Chapter1 mg1

TEL252E
Signals and Systems

Prof.Dr. Muhittin Gkmen
Dept. Of Computer Eng.
MG2011
TEL252E Signals and Systems
Lecture 1
1) General Informations about course
2) Signals
3) Systems
4) Examples
MG2011
General Information
Course Name:
Course Staff:
Instructors: Prof.Dr.Muhittin Gkmen
Room #: EEB 4316
e-mail:gokmen@cs.itu.edu.tr
Do.Dr. Sabih Atadan
Room #:EEB 5109
e-mail: satadan@cs.itu.edu.tr
Teaching Assistant: Muzaffer Ege Alper
Room #: Research Lab3
MG2011
General Information (Cntd)

Course Material:
Lecture Notes
Oppenheim, A. V., and A. S. Willsky, with
S. H. Nawab. Signals and Systems. 2nd ed.
New Jersey: Prentice-Hall, 1997.
ISBN: 0138147574.
MG2011
General Information
Evaluation:
4 Homeworks %10
2 Quizes
%15
Mid-Term
%35
Final
%40
Detailed information can be found in the course
web-page:
http://www.cs.itu.edu.tr/~canberk/tel252e.htm
MG2011
Calendar
Week 1 Introduction
Week 2 Continuous-Time and Discrete-Time Signals and Systems. System Properties.
Singular functions.
Week 3 Convolution. Periodic Signals.
Week 4 Continuous- and Discrete-Time Fourier Series.
Week 5 Continuous-Time Fourier Transform.
Week 6 Continuous-Time Fourier Transform (cont.). Discrete-Time Fourier Transform.
Week 7 Discrete-Time Fourier Transform (cont.).
Week 8 First and Second Order Continuous- and Discrete-Time Systems. Ideal and Non-Ideal
Filters.
Week 9 Midterm Exam
Week 10 Sampling. Impulse-Train Sampling. Sampling Theorem and Aliasing. Zero and First
Order Hold. Analog-to-Digital and Digital-to-Analog Conversions.
Week 11 Laplace Transforms, Unilateral and Bilateral z-Transforms, Region of Convergence
(ROC). The relationships between Laplace Transform, (Continuous and Discrete) Fourier
Transforms and z-Transform.
Week 12 Transfer Functions using the Laplace- and z-Transforms, Pole-Zero Plot in s- and zplanes, Stability.
Week 13 Constant Coefficient Linear Differential and Difference Equations.
Week 14 Block Diagram Representation of Continuous- and Discrete-Time Systems. Direct
Form, Series and Cascade Filter Realizations. Feedback Structure in s-Domain.
MG2011
Course Outline (Tentative)
Fundamental Concepts of Signals and Systems
Linear Time-Invariant (LTI) Systems
Fourier Transform & Properties

Modulation (An application example)
Discrete-Time Frequency Domain Methods
Response to complex exponentials

Harmonically related complex exponentials
Fourier Integral
Convolution integral and sum

Properties of LTI Systems
Fourier Series
Signals
Systems
DT Fourier Series
DT Fourier Transform
Sampling Theorem
Z Transform
Stability analysis in z domain
MG2011
Chapter I
Signals and Systems
MG2011
SIGNALS
Signals are functions of independent variables
that carry information about the behavior or
nature of some phenomenon
For example:
Electrical signals --- voltages and currents
in a circuit
Acoustic signals --- audio or speech signals
(analog or digital)
Video signals --- intensity variations in an
image (e.g. a CAT scan)
Biological signals --- sequence of bases in a
gene
MG2011
What is Signal?
Signal is the variation of a physical phenomenon / quantity with respect

to one or more independent variable
A signal is a function.
Example 1: Voltage on a capacitor as a function of time.
R
+
Vs I
Vc
-
RC circuit
MG2011
10
What is Signal?
Example 2 : Closing value of the stock exchange
index as a function of days
Example 3:Image as a function of x-y coordinates

