< Previous Lecture Next Lecture >

Class 2 (Continuous-Time and Discrete-Time Sig

Types of signals

Continuous / Discrete

A continuous-time signal x(t) is represented by an uncountably infinite numbe

variable points (e.g., an uncountably infinite number of values across time).

A discrete-time signal x[n] is represented by an countable (potentially infinite, but

number of dependent variable values (e.g., five values across time).

Notation warning: Some texts use the same bracket notation for both continuous-tim
time signals (i.e., x(t) is continuous-time and x(n) is discrete-time). Furthermore, t
discrete-time using m (common in controls literature) rather than n (common in s
literature). In this course, we will denote continuous-time signal with round b
discrete-time signals by square brackets [⋅] . Further, we will typically use t to den
time signals and n to denote discrete-time signals. Note though that the dependent v
need to be time. It could be anything -- time, space, speed, force, etc.

Even / Odd

The continuous-time signals x(t) is even when

x(−t) = x(t)

The continuous-time signals x(t) is odd when

x(−t) = −x(t)

The discrete-time signals x[n] is even when

x[−n] = x[n]

The discrete-time signals x[n] is odd when

x[−n] = −x[n]
Causal / Acausal

A continuous-time signal x(t) is said to be causal when

x(t) = 0 for t < 0

A continuous-time signal x(t) is said to be anticausal when

x(t) = 0 for t ≥ 0

A discrete-time signal x[n] is said to be causal when

x[n] = 0 for n < 0

A discrete-time signal x[n] is said to be anticausal when

x[n] = 0 for n ≥ 0

Signals that are not causal are also called acausal. Anticausal signals are are a type a
Periodic / Aperiodic

A continuous-time signal x(t) is said to be periodic if some T0 exists such that

x(t) = x(t + mT0 )

where T0 is finite and m can equal any integer value m = … − 2, −1, 0, 1, 2, ….

A discrete-time signal x[n] is said to be periodic if some N0 exists such that

x[n] = x[n + mN0 ]

where N0 is finite and m can equal any integer value m = … − 2, −1, 0, 1, 2, ….

Fundamental period/frequency: The smallest T0 or N0 that satisfies the abov

known as the fundamental period of the signal. The reciprocal value (f0 = 1/T0 o
known as the fundamental frequency of the signal. We will discuss how to find t
frequency in the next lecture.

Caveat for discrete-time signals: Note that determining periodicity for discrete
more complicated than it may initially seem. Some signals that are periodic in con
not satisfy periodicity in discrete-time (for example, x[n] = cos(n) does not satisfy
property).
Measures of signal "size"

Energy

The energy of a continuous-time signal x(t)

∞
2
Ex = ∫ |x(t)| dt
−∞

The energy of a discrete-time signal x[n] is

∞
2
Ex = ∑ |x[n]|

n=−∞

An energy signal is a signal with finite energy (i.e., 0 < Ex < ∞ ).

Physical Interpretation: Energy, in this context, does not refer to a specific ph

Instead, it describes the "size" of a signal. Our energy, however, can be related to elec
x(t) is a voltage signal across a load of resistance R , then the energy supplied to tha

Power (sometimes referred to as average power)

The power of a continuous-time signal x(t)

T /2
1 2
Px = lim ∫ |x(t)| dt .
T →∞ T −T /2

When the signal is periodic, the power simplifies to

T /2
1 2
Px = ∫ |x(t)| dt ,
T −T /2
 where T is a period of the periodic signal. This is equivalent to saying that the pow
signal is equal to the average energy in one period in the signal.

The power of a discrete-time signal x[n] is

N
1 2
Px = lim ∑ |x[n]| .
N →∞ 2N + 1
n=−N

When the signal is periodic, the power simplifies to

N −1
1 2
Px = ∑ |x[n]| ,
N
n=0

where N is a period of the periodic signal. This is equivalent to saying that the pow
signal is equal to the average energy in one period in the signal.

A power signal is a signal with finite power (i.e., 0 < Px < ∞ ).

