1

CLASS 2

(Sections 1.3)

Exponential and Sinusoidal Signals

∙ They arise frequently in applications, and many other signals can be constructed from them.

Continuous-time complex exponential and sinusoidal signals:

𝑥(𝑡) = 𝐶𝑒𝑎𝑡

where 𝐶 and 𝑎 are in general complex numbers.

Real exponential signals: 𝐶 and 𝑎 are reals.

C>0 and a>0. C>0 and a<0.

at

at
Ce

Ce

C C
0 0
0 0
t t

∙ The case 𝑎 > 0 represents exponential growth. Some signals in unstable systems exhibit exponential

growth.

∙ The case 𝑎 < 0 represents exponential decay. Some signals in stable systems exhibit exponential

decay.
 2

Periodic complex exponential:

𝑒𝑗𝑤0 𝑡

√
where 𝑗 = −1, 𝑤0 ∕= 0 is real, and 𝑡 is the time.

Euler’s formula: 𝑒𝑗𝑤0 𝑡 = cos(𝑤0 𝑡) +𝑗 sin(𝑤0 𝑡) . Note that

| {z } | {z }
Re{𝑒𝑗𝑤0 𝑡 } Im{𝑒𝑗𝑤0 𝑡 }

1
jw t
sin(w t) e 0
0
Im

0 cos(w0t)

−1
−1 0 1
Re

∙ ∣𝑒𝑗𝑤0 𝑡 ∣ = 1 and ∠𝑒𝑗𝑤0 𝑡 = 𝑤0 𝑡.

∙ 𝑒𝑗2𝜋𝑘 = 1, for 𝑘 = 0, ±1, ±2, . . . .

Since
( )
2𝜋 𝑤
𝑗𝑤0 𝑡+ ∣𝑤 𝑗2𝜋 ∣𝑤0 ∣
𝑒 ∣ 0 = 𝑒𝑗𝑤0 𝑡 𝑒 0 = 𝑒𝑗𝑤0 𝑡 𝑒|𝑗2𝜋sign(𝑤0)
{z } = 𝑒
𝑗𝑤0 𝑡

=1

we have
2𝜋
𝑒𝑗𝑤0 𝑡 is periodic with fundamental period .
∣𝑤0 ∣

Note that

∙ 𝑒𝑗𝑤0 𝑡 and 𝑒−𝑗𝑤0 𝑡 have the same fundamental period.

∫∞ ∫∞
∙ Energy in 𝑒𝑗𝑤0 𝑡 : −∞
∣𝑒𝑗𝑤0 𝑡 ∣𝑑𝑡 = −∞
1.𝑑𝑡 = ∞
1
∫𝑇 1
∫𝑇
∙ Average Power in 𝑒𝑗𝑤0 𝑡 : lim𝑇 →∞ 2𝑇 −𝑇
∣𝑒𝑗𝑤0 𝑡 ∣𝑑𝑡 = lim𝑇 →∞ 2𝑇 −𝑇
1.𝑑𝑡 = 1.
2𝜋
∙ {𝑒𝑗𝑘𝑤0 𝑡 }𝑘=0,±1,... , are all periodic with period ∣𝑤0 ∣
. They are called a harmonically related set of

complex exponentials with 𝑒𝑗𝑘𝑤0 𝑡 being the 𝑘−th harmonic.

 3

Sinusoidal signals:

𝐴 cos(𝑤0 𝑡 + 𝜙) and 𝐴 sin(𝑤0 𝑡 + 𝜙).

where 𝐴 is real, 𝑤0 is real, 𝜙 is real, and 𝑡 is the time. (Graph one of the signals!)

∙ They arise in systems that conserve energy such as an ideal LC circuit or an ideal mass-spring system.

∙ – Periodic with the same fundamental period 𝑇0 = 2𝜋/∣𝑤0 ∣

– ∣𝑤0 ∣ is the fundamental frequency

– 𝑓0 := 1/𝑇0 = ∣𝑤0 ∣/(2𝜋) is the number of cycles per unit time (large 𝑓0 means more oscillatory)

– ∣𝐴∣ is the amplitude

– ∣𝜙∣ is the size of the phase shift.

∙ Since

𝑒𝑗(𝑤0 𝑡+𝜙) = cos(𝑤0 𝑡 + 𝜙) + 𝑗 sin(𝑤0 𝑡 + 𝜙)

we can write

𝐴 cos(𝑤0 𝑡 + 𝜙) = 𝐴Re(𝑒𝑗(𝑤0 𝑡+𝜙) )

𝐴 sin(𝑤0 𝑡 + 𝜙) = 𝐴Im(𝑒𝑗(𝑤0 𝑡+𝜙) ).