(e.g. 256 X 256 pixel image)
Index
M T
Fig. Stock exchange
MG2011
11
THE INDEPENDENT VARIABLES

Can be continuous
Trajectory of a space shuttle
Mass density in a cross-section of a brain
Can be discrete
DNA base sequence
Digital image pixels
Can be 1-D, 2-D, N-D
For this course: Focus on a single (1-D) independent variable
which we call time.
Continuous-Time (CT) signals: x(t), t continuous values
Discrete-Time (DT) signals:
x[n], n integer values only
MG2011
12
CT Signals
Most of the signals in the physical world are CT

signalsE.g. voltage & current, pressure,
temperature, velocity, etc.
MG2011
13
DT Signals
x[n], n integer, time varies discretely
Examples of DT signals in nature:

DNA base sequence
Population of the nth generation of certain
species
MG2011
14
Many human-made DT Signals

Ex.#1 Weekly Dow-Jones
industrial average
Ex.#2 digital image
Courtesy of Jason Oppenheim.

Used with permission.
Why DT? Can be processed by modern digital computers

and digital signal processors (DSPs).
MG2011
15
Continuous-Time vs. Discrete Time
Signals are classified as continuous-time (CT) signals and

discrete-time (DT) signals based on the continuity of the
independent variable!
In CT signals, the independent variable is continuous (See

Example 1 (Time))
In DT signals, the independent variable is discrete (See
Ex 2 (Days), Example 3 (x-y coordinates, also a 2-D
signal))
MG2011
DT signal is defined only for specified time instants!

also referred as DT sequence!
16
The postfix (-time) is accepted as a convention, although some

independent variables are not time
To distinguish CT and DT signals, t is used to denote CT

independent variable in (.), and n is used to denote DT
independent variable in [.]
Discrete x[n], n is integer
Continuous x(t), t is real
Signals can be represented in mathematical form:
x(t) = et, x[n] = n/2
0
y(t) = 2
t
Discrete signals can also be represented as sequences:
MG2011
,t < 5
,t 5
{y[n]} = {,1,0,1,0,1,0,1,0,1,0,}
17

Graphically,
x[n]
x(t )
x[2] ...
x[1]
x[2]
x[1] x[0]
... x[3]
... -3 -2 -1
0 1 2 3 4 ...
(Fig.1.7 Oppenheim)
It is meaningless to say 3/2th sample of a DT signal because it is not defined.
The signal values may well also be complex numbers (e.g. Phasor of the
capacitor voltage in Example 1 when the input is sinusoidal and R is time
varying)
MG2011
18
Signal Energy and Power
In many applications, signals are directly related to physical quantities

capturing power and energy in a physical systems
t2
Total energy of a CT signal x(t ) over
t1 t t 2is
t1
t2
The time average of total energy is

average power of x(t ) over t1 t t 2
x(t ) dt
1
2
(
)
x
t
dtand referred to as
(t2 t1 ) t1
Similarly, total energy of a DT signal x[n] over
n1 n nis2
n2
x[n]
n1
Average power of x[n] over
n1 n n2is
n2
1
2
x[n]
(n2 n1 + 1) n1
MG2011
19
For infinite time intervals:
Energy: accumulation of absolute of the signal
E = lim x(t ) dt =
2
T T
=
E
=
x[n]
lim
2
T n = N
x(t ) dt
Total energy in CT signal
x[ n]
Total energy in DT signal
n =
E < are of finite energy
Signals with
In order to define the power over infinite intervals we need to take limit of
the average:
Note: Signals with E <

1
E
2
=
(
)
lim
P = lim
x
t
dt
T 2T
T 2T
have P = 0
T
N
1
E
2
[
]
=
lim
P = lim
x
n
N 2 N + 1
N 2 N + 1
n= N
MG2011
20
Energy signal iff 0<E<, and so P=0
e.g:
{x[n]} = {... 1,1, 1,1,1,1...}
Neither energy nor power, when both E and P are infinite
0, t < 0
x(t ) = t
e , t 0
Power signal iff 0<P<, and so E=
e.g:
e.g:
x(t ) = e t
Exercise: Calculate power and energy for the above signals