Physical Interpretation: Power, in this context, does not refer to a specific ph

Instead, it describes the "size" of a periodic signal. Our power, however, can be rela
power. If x(t) is a voltage signal across a load of resistance R , then the power supp
is Px /R.

Important signal operations

Time shifting

The continuous-time signal y(t) = x(t − T ) is the signal x(t) shifted to the right by

The continuous-time signal y(t) = x(t + T ) is the signal x(t) shifted to the left by T

The discrete-time signal y[n] = x[n − N ] is the signal x[n] shifted to the right by N

The discrete-time signal y[n] = x[n + N ] is the signal x[n] shifted to the left by N s
 Time scaling

The continuous-time signal y(t) = x(at) is the signal x(t) condensed by a factor of a

The continuous-time signal y(t) = x(t/a) is the signal x(t) expanded by a factor of a

The discrete-time signal y[n] = x[an] is the signal x[n] condensed by a factor of a.

The discrete-time signal y[n] = x[n/a] is the signal x[n] expanded by a factor of a.

Time reversal

The continuous-time signal y(t) = x(−t) is the time-reversed signal of x(t) .

The discrete-time signal y[n] = x[−n] is the time-reversed replica of signal x[n] .

Discrete-time Power: From Limit Representation to Per

Representation

Let N0 be defined as the fundamental period of a periodic signal. Using the assumption
periodic, the limit representation of discrete-time power can be simplified to
 N
1 2
Px = lim ∑ |x[n]|
N →∞ 2N + 1
n=−N

N0
1 2
2N 2
= lim [|x[0]| + ∑ |x[n]| ]
N →∞ 2N + 1 N0
n=1

We broke up the summation into two terms. The first term is the value of the signal at n
term is the energy in one period multiplied by , which is the ratio of the total ran
2N

N0

and (i.e., ) and the fundamental period. That is, for any given , there are
2N
N 2N N
N0

−N and N . Note that this simplification is not true for any N but is true for the
approaches infinity.

To complete our proof, we distribute the 1/(2N + 1) so that

N0 2
1 2 2N 2
|x[0]| 2N
lim [|x[0]| + ∑ |x[n]| ] = lim +
N →∞ 2N + 1 N0 N →∞ 2N + 1 N0 (2N + 1)
n=1

When we evaluate the limit, the first term in the above equation goes to zero as N go
second term, after applying L'Hopsital's rule, converges to

N0
1 2
∑ |x[n]| .
N0
n=1

So, the final solution is

N0
1 2
Px = ∑ |x[n]|
N0
n=1

N0 −1
1 2
= ∑ |x[n]|
N0
n=0

Periodic signals

The fundamental period

For a continuous-time, periodic signal, the fundamental period is the smallest T0 va

x(t) = x(t + mT0 )

for any integer value m . That is, this is a small length of time for which the signal rep

For a discrete-time, periodic signal, the fundamental period is the smallest N0 valu

x[n] = x[n + mN0 ]

for any integer value of m . Again, this is a small length of time for which the signal r
this value must be an integer. If there are no integer-valued periods, the signal is not
The fundamental frequency

For a continuous-time or discrete-time, periodic signal, the fundamental fre

reciprocal of the fundamental period, such that

f0 = 1/T0 or f0 = 1/N0 .

The fundamental frequency specifices such that all of frequencies in the signal
where m is some m is some integer. These frequencies are known as harmonics

For a continuous-time or discrete-time, periodic signal, the fundamental angula

defined by

ω0 = 2πf0 .

Angular frequency is notationally convenient since the functions cos(ω1 t) and sin

angular frequencies of ω1 and ω2 , respectively.

Computing the fundamental period for a sum of periodic signals

For a single periodic signal, the fundamental period is the smallest period in the sign
the fundamental period of x(t) = cos(2πt) is T0 = 1. For a sum of periodic signa
the fundamental period is more complicated.

The fundamental periodT0 of x1 (t) + x2 (t) + x3 (t) + … + xN (t)is the least com
of the individual periods T1 , T2 , T3 , … , TN. Conceptually, this implies that the fund
is the first common period in each component signal.

