
0, 𝑡𝑡 > 1/2.

Triangle Function
1 1 𝑡𝑡 −1 1 𝑡𝑡
1 − |𝑡𝑡|, 𝑡𝑡 < 1, −
tri 𝑡𝑡 = � 2 2
0, 𝑡𝑡 > 1.
Delta Functions
One of the waveforms that plays a One way to view the delta function is
fundamental role in the study of as an infinitely tall, infinitely narrow
signals and systems is the Dirac Delta pulse with a unit area underneath.
Function a.k.a. the unit impulse 1/∆
function, 𝛿𝛿(𝑡𝑡).

𝑡𝑡
Another way to view a delta function −∆/2 ∆/2
is as the derivative of the unit step

∆→ 0
function.
𝑢𝑢(𝑡𝑡)
𝛿𝛿(𝑡𝑡)
1
𝑑𝑑/𝑑𝑑𝑑𝑑 1
𝑡𝑡
𝑡𝑡
Delta Functions
One of the important properties of delta functions is known as the sifting property.
Consider evaluating the following integral for some arbitrary function 𝑥𝑥(𝑡𝑡).
𝑏𝑏
∫𝑎𝑎 𝛿𝛿 𝑡𝑡 − 𝑡𝑡𝑜𝑜 𝑥𝑥 𝑡𝑡 𝑑𝑑𝑑𝑑 (assume that 𝑎𝑎 < 𝑡𝑡𝑜𝑜 < 𝑏𝑏)

𝑑𝑑 Using the tall thin pulse interpretation,

Using the 𝛿𝛿 𝑡𝑡 − 𝑡𝑡𝑜𝑜 = 𝑢𝑢(𝑡𝑡 − 𝑡𝑡𝑜𝑜 )
𝑑𝑑𝑑𝑑
the integrand 𝛿𝛿 𝑡𝑡 − 𝑡𝑡𝑜𝑜 𝑥𝑥 𝑡𝑡 is 0
interpretation:
𝑏𝑏 everywhere except at 𝑡𝑡 = 𝑡𝑡𝑜𝑜 and is
𝑑𝑑
� 𝑢𝑢(𝑡𝑡 − 𝑡𝑡𝑜𝑜 )𝑥𝑥 𝑡𝑡 𝑑𝑑𝑑𝑑 equivalent to 𝛿𝛿 𝑡𝑡 − 𝑡𝑡𝑜𝑜 𝑥𝑥 𝑡𝑡𝑜𝑜 so that
𝑎𝑎 𝑑𝑑𝑑𝑑 𝑏𝑏
𝑏𝑏 � 𝛿𝛿 𝑡𝑡 − 𝑡𝑡𝑜𝑜 𝑥𝑥 𝑡𝑡 𝑑𝑑𝑑𝑑
𝑏𝑏 𝑑𝑑𝑑𝑑
= 𝑥𝑥(𝑡𝑡) 𝑢𝑢(𝑡𝑡 − 𝑡𝑡𝑜𝑜 )� − � 𝑢𝑢 𝑡𝑡 − 𝑡𝑡𝑜𝑜 𝑑𝑑𝑑𝑑 𝑎𝑎
𝑎𝑎 𝑎𝑎 𝑑𝑑𝑑𝑑 𝑏𝑏
Assuming that 𝑎𝑎 < 𝑡𝑡𝑜𝑜 < 𝑏𝑏, = � 𝛿𝛿 𝑡𝑡 − 𝑡𝑡𝑜𝑜 𝑥𝑥 𝑡𝑡𝑜𝑜 𝑑𝑑𝑑𝑑
𝑏𝑏 𝑎𝑎
𝑑𝑑𝑑𝑑 𝑏𝑏
= 𝑥𝑥 𝑏𝑏 − � 𝑑𝑑𝑑𝑑
𝑡𝑡𝑜𝑜 𝑑𝑑𝑑𝑑 = 𝑥𝑥 𝑡𝑡𝑜𝑜 � 𝛿𝛿 𝑡𝑡 − 𝑡𝑡𝑜𝑜 𝑑𝑑𝑑𝑑 = 𝑥𝑥 𝑡𝑡𝑜𝑜
𝑎𝑎
= 𝑥𝑥 𝑏𝑏 − 𝑥𝑥 𝑏𝑏 − 𝑥𝑥 𝑡𝑡𝑜𝑜
Since the area underneath the delta
= 𝑥𝑥 𝑡𝑡𝑜𝑜
function is one.
Operations on Signals
Time reversal
𝑥𝑥(𝑡𝑡) 𝑥𝑥(−𝑡𝑡)
𝑥𝑥(𝑡𝑡) → 𝑥𝑥(−𝑡𝑡) time
reversal
𝑡𝑡 𝑡𝑡
Time scaling
𝑥𝑥(𝑡𝑡) → 𝑥𝑥(𝑎𝑎𝑎𝑎) 𝑥𝑥(𝑡𝑡) 𝑥𝑥(𝑎𝑎𝑎𝑎)
time
0 < 𝑎𝑎 < 1 → expand in time 𝑡𝑡 scaling 𝑡𝑡
1 1/𝑎𝑎
𝑎𝑎 > 1 → compress in time