MG2011
21
Transformation of Independent
Variable
Sometimes we need to change the independent variable

axis for teoretical analysis or for just practical purposes
(both in CT and DT signals)
x[n] x[n n0 ]
Time shift
Time reversal x(t ) x(t )
Time scaling
MG2011
x(t ) x(t / 2)
(reverse playing of magnetic tape)

(slow playing, fast playing)
22
Examples of Transformations
x[n]
Time shift
...
...
x[n-n0]
........
...
n0
If n0 > 0 x[n-n0] is the delayed version of x[n]

(Each point in x[n] occurs later in x[n-n0])
MG2011
23
x(t)
Time shift
x(t-t0)
t
t0
t0 < 0 x(t-t0) is an advanced version of x(t)

MG2011
24
x(t)
Time reversal
t
x(-t)
Reflection about t=0

MG2011
25
Time scaling
x(t)
t
x(2t)
compressed!
t
x(t/2)
stretched!
t
MG2011
26
It is possible to transform the independent variable with a general nonlinear

function h(t) ( we can find x(h(t)) )
However, we are interested in 1st order polynomial transforms of t, i.e.,
x(t+)
x(t)
1
Given the signal x(t):
t
2
x(t+1)
1
Let us find x(t+1):

(It is a time shift to the left)
-1
x(-t+1)
1
Let us find x(-t+1):

(Time reversal of x(t+1))
MG2011
-1
t
27
For the general case, i.e., x(t+),
1. first apply the shift (),
2. and then perform time scaling (or reversal) based on .
x(t+1)
Example: Find x(3t/2+1)
t
-1
x((3/2)t+1)
1
t
-2/3
MG2011
2/3
28
Periodic Signals
A periodic signal satisfies:
x(t ) = x(t + T ) t , T > 0
Example: A CT periodic signal

x (t )
2T
2T
x(t ) = x(t + mT ) for m Z +
If x(t) is periodic with T then
Thus, x(t) is also periodic with 2T, 3T, 4T, ...
The fundamental period T0 of x(t) is the smallest value of T for which

x(t ) = x(t + T ) t , T > 0 holds
MG2011
29
Periodic Signals
A non-periodic signal is called aperiodic.

For DT we must have
Period must be
integer!
x[n + N ] = x[n] n, N > 0

Here the smallest N can be 1,
a constant signal
The smallest positive value N0 of N is the fundamental

period
MG2011
30
Even and Odd Signals
If
x(t ) = x(t ) or x[n] = x[n]
even signal (symmetric wrt y-axis)
If
x(t ) = x(t ) or x[n] = x[n]
odd signal (symmetric wrt origin)

x(t)
x(t)
odd
even
Decomposition of signals to even and odd parts:
1
{
}
EV x(t ) = [x(t ) + x(t )]
2
OD { x(=
t )}
MG2011
x(t ) = EV {x(t )}+ OD{x(t )}
1
[ x(t ) x(t )]
2
31
Exponential and Sinusoidal Signals
Occur frequently and serve as building blocks to construct many

other signals
CT Complex Exponential:
at
x(t ) = Ce
where a and C are in general complex.
Depending on the values of these parameters, the complex

exponential can exhibit several different characteristics
a>0
Real Exponential
(C and a are real)
x(t)
a<0
x(t)
C
t
MG2011
32
Periodic Complex Exponential (C real, a purely imaginary)
x(t ) = e
Is this function periodic?
x(t ) = e
jw0t
=e
jw0 ( t +T )
for periodicity
MG2011
jw0t
=e
jw0t
.e
jw0T
T=
2n
n Z +
=1
The fundamental period is T0 =
Thus, the signals ej0t and e-j0t have the same fundamental
period
33
By using the Eulers relations:
e j = cos + j sin
e j = cos j sin
= 0 t +
We can express: (put
e j0t = cos 0t + j sin 0t
A j (0t + ) j (0t + )
A j j 0 t
+e
= e e + e j e j 0 t
A cos(0t + ) = e
2
2
A j (0t + ) j (0t + )
A j j 0 t
e
=
A sin(0t + ) =
e
e e e j e j 0 t
2j
2j
Sinusoidals in terms of complex exponentials