Computing a Least Common Multiple (for integers)

To determine T0 = LCM(T1 , T2 , … , TN ) , where T1 , T2 , … , TN are integers:

1. Perform prime factorization on each period

Example 1: 120 = 2 × 2 × 2 × 3 × 5
Example 2: 102 = 2 × 3 × 17

2. Combine equal terms in each factorization (i.e., 120 = 23 × 3 × 5)

Example 1: 120 = 2 × 3 × 5
3

Example 2: 102 = 2 × 3 × 17

3. The least common multiple is the product of the highest-power unique factors
 Example: LCM(120, 102) = 23 × 3 × 5 × 17 = 2040

Computing a Least Common Multiple (for rational numbers)

To determine , where and

n1 n2 nN
LCM ( , ,…, ) n1 , n2 , … , nN d1 , d2 , … , dN
d1 d2 dN

compute

n1 n2 nN LCM (n1 , n2 , … , nN )
LCM ( , ,…, ) =
d1 d2 dN GCD (d1 , d2 , … , dN )

See the next section on how to compute the greatest common divisor (GCD).

Computing the fundamental frequency for a sum of periodic sig

The fundamental frequency f0 of x1 (t) + x2 (t) + x3 (t) + … + xN (t) is the gre

divisor of the individual frequencies f1 , f2 , f3 , … , fN. Similarly, the fundam
frequency ω0 of x1 (t) + x2 (t) + x3 (t) + … + xN (t) is the greatest common
individual angular frequencies ω1 , ω2 , ω3 , … , ωN.

Computing a Greatest Common Divisor (for integers)

1. Perform prime factorization on frequencies

Example 1: 120 = 2 × 2 × 2 × 3 × 5
Example 2: 102 = 2 × 3 × 17

2. The greatest common divisor is the product of the common factors

Example: GCD(120, 102) = 2 × 3 = 6

Computing a Greatest Common Divisor (for rational numbers)

To determine , where and

n1 n2 nN
GCD ( , ,…, ) n1 , n2 , … , nN d1 , d2 , … , dN
d1 d2 dN

compute

n1 n2 nN GCD (n1 , n2 , … , nN )
GCD ( , ,…, ) =
d1 d2 dN LCM (d1 , d2 , … , dN )
See the previous section on how to compute the least common multiple (LCM).

Special functions / signals

Impulse functions

The Dirac δ signal (continuous-time impulse signal) is defined by

∞ for t = 0
δ(t) = {
0 for t ≠ 0

where
∞

∫ δ(t)dt = 1
−∞

The Kronecker δ signal (discrete-time impulse signal) is defined by

1 for n = 0
δ[n] = {
0 for n ≠ 0

The properties of impulse signals:

Energy: Ex = not well defined
Power: Px = 0
Even / Odd?: Even (Understanding why is beyond the purview of this course. You
tested on this.)
Periodic?: No
Causal?: Yes
 Heaviside step functions

The continuous-time step function u(t) is defined by

t
1 for t ≥ 0
u(t) = ∫ δ(τ )dτ = {
−∞ 0 for t < 0

The discrete-time step function u[n] is defined by

n
1 for n ≥ 0
u[n] = ∑ δ[k] = {
0 for n < 0
k=−∞

The properties of the Heaviside step functions:

Energy: Ex = ∞
Power: Px = 1/2
Even / Odd?: Neither
Periodic?: No
Causal?: Yes

Other Important Signals

Cosines and sines

The continuous-time cosine and sine signals are defined by

x1 (t) = cos(ω0 t) , x2 (t) = sin(ω0 t)

The properties of the continuous-time cosines or sines:

Enegry: Ex = ∞
Power: Px = 1/2
Even / Odd?: cos(ω0 t) is even and sin(ω0 t) is odd
Periodic?: Yes
Fundamental frequency: f0 = ω0 /(2π)
Fundamental period: T0 = 1/f0