Time shifting
𝑥𝑥(𝑡𝑡) → 𝑥𝑥(𝑡𝑡 − 𝑡𝑡𝑜𝑜 ) 𝑥𝑥(𝑡𝑡) 𝑥𝑥(𝑡𝑡 − 𝑡𝑡𝑜𝑜 )
time
𝑡𝑡𝑜𝑜 > 0 → shift right 𝑡𝑡 shifting 𝑡𝑡
𝑡𝑡𝑜𝑜 < 0 → shift left 1 𝑡𝑡𝑜𝑜
𝑡𝑡𝑜𝑜 + 1
Examples
Sketch each of the following
a) rect(𝑡𝑡/6)
b) rect(𝑡𝑡 + 2)
c) rect(3𝑡𝑡 + 2)
d) rect(2 − 𝑡𝑡/2)
Operations on Signals
Given a signal 𝑥𝑥(𝑡𝑡), a time transformation of the form 𝑥𝑥(𝑎𝑎𝑎𝑎 − 𝑏𝑏) can be written as
one of two equivalent two step operations.

Shift then scale Scale then shift

Step 1: Shift by 𝑏𝑏: 𝑥𝑥(𝑡𝑡) → 𝑥𝑥(𝑡𝑡 − 𝑏𝑏) Step 1: Scale by 𝑎𝑎: 𝑥𝑥(𝑡𝑡) → 𝑥𝑥(𝑎𝑎𝑎𝑎)
Step 2: Scale by 𝑎𝑎: 𝑥𝑥(𝑡𝑡 − 𝑏𝑏) → x(𝑎𝑎𝑎𝑎 − 𝑏𝑏) Step 2: Shift by 𝑏𝑏/𝑎𝑎:
𝑏𝑏
𝑥𝑥 𝑎𝑎𝑡𝑡 → 𝑥𝑥 𝑎𝑎 𝑡𝑡 − = 𝑥𝑥(𝑎𝑎𝑎𝑎 − 𝑏𝑏)
𝑎𝑎
Example:
Suppose 𝑥𝑥(𝑡𝑡) is as shown.
(a) Sketch 𝑥𝑥(2𝑡𝑡 − 1). Try it both ways.
(b) Then sketch 𝑥𝑥(1 − 2𝑡𝑡).
𝑥𝑥(𝑡𝑡)

1 2 3 𝑡𝑡
 Even and Odd Functions
rect(𝑡𝑡) tri(𝑡𝑡)
A function 𝑥𝑥 𝑡𝑡 is even if 𝑥𝑥 𝑡𝑡 = 𝑥𝑥(−𝑡𝑡).
 If you flip it horizontally (time reverse it)
you get the same thing. 1 1 𝑡𝑡 −1 1
−
2 2
A function 𝑥𝑥 𝑡𝑡 is odd if 𝑥𝑥 𝑡𝑡 = −𝑥𝑥(−𝑡𝑡).
 If you flip it horizontally (time reverse it) sgn(𝑡𝑡)
and vertically (negate it) you get the 𝑡𝑡
same thing.
𝑥𝑥(𝑡𝑡)
For any function 𝑥𝑥(𝑡𝑡) that is not necessarily
either even or odd, note that 𝑡𝑡
𝑥𝑥 𝑡𝑡 +𝑥𝑥(−𝑡𝑡)
 𝑥𝑥𝑒𝑒 𝑡𝑡 = is even, 𝑥𝑥𝑒𝑒 (𝑡𝑡)
2
𝑥𝑥 𝑡𝑡 −𝑥𝑥(−𝑡𝑡)
 𝑥𝑥𝑜𝑜 𝑡𝑡 = is odd, 𝑡𝑡
2