MG2011
34

Alternatively,
Acos(0t + ) = ARe(e j ( t + ) )
0
Asin(0t + ) = AIm(e j ( t + ) )
0
x (t )
T0 =
0
CT sinusoidal signal
A cos
A cos(0t + )
x(t ) = A cos(0t + )
MG2011
35
Complex periodic exponential and sinusoidal signals are of infinite

total energy but finite average power
E period =
T +T0
Pperiod
j 0 t 2
T +T0
dt = 1 dt =(T + T0 ) T = T0
T
1
=
E period = 1
(T + T0 ) T
As the upper limit of integrand is increased as T + 2T0 , T + 3T0 ,...E period

However, always Pperiod = 1
Thus,
Finite average power!
T
1
j t 2
P = lim
e
dt = 1 IMPORTANT
T 2T
T
0
MG2011
36
Harmonically Related Complex

Exponentials
Set of periodic exponentials with fundamental

frequencies that are multiplies of a single
positive frequency 0
xk (t ) = e jk0t for k = 0,1,2,...

k = 0 xk (t ) is a constant
k 0 xk (t ) is periodic with fundamental frequency k 0
T0
2
and fundamental period
= , where
k 0 k
MG2011
T0 =
0
37

Exponentials
kth harmonic xk(t) is still periodic with T0 as well

Harmonic (from music): tones resulting from
variations in acoustic pressures that are integer
multiples of a fundamental frequency
Used to build very rich class of periodic signals
MG2011
38
General Complex Exponential Signals

Here, C and a are general complex numbers
Say, C = C e j and a = r + j0 x(t ) = Ce at
Then
x(t ) = Ce at = C e rt e j (0t + ) = C e rt cos(0t + ) + j C e rt sin(0t + )
(Real and imaginary parts) Growing and damping sinusoids for r>0 and r<0
x(t )
t, r > 0
MG2011
envelope
t, r < 0
x(t ) = Ce rt cos(0t + )
39
DT Complex Exponential and Sinusoidal Signals

x[n] = C n where C and are in general complex numbers
It is more convenient and customary to use instead of en
Real exponential signals : C and are real
for > 1,0 < < 1,1 < < 0, < 1
(Sign alternation for < 0)
>1
MG2011
0 < <1
1 < < 0
< 1
40
DT Sinusoidal Signals
Consider e j0 n we then have similar to the CT case
A
A cos(0 n + ) = (e j e j0 n + e j e j0 n )
2
Infinite energy, finite average power with 1.
>1
General complex exp signals
If C and are in polar form as

C = C e j , = e j 0
then
C n = C cos(0 n + ) + j C sin(0 n + )
n
<1
Real and imaginary parts of DT general complex exp are

sinusoidals (growing > 1, and decaying < 1)
MG2011
41
Periodicity Properties of DT Signals

Consider the DT complex exp : e j0 n
j2n
Let' s find e j(0 + 2 )n = e j0 n e
=1
SO THE FN WITH FREQ 0 IS THE SAME AS THE FN WITH 0 + 2

This is very different from CT complex exp.
CT exp has distinct freq values 0
DT exp has identical freq values 0 + 2k , k Z
Result:
It is sufficient to consider an interval from 0 to 0 +2 to completely
characterize the DT complex exponential!
MG2011
42
One usually takes 0 0 < 2 or - 0 <