∙ Recall, for any complex number 𝑧,

𝑧 = Re(𝑧) + 𝑗Im(𝑧) 𝑧 ∗ = Re(𝑧) − 𝑗Im(𝑧)

therefore
𝑧 + 𝑧∗ 𝑧 − 𝑧∗
Re(𝑧) = Im(𝑧) = .
2 2𝑗

Hence, we can also write

𝐴 ( 𝑗(𝑤0 𝑡+𝜙) ( 𝑗(𝑤0 𝑡+𝜙) )∗ ) 𝐴 ( 𝑗(𝑤0 𝑡+𝜙) )

𝐴 cos(𝑤0 𝑡 + 𝜙) = 𝑒 + 𝑒 = 𝑒 + 𝑒−𝑗(𝑤0 𝑡+𝜙)
2 2
𝐴 𝐴
= 𝑒𝑗𝜙 𝑒𝑗𝑤0 𝑡 + 𝑒−𝑗𝜙 𝑒−𝑗𝑤0 𝑡
2 2
𝐴 𝑗(𝑤0 𝑡+𝜙) ( 𝑗(𝑤0 𝑡+𝜙) )∗ ) 𝐴 −𝑗𝜋/2 ( 𝑗(𝑤0 𝑡+𝜙)
( )
𝐴 sin(𝑤0 𝑡 + 𝜙) = 𝑒 − 𝑒 = 𝑒 𝑒 − 𝑒−𝑗(𝑤0 𝑡+𝜙)
2𝑗 2
𝐴 𝐴
= 𝑒𝑗(𝜙−𝜋/2) 𝑒𝑗𝑤0 𝑡 − 𝑒−𝑗(𝜙+𝜋/2) 𝑒−𝑗𝑤0 𝑡 .
2 2
 4

General complex exponential signals:

𝐶𝑒𝑎𝑡

where 𝐶 and 𝑎 are complex numbers.

If

𝐶 = ∣𝐶∣𝑒𝑗𝜃 and 𝑎 = 𝑟 + 𝑗𝑤0

then

𝐶𝑒𝑎𝑡 = ∣𝐶∣𝑒𝑗𝜃 𝑒(𝑟+𝑗𝑤0 )𝑡 = ∣𝐶∣𝑒𝑟𝑡 𝑒𝑗(𝑤0 𝑡+𝜃) = ∣𝐶∣𝑒𝑟𝑡 cos(𝑤0 𝑡 + 𝜃) +𝑗 ∣𝐶∣𝑒𝑟𝑡 sin(𝑤0 𝑡 + 𝜃) .
| {z } | {z }
Re(𝐶𝑒𝑎𝑡 ) Im(𝐶𝑒𝑎𝑡 )

r>0. r<0.

|C|ert |C|ert
Re(Ce )

Re(Ce )
at

at

0 0

−|C|ert −|C|e
rt

0 0
t t

∙ If 𝑟 = 0, the real and imaginary part are sinusoidals.

∙ If 𝑟 > 0, the real and imaginary part are sinusoidals multiplied by a growing exponential.

Such signals arise in unstable systems.

∙ If 𝑟 < 0, the real and imaginary part are sinusoidals multiplied by a decaying exponential.

Such signals arise in stable systems, for example, in RLC circuits, or in mass-spring-friction system,

where the energy is dissipated due to the resistors, friction, etc.

 5

Discrete-time complex exponential and sinusoidal signals:

𝑥[𝑛] = 𝐶𝑒𝛽𝑛

where 𝐶 and 𝛽 are complex numbers.

Analogous to the continuous-time case with the following differences: (𝑤0 is real below)

∙ 𝑒𝑗𝑤0 𝑡 = 𝑒𝑗𝑤1 𝑡 are different signals if 𝑤0 ∕= 𝑤1 , whereas

𝑒𝑗𝑤0 𝑛 = 𝑒𝑗𝑤1 𝑛 if 𝑤0 − 𝑤1 = 2𝑘𝜋, for some 𝑘 ∈ {0, ±1, . . . }.

(Explain this on the unit circle!)

Therefore, it is sufficient to consider only the case 𝑤0 ∈ [0, 2𝜋) or 𝑤0 ∈ [−𝜋, 𝜋).

∙ As 𝑤0 increases 𝑒𝑗𝑤0 𝑛 oscillates at higher frequencies, whereas this is not the case for 𝑒𝑗𝑤0 𝑛 .

In the figure below, the frequency of oscillations increases as 𝑤0 changes from 0 to 𝜋 then it decreases

as 𝑤0 changes from 𝜋 to 2𝜋.