 𝑥𝑥 𝑡𝑡 = 𝑥𝑥𝑒𝑒 𝑡𝑡 + 𝑥𝑥𝑜𝑜 𝑡𝑡 . 𝑥𝑥𝑜𝑜 (𝑡𝑡)

 Any signal can be decomposed into the 𝑡𝑡

sum of an even part and an odd part.
 Example
Decompose the signals shown into and even and odd parts.
𝑥𝑥(𝑡𝑡) 𝑦𝑦(𝑡𝑡)
2 1
... ...
1
𝑡𝑡
−1 1 −2 −1 1 2 3 4 𝑡𝑡
 Systems
In general, a system is a device that takes 𝑥𝑥(𝑡𝑡) 𝑦𝑦(𝑡𝑡)
an input (signal) and produces an output System
(signal). The following are properties that
can be used to characterize/classify systems.  Stability – A system is stable if for
any 𝑥𝑥(𝑡𝑡) such that 𝑥𝑥(𝑡𝑡) < 𝑐𝑐1 , then
 Memory – A system has memory if 𝑦𝑦(𝑡𝑡𝑜𝑜 ) 𝑦𝑦(𝑡𝑡) < 𝑐𝑐2 , for any finite constants
𝑐𝑐1 and 𝑐𝑐2 .  bounded inputs
depends on values of 𝑥𝑥(𝑡𝑡) for 𝑡𝑡 ≠ 𝑡𝑡𝑜𝑜 . If
𝑦𝑦(𝑡𝑡𝑜𝑜 ) depends only on 𝑥𝑥(𝑡𝑡𝑜𝑜 ) then the produce bounded outputs.
system is memoryless.
 Time-Invariance – A system is time
invariant if for any input 𝑥𝑥(𝑡𝑡) that
 Invertibility – A system is invertible if
each distinct input 𝑥𝑥(𝑡𝑡) produces a produces an output 𝑦𝑦(𝑡𝑡) then the
distinct output 𝑦𝑦(𝑡𝑡).  two different input 𝑥𝑥(𝑡𝑡 − 𝑡𝑡𝑜𝑜 ) produces 𝑦𝑦(𝑡𝑡 − 𝑡𝑡𝑜𝑜 ).
 Time shifted inputs produce time
inputs do not produce the same output.
shifted outputs.
 Causality – A system is causal if 𝑦𝑦(𝑡𝑡𝑜𝑜 )
depends on values of 𝑥𝑥(𝑡𝑡) only for 𝑡𝑡 ≤ 𝑡𝑡𝑜𝑜 .  Linearity – If 𝑥𝑥1 (𝑡𝑡) → 𝑦𝑦1 (𝑡𝑡) and
 Current output depends on current 𝑥𝑥2 (𝑡𝑡) → 𝑦𝑦2 (𝑡𝑡), the system is linear if
𝑎𝑎𝑥𝑥1 𝑡𝑡 + 𝑏𝑏𝑥𝑥2 (𝑡𝑡) → 𝑎𝑎𝑦𝑦1 𝑡𝑡 + 𝑏𝑏𝑦𝑦2 (𝑡𝑡).
and past inputs, not future inputs.
 Superposition holds
 Example
Suppose a system has an input/output relationship given by
𝑡𝑡
𝑦𝑦 𝑡𝑡 = ∫−∞ 𝑥𝑥 𝑢𝑢 𝑑𝑑𝑑𝑑.
This is known as an integrator. Determine if this system possesses any of the six
properties listed on the previous slide.
 Example
Suppose a system has an input/output relationship given by
𝑦𝑦 𝑡𝑡 = 𝑥𝑥 2 (𝑡𝑡).
Determine if this system possesses any of the six properties of systems.
 An Important Class of Systems
Suppose a system has an input/output relationship described by a differential
equation
𝑑𝑑𝑚𝑚 𝑦𝑦 𝑑𝑑𝑛𝑛 𝑥𝑥
∑𝑀𝑀 𝑏𝑏
𝑚𝑚=0 𝑚𝑚 𝑚𝑚 = ∑𝑁𝑁
𝑎𝑎
𝑛𝑛=0 𝑛𝑛 𝑛𝑛 .
𝑑𝑑𝑡𝑡 𝑑𝑑𝑡𝑡