For CT exp as 0 the rate of oscillation indefinitely
For DT exp as 0 from 0 to , the rate of oscillation ,
as 0 more until 2 , the rate of oscillation to zero.
Hence, low freq (slow varying) DT complex exp is around 0 = 0 and 0 = 2
High freq (rapidly varying) DT complex exp is around 0 =
What about the periodicity of DT complex exp?
MG2011
43

j 0 N
Periodicity condition : e j0 ( n + N ) = e j0 n e
(*)
must be unity
This holds if 0 N is an integer multiple of 2 . (**)

In other words some integer m we must have 0 N = 2m
0 m
Or equivalently
= (* * *)
2 N
We have the conditions from (*) and (**) that m and N
must be integers.
0 m
So DT exp is periodic when
= is a rational number,
2 N
not periodic otherwise!!!
MG2011
44

Take the common factor out
2 0
=
The fundamental frequency is then
N
m
2
(* * **)
The fundamental period is then N = m
0
Therefore to find the fund freq of an complex exp we need to express
0
as in (* * *)
2
(The same development is also valid for DT sinusoidal signals.)!!
MG2011
45

Examples
Ex : x[n] = cos(2n
) is periodic with fund period 12.
12
0 1
2
) = cos(0 n) 0 =
no factors in common,
=
12
12
2 12
12
so by using (* * **) , N = 1 = 12
1
x[n] = cos(2n
Ex : x[n] = cos(4n
x[n] = cos(4n
) is periodic with fundamental period 6.
12
0 2 (n = 1)
4
) = cos(0 n) 0 =
,
= =
12
12
2 12 ( N = 6)
12
then using (* * **) , N = 1 = 6
2
MG2011
46

Examples
OBSERVATION:
With no common factors between N and m, N in (***) is the

fundamental period of the signal
Hence, if we take common factors out
0 1
=
N =6
2 6
Comparison of Periodicity of CT and DT Signals:
Consider x(t) and x[n]
x(t) = cos ( 2t
12
) x[n] = cos ( 2n
12
x(t) is periodic with T=12, x[n] is periodic with N=12.

MG2011
47

Examples
x(t ) = cos 8t
)
31
and
x[n] = cos 8n
But, if
x(t) is periodic with 31/4.

In DT there can be no fractional periods, for x[n] we have
31
0
4
=
2
31
then N=31.
If
( 6 ) and x[n] = cos(n 6 )
x(t ) = cos t
x(t) is periodic with 12, but x[n] is not periodic, because

no way to express it as in (***)
0
1 there is
=
2 12
Study Fig.1.27 page 27, Table 1.1 in Opp. Example 1.6 as well
MG2011
48

Exponentials (Discrete Time)
Set of periodic exponentials with a common period

N
2
Signals at frequencies multiples of
N
(from 0N=2m)
k [ n ] = e
MG2011
2
jk
N
for k = 0,1,2,...
In CT, all of the HRCE, e jk0t for k = 0,1,2,... are

distinct
Different in DT case!
49

Exponentials (Discrete Time)
Lets look at (k+N)th harmonic:
k + N [ n] = e
2
j ( k + N )
n
N
=e
2
jk
n
N
j 2n
.e = k [n]
=1
Only N distinct periodic exponentials in k[n] !!

That is,
0 [n] = 1, 1[n] = e
2
n
N
, 2 [ n] = e
4
n
N
, , N 1[n] = e
j ( N 1)
2
n
N
0 [n] = N [n], 1[n] = N 1[n]

MG2011
50
Unit Impulse and Unit Step Functions
Basic signals used to construct and represent other signals
DT unit impulse:
0, n 0
[ n] =
1, n = 0
[n]
Unit impulse (unit sample)
- - -
- - n
DT unit step:
0, n < 0
u[ n] =
1, n 0
u[n]
- - -
- - -
Relation between DT unit impulse and unit step (?):
[n] = u[n] u[n 1]