∙ 𝑒𝑗𝑤0 𝑡 is periodic with fundamental period 2𝜋/∣𝑤0 ∣, whereas

𝑒𝑗𝑤0 𝑛 is periodic ⇔ 𝑒𝑗𝑤0 𝑛 = 𝑒𝑗𝑤0 (𝑛+𝑀 ) for some integer 𝑀 > 0, for all 𝑛

⇔ 𝑒𝑗𝑤0 𝑀 = 1 for some integer 𝑀 > 0

⇔ 𝑤0 𝑀 = 2𝜋𝑚 for some integers 𝑚, 𝑀 > 0

𝑤0
⇔ is rational.
2𝜋
𝑤0 𝑚
∙ If 2𝜋
= 𝑀
for some integers 𝑚 and 𝑀 which have no common factors, then the fundamental period
2𝑚𝜋
is 𝑀 = 𝑤0
because
2𝜋𝑚
𝑒𝑗𝑤0 (𝑛+𝑁 ) = 𝑒𝑗𝑤0 𝑛 𝑒𝑗 𝑀
𝑁
.

The same observations hold for discrete-time sinusoids.

 6

w =0π/8 w =1π/8 w =2π/8

0 0 0

n=4
n=5 n=3 n=2
n=6 n=2 n=3 n=1
Im n=7 n=1

Im

Im
n=0 n=8 n=0 n=4 n=0
n=9 n=15
n=10 n=14 n=5 n=7
n=11n=13
n=12 n=6
Re Re Re

w =4π/8 w =8π/8 w =12π/8

0 0 0

n=1 n=3
Im

Im

Im
n=2 n=0 n=1 n=0 n=2 n=0

n=3 n=1
Re Re Re

w0=14π/8 w0=15π/8 w0=16π/8

n=6 n=12
n=11n=13
n=5 n=7 n=10 n=14
n=9 n=15
Im

Im

Im

n=4 n=0 n=8 n=0 n=0

n=7 n=1
n=3 n=1 n=6 n=2
n=2 n=5 n=3
n=4
Re Re Re

Fig. 1. To determine the fundamental period, count the number of steps to get back to 1!
 7

Examples:

1) Is 𝑥[𝑛] = 𝑒𝑗𝑛2𝜋/3 + 𝑒𝑗𝑛3𝜋/4 periodic? If it is periodic, what’s its fundamental period?

For 𝑒𝑗𝑛2𝜋/3 , 𝑤0 /(2𝜋) = 1/3, so 𝑒𝑗𝑛2𝜋/3 is periodic with fundamental period 3.

For 𝑒𝑗𝑛3𝜋/4 , 𝑤0 /(2𝜋) = 3/8, so 𝑒𝑗𝑛3𝜋/4 is periodic with fundamental period 8.

𝑥[𝑛] is periodic with fundamental period 24 = 𝑙𝑐𝑚(3, 8).

2) Is 𝑥[𝑛] = sin(3𝑛/4) periodic? If it is periodic, what’s its fundamental period?

𝑤0 3
Since 2𝜋
= 8𝜋
is irrational, 𝑥[𝑛] is not periodic; see the figure where 𝑥[𝑛] = 0 only at 𝑛 = 0.

sin(3t/4) and sin(3n/4)

1

−1
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
t

3) Is 𝑥[𝑛] = sin(8𝜋𝑛/31) periodic? If it is periodic, what’s its fundamental period?

Since 𝑤0 /(2𝜋) = 4/31, 𝑥[𝑛] is periodic with fundamental period 31; see the figure where 𝑥[0] =

𝑥[31] = 0. Note that the continuous-time signal sin(8𝜋𝑡/31) has fundamental period 31/4, hence it

is 0 at 𝑡 = 31/4. But 𝑥[𝑛] has no 31/4−th sample and it misses 0 between 𝑥[7] and 𝑥[8].

sin(8π t/31) and sin(8π n/31)

1

−1
0 1 2 3 4 5 6 7 8 9 1011121314151617181920212223242526272829303132333435
t
 8

Harmonically related discrete-time periodic exponentials:

𝜙𝑘 [𝑛] = {𝑒𝑗𝑘(2𝜋/𝑁 )𝑛 }𝑘=0,±1,... , are all periodic with period 𝑁 .

However, unlike the continuous-time signals, these signals are not all distinct because

𝜙𝑘+𝑁 [𝑛] = 𝑒𝑗(𝑘+𝑁 )(2𝜋/𝑁 )𝑛 = 𝑒𝑗𝑘(2𝜋/𝑁 )𝑛 𝑒𝑗2𝜋𝑛 = 𝜙𝑘 [𝑛].

This implies that there are only 𝑁 distinct signals in this set, for example,

𝜙0 [𝑛] = 1

𝜙1 [𝑛] = 𝑒𝑗2𝜋𝑛/𝑁

𝜙2 [𝑛] = 𝑒𝑗4𝜋𝑛/𝑁
..
.

𝜙1 [𝑛] = 𝑒𝑗2(𝑁 −1)𝜋/𝑁 .