1 Signals PDF

Note: All physical signals are real.
But often we
can use complex numbers to represent a signal
since
Part 1. SIGNALS & SYSTEMS

1.1. SIGNALS
Signals as functions of time
They capture properties of the signal in a

more compact form
1. Signal: a function of certain variables
They are convenient for analysis (as we will

see through the course)
a function of time: price of stocks

a function of spatial coordinations: brightness &
color of an image as a function of space
We are dealing with: FUNCTIONS OF ONE
VARIABLE and the variable is time.
2. Interpretations of time:
Discrete time: The signal is a function of integers.
Our notation:
x[n] : . . . x[1], x[0], x[1], x[2] . . .
Domain of discrete signals: Z
Continuous time: The signal is a function of real
numbers x(t).
Domain of continuous signals: R
The range (value of the signal x) of both could be real
R or complex C.
ES156 Harvard SEAS
ES156 Harvard SEAS
Signals with finite power
Signal Energy and Power

In continuous time:
Example: Electrical circuits

i(t)
P (t)
E[t1 ,t2 ]
V (t)2
= V (t)i(t) =
R
Z t2
=
P (t)dt
Example:
Energy is defined as
Average power of the signal in this period is
Average power is defined as
E[n1 ,n2 ] =
t1
t2
t1
E =
ES156 Harvard SEAS
|x[n]|
n2
X
1
2
|x[n]|
n2 n1 + 1 n=n
|x(t)| dt
The total energy of x[n] is
Signals with finite energy

+
n2
X
n=n1
We are interested in 2 classes of signals:
x(t) = A sin(t)
Define the energy of a signal x(t) in a period [t1 , t2 ] as

Z t2
2
E[t1 ,t2 ] =
|x(t)| dt
|x(t)| dt <
In discrete time:
t1
1
t2 t 1
+T
Finite energy signals must have zero average power.

Otherwise
P > 0 E =
V (t)
R
1
= lim
T 2T
E = lim
|x(t)| dt <
N
X
n=N
|x[n]|
The signal has finite total energy if E < .

3
ES156 Harvard SEAS
The average total power of the signal is
Transformation of Signals
n
X
1
2
|x[m]| = P
lim
n 2n + 1
m=n
1. Time shift
x(t) x(t t0 )
The signal has finite average power if P < .
time shift or translation of the signal by t0

seconds
Examples:
2. Time reversal
E <
1
n
n 6= 0
x[n]
X
1
= 2
=
n2
3
n=1
x(t) x(t)
3. Time scaling
n=0
x(t) , > 0
compression by a factor of if > 1
expansion by a factor of if < 1
E = , 0 < P <
x[n]
= (1)n for n in Z
4. Combinations of the above

x(t + ) = x((t + /))
translation by /
an expansion (or contraction) by ||
n
X
1
lim
1=1
n 2n + 1
m=n
reversion if is negative.
Note: For any signal (real or complex), the energy

and power are always real.
ES156 Harvard SEAS
ES156 Harvard SEAS
Even and Odd Signals
Periodic Signals
A continuous signal x(t) is periodic with period
T0 > 0 if
x(t + T0 ) = x(t)
Signals also could be ODD or EVEN

A signal is odd if
x(t) = x(t) , or x[n] = x[n]
for all t R.
T0 is the fundamental period of x(t) if T0 is the
smallest period of x(t).
A signal is even if
x(t) = x(t) , or x[n] = x[n]
Example: ej0 t , where 0 is a real positive

2
number has T0 =
.
0
Examples:
Same definition holds for discrete signals. If N0 > 0 is

an integer such that
Odd signal: x(t) = t

Even signal: x(t) = |t|.
x[n + N0 ] = x[n]
Remark: Any signal can be written as a sum of an

odd and an even signal, as
x(t)
for all n Z, then N0 is a period of x[n]
x(t) + x(t) x(t) x(t)

+
2
2
= xeven (t) + xodd (t)
If N0 is the smallest number with this property,

N0 is a fundamental period.
Remark:
xeven (t): even part of x(t),
For discrete signals if there is a period, there is

always a fundamental period. This is not
necessarily true for the continuous case.
xodd (t): odd part of x(t).
Example: x(t) is a constant.