The output of the system can be broken into two parts:

 Natural response – Due to initial conditions (found assuming 𝑥𝑥 𝑡𝑡 = 0),
 Forced Response – Due to the input (found assuming all ICs are 0).
For most of this course, we will concern ourselves with the forced response of such
systems and will assume that the initial conditions are all zero.

It can be shown that such a system is linear and time-invariant (LTI). The other 4
properties (memory, invertibility, causality, stability) may or may not hold depending
on the coefficients of the ODE.
For a general LTI system, it is possible to determine the output of the system in
terms of the input. The derivation follows.
 Output of LTI Systems
To start the derivation, we approximate the input 𝑥𝑥(𝑡𝑡)
with a series of rectangular pulses. Let 𝑝𝑝∆ 𝑡𝑡 =
1 𝑡𝑡 1
𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟  rectangle of height , width ∆ and area
∆ ∆ ∆
𝑡𝑡
1. Then
∞

𝑥𝑥 𝑡𝑡 ≅ � 𝑥𝑥 𝑘𝑘∆ � 𝑝𝑝∆ 𝑡𝑡 − 𝑘𝑘∆ � ∆.

𝑘𝑘=−∞
Note that as ∆→ 0,
 The approximation becomes exact, and 𝑝𝑝∆ 𝑡𝑡 ℎ∆ (𝑡𝑡)
 𝑝𝑝∆ 𝑡𝑡 → 𝛿𝛿(𝑡𝑡). System

Let ℎ∆ (𝑡𝑡) be the response (output) of the system to ∆→ 0

an input of 𝑝𝑝∆ 𝑡𝑡 . Also, let ℎ 𝑡𝑡 = lim ℎ∆ (𝑡𝑡) so that
∆→0
ℎ 𝑡𝑡 is the response of the system to a unit impulse 𝛿𝛿 𝑡𝑡 ℎ(𝑡𝑡)
at the input. We call ℎ 𝑡𝑡 the impulse response of System
the system.
 Output of LTI Systems
Starting with 𝑝𝑝∆ 𝑡𝑡 → ℎ∆ 𝑡𝑡 , then using the properties of linearity and time
invariance:
𝑝𝑝∆ 𝑡𝑡 − 𝑘𝑘∆ → ℎ∆ 𝑡𝑡 − 𝑘𝑘∆ (due to time invariance)

∆ � 𝑥𝑥 𝑘𝑘∆ � 𝑝𝑝∆ 𝑡𝑡 − 𝑘𝑘∆ → ∆ � 𝑥𝑥 𝑘𝑘∆ � ℎ∆ 𝑡𝑡 − 𝑘𝑘∆ (due to linearity)

∑∞ ∞
𝑘𝑘=−∞ ∆ � 𝑥𝑥 𝑘𝑘∆ � 𝑝𝑝∆ 𝑡𝑡 − 𝑘𝑘∆ → ∑𝑘𝑘=−∞ ∆ � 𝑥𝑥 𝑘𝑘∆ � ℎ∆ 𝑡𝑡 − 𝑘𝑘∆ (due to linearity)

Passing to the limit as ∆→ 0, the input becomes 𝑥𝑥(𝑡𝑡) and the output becomes
∞
𝑦𝑦 𝑡𝑡 = � 𝑥𝑥 𝑢𝑢 ℎ 𝑡𝑡 − 𝑢𝑢 𝑑𝑑𝑑𝑑 .
−∞

This is known as a convolution integral and we often use the shorthand notation
𝑦𝑦 𝑡𝑡 = 𝑥𝑥 𝑡𝑡 ∗ ℎ(𝑡𝑡).

This result tells us that if we can determine the impulse response of a system, we
can then determine the response of the system to any arbitrary input.

The next set of notes will elaborate on how to evaluate convolution integrals.