Causal?: No
Stepped exponential

A continuous-time exponential signal is defined by

a0 t
x(t) = e u(t)

The properties of the exponential signals are:

∞ for a0 ≥ 0
Enegry: Ex = {
−1/(2a0 ) for a0 < 0

Power: Px = 0
Even / Odd?: No
Periodic?: No
Causal?: Yes

Complex exponentials

The continuous-time complex exponential signal is defined by

(jω0 +a0 )t
x(t) = e

= cos(ω0 t) + j sin(ω0 t)

The properties of the continuous-time complex exponential:

Enegry: Ex = ∞
Power: Px = 1
Even / Odd?: No, but its components are: cos(ω0 t) is even and sin(ω0 t) is odd
Periodic?: Yes
 Fundamental frequency: f0 = ω0 /(2π)
Fundamental period: T0 = 1/f0

Causal?: No

Stepped general exponential

The continuous-time general exponential signal is defined by

(jω0 +a0 )t
x(t) = e u(t)

a0 t a0 t
= [e cos(ω0 t) + je sin(ω0 t)] u(t)

The properties of the general exponential signals are:

∞ for a0 ≥ 0
Enegry: Ex = {
−1/(2a0 ) for a0 < 0

Power: Px = 0
Even / Odd?: No
Periodic?: No
Causal?: Yes

Input-Output system models

Input-output system model

Throughout the course, we will visualize systems using block diagrams, such as the o
 Block diagram illustrating a generic system

In this diagram, the x(t) is the input, which the system manipulates, and y(t) is
general systems, we can represent a system by a function H{⋅} that operates on si
the system above can be formally expressed as

y(t) = H{x(t)}.

Note that this notation will disappear when we start focusing on linear, time-invar
class.

System properties

Continuous-time or discrete-time

A continuous-time system whose inputs and outputs are continuous-time signals.

system is one whose inputs and outputs are discrete-time signals.

Linear or nonlinear

A system is linear if it obeys the law of superposition. Formally, if H{⋅} is a system w

or x2 (t) and output y2 (t) or y1 (t) such that

y1 (t) = H{x1 (t)}

y2 (t) = H{x2 (t)},

then the system is linear if

y(t) = H{ax1 (t) + bx2 (t)}

= aH{x1 (t)} + bH{x2 (t)}

= ay1 (t) + by2 (t),

where a and b are arbitrary scalar numbers.

 Block diagrams illustrating the linearity property

Time invariant or time varying

A system is time-invariant if the system does not change with time. Formally, if H{
time t with input x(t) and output y(t), such that y(t) = H{x(t)}, then the
invariant if

y(t + τ ) = H{x(t + τ )}

for any arbitrary time delay τ . That is, if we delay our input, we expect the output t
the same amount. In a time-invariant system, an input delay may not affect the outp
other way.
Block diagrams illustrating input and output for a time-invariant system.

Block diagrams illustrating input and output for a non-time-invariant

system.

Memoryless or with memory

A system is instantaneous (or memoryless) if its output at time t only depends on th

t. A system is dynamic (or has memory) if the output on t depends on the input at

previous inputs t − τ for τ > 0. That is, the system "remembers" previous input
Block diagrams illustrating input and output for a memoryless system.

Block diagrams illustrating input and output for a system with memory.

Causal or non-causal

A system is causal (or physical or non-anticipative) if its output at time t only

current input at time t and its past outputs at time t − τ for τ > 0. Hence, the
depend on future inputs.

All real-time systems are causal systems.

Block diagrams illustrating input and output for a causal system.

Block diagrams illustrating input and output for a non-causal system.

Bounded-input, bounded-output (BIBO) stable or unstable

 A system is BIBO stable if any amplitude-bounded input (i.e., the signal x(t) n
approaches ∞ for all t) yields and amplitude-bounded output (i.e., the signal y(t) n
approaches ∞ for all t).

Note that there are many different types of stability criteria for systems. We will dis
much greater detail after learning about the Laplace transform.

Built by Joel B. Harley, Assistant Professor, Univeristy of Florida