ES156 Harvard SEAS
ES156 Harvard SEAS
Exponential and Sinusoidal Functions

Examples:
x(t) = cos 0 t is periodic with period
A general continuous exponential function is

2
0
x(t) = Cet
x[n] = cos 0 n could be periodic, an example is
0 = 2, but it could be aperiodic if 0 = 2.
where C and are complex numbers.

Write C = |C|ej , = r + j0 , then
Cet
=
=
=
|C|ej ert ej0 t
|C|ert ej(0 t+)
|C|ert cos(0 t + ) + j|C|ert sin(0 t + )
Special cases
0 = 0 and = 0: real exponential (aperiodic)
x(t) = |C|ert
r = 0: constant signal
r > 0: growing exponential
r < 0: decaying exponential.
r = 0: signal is periodic with fundamental period
T0 = 2/0
ejw0 t+ = cos(w0 t + ) + j sin(w0 t + )
0 is called the fundamental frequency
ES156 Harvard SEAS
ES156 Harvard SEAS
10
Harmonically related complex exponentials are

those with a common period T0 .
Discrete Exponentials
Definition
Harmonics of a periodic exponential signal

occur at integer multiples of its fundamental
frequency.
ejk0 t ,
x[n] =
=
Cen = |C|ern ej(0 n+)
|C|ern [cos(0 n + ) + j sin(0 n + )]
Remarks:
k = 1, 2, . . .
ej0 n = ej(0 +2)n , so we only deal with

0 0 2 or 0 .
rt
r > 0 and x(t) = e cos(0 t + ): signal is an

expanding sinusoidal.
0 close to 0 and 2 are low frequencies
rt
r < 0 and x(t) = e cos(0 t + ): signal is an

decaying sinusoidal.
0 close to are high frequencies

ej0 n is not necessarily periodic. It is periodic
with period N if and only if
expanding sinusoidal
ert
0 N = 2m , m Z
In other words,
rt
is a rational number
m
N.
There are only N distinct complex exponentials

x[n] with period N. (these are harmonics of each
other)
decaying sinusoidal
If x[n] = Cej0 n with period N, then

0
0 1
N 1
2 = N , N , . . . , N .
rt
While in the continuous case, if x(t) = e(j0 t)

4 8
with period T , then 0 could be 2
T , T , T ,...
rt
ES156 Harvard SEAS
0
2
11
ES156 Harvard SEAS
12
Unit Impulse and Unit Step Functions
Continuous Impulse Function
1 if n = 0
Unit impulse function: [n] =
0 otherwise
1 if n 0
Unit step function: u[n] =
0 if n < 0
No measurement device is accurate enough to have a

resolution of arbitrary accuracy. This forces us to
think of signals that are very short in time duration
and have impulse behavior.
The continuous impulse function is an abstraction
of this concept.
Note that u[n] u[n 1] = [n], so [n] is the

derivative of u[n].
4x[n] =
x[n] x[n 1]
n (n 1)
Applying to the unit step function

n
X
m=
(u[m] u[m 1]) = u[n] =
n
X
4 (t) =
[m]
m=
k=0
1
4,
0,
0t4
otherwise
So to a capacitor, 4 (t) would deliver 1 unit of

charge, no matter how small 4 is. The ideal
capacitor charge is given by:
[n k]
Sampling effect:
q(t)
x[n][n n0 ] = x[n0 ][n n0 ] , for any n0 Z

ES156 Harvard SEAS
Example: If 4 (t) were an electric current, it

R +
would deliver 1 unit of charge: 4 (t) = 1.
Let m = n k (k is a new dummy variable), then

u[n] =
Definition of discrete derivative:
Specifically, consider the function 4 (t)
13
ES156 Harvard SEAS
14
We define the unit impulse function (also known as

Diracs delta function) as (t) = lim40 4 (t).
0, t 6= 0
(t) = lim 4 (t) =
, t = 0
40
Multiplication: Provided f (t) continuous at

t = t0 ,
f (t)(t t0 ) = f (t0 )(t t0 )
In particular,
(t) is a function that is zero for t < 0, jumps to

at t = 0, and immediately drops to zero at
t > 0 and has area under its curve
Z
(t)dt = 1.
f (t)(t) = f (0)(t)
(t) is the shortest current into a capacitor that

carries one unit of charge.
Good properties of the () function:
Integration: Assuming f is continuous at t
Z +
f ( )(t )d = f (t)
Scaling:
Time-reversal:
ES156 Harvard SEAS
(at) =
1
(t)
|a|
(t) = (t)
15
ES156 Harvard SEAS
16
Unit Step Function
Unit ramp function
Rt
Recall that 4 (t) was given by the following
function, which we call u4 (t):
The unit ramp function is defined as
t, t 0
r(t) =
0, t < 0
It is the integral of the unit step function

Z t
u( )d
r(t) =
We define the unit step function u(t) to be
1, t > 0
u(t) =
0, t < 0
u(t)
Note that
4 (t) =
u04 (t)
and u4 (t) =
In the limit as 4 0, we have:

du(t)
and u(t) =
(t) =
dt
ES156 Harvard SEAS
4 ( )d
( )d
17
ES156 Harvard SEAS
18

1 Signals PDF

Uploaded by

Copyright:

Available Formats

1 Signals PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

1 Signals PDF

Uploaded by

Copyright:

Available Formats

Note: All physical signals are real.

Part 1. SIGNALS & SYSTEMS

Signals as functions of time

They capture properties of the signal in a

1. Signal: a function of certain variables

They are convenient for analysis (as we will

a function of time: price of stocks

ES156 Harvard SEAS

Signals with finite power

Signal Energy and Power

Example: Electrical circuits

Average power of the signal in this period is

Average power is defined as

The total energy of x[n] is

Signals with finite energy

We are interested in 2 classes of signals:

Define the energy of a signal x(t) in a period [t1 , t2 ] as

Finite energy signals must have zero average power.

The signal has finite total energy if E < .

ES156 Harvard SEAS

The average total power of the signal is

The signal has finite average power if P < .

time shift or translation of the signal by t0

expansion by a factor of if < 1

4. Combinations of the above

an expansion (or contraction) by ||

Note: For any signal (real or complex), the energy

ES156 Harvard SEAS

Even and Odd Signals

Signals also could be ODD or EVEN

Example: ej0 t , where 0 is a real positive

Same definition holds for discrete signals. If N0 > 0 is

Odd signal: x(t) = t

Remark: Any signal can be written as a sum of an

for all n Z, then N0 is a period of x[n]

x(t) + x(t) x(t) x(t)

If N0 is the smallest number with this property,

xeven (t): even part of x(t),

For discrete signals if there is a period, there is

xodd (t): odd part of x(t).

Example: x(t) is a constant.

ES156 Harvard SEAS

Exponential and Sinusoidal Functions

A general continuous exponential function is

x[n] = cos 0 n could be periodic, an example is

0 = 2, but it could be aperiodic if 0 = 2.

where C and are complex numbers.

|C|ej ert ej0 t

|C|ert ej(0 t+)

|C|ert cos(0 t + ) + j|C|ert sin(0 t + )

ES156 Harvard SEAS

Harmonically related complex exponentials are

Harmonics of a periodic exponential signal

Cen = |C|ern ej(0 n+)

|C|ern [cos(0 n + ) + j sin(0 n + )]

ej0 n = ej(0 +2)n , so we only deal with

r > 0 and x(t) = e cos(0 t + ): signal is an

0 close to 0 and 2 are low frequencies

r < 0 and x(t) = e cos(0 t + ): signal is an