(DT unit impulse is the first difference of the DT step)
MG2011
51
u[n] = [n - k]
0
Interval of summation
[n-k]
- - n
(DT step is the

running sum of DT
unit sample)
n<0
- - k
0
Interval of summation
[n-k]
- - -
- - -
x[n] [n] = x[0] [n]
n>0
k
More generally for a unit impulse [n-n0] at n0 :
x[n] [n n0 ] = x[n0 ] [n n0 ]
MG2011
Sampling property
52

(Continuous-Time)
CT unit step:
0, t < 0
u (t ) =
1, t > 0
u(t)
t
(t)
CT impulse:
du (t )
(t ) =
dt
t
u (t ) = ( )d
MG2011
t
0
CT unit impulse is the

1st derivative of the
unit sample
CT unit step is the running integral

of the unit impulse
53
Continuous-Time Impulse
CT impulse is the 1st derivative of unit step

du (t )
(t ) =
dt
There is discontinuity at t=0, therefore we define
u (t ) = lim u (t )
u (t )
1
0
du (t )
(t ) =
dt
(t ) = lim (t )
0
(t )
(t )
1/
0
MG2011
u (tas
)
t
0
54
Continuous-Time Impulse
REMARKS:
Signal of a unit area
Derivative of unit step function
Sampling property x(t ) (t t0 ) = x(t0 ) (t t0 )

The integral of product of (t) and (t) equals (0) for
any (t) continuous at the origin and if the interval of
integration includes the origin, i.e.,
t2
( ) ( )d
= (0)
for t1 < 0 < t 2
t1
MG2011
55
CT and DT Systems
What is a system?
A system: any process that results in the transformation

of signals
A system has an input-output relationship
Discrete-Time System: x[n] y[n] : y[n] = H[x[n]]
x[n]
DT
System
y[n]
Continuous-Time System: x(t) y(t) : y(t) = H(x(t))

x(t)
CT
y(t)
System
MG2011
56
CT and DT Systems
Examples
In CT, differential equations are examples of systems

Zero state response of the capacitor voltage in a series
RC circuit
R
Vs I
+
V
c
dvc (t ) 1
1
+
vc (t ) =
vs (t )
dt
RC
RC
vc(t): output, vs(t): input
RC circuit
In DT, we have difference equations

Consider a bank account with %1 monthly interest rate
y[n] = 1.01 y[n 1] + x[n]
added on:
MG2011
y[n]: output: account balance at the end of each month

x[n]: input: net deposit (deposits-withdrawals)
57
Interconnection of Systems
Series (or cascade) Connection: y(t) = H2( H1( x(t) ) )

x(t)
y(t)
System 2
H2
e.g. radio receiver followed by an amplifier
System 1
H1
Parallel Connection:
x(t)
y(t) = H2( x(t) ) + H1( x(t) )
System 1
H1
y(t)
System 1
H2
MG2011
e.g. phone line connecting parallel phone microphones

58
Interconnection of Systems
Previous interconnections were feedforward systems
The systems has no idea what the output is
Feedback Connection: y(t) = H2( y(t) ) + H1( x(t) )

x(t)
System 1
H1
y(t)
System 2
H2
In feedback connection, the system has the knowledge of output

e.g. cruise control
Possible
MG2011
to have combinations
of connections..
59
System Properties
Memory vs. Memoryless Systems
Memoryless Systems: System output y(t) depends only

on the input at time t, i.e. y(t) is a function of x(t).
e.g. y(t)=2x(t)
Memory Systems: System output y(t) depends on input

at past or future of the current time t, i.e. y(t) is a function
of x() where - < <.
Examples:
A resistor:
y(t) = R x(t)
A capacitor:
1
y (t ) = x( )d
C
A one unit delayer: y[n] = x[n-1]
MG2011
An accumulator:
y[n] =
x[k ]
k =
60
System Properties
Invertibility
A system is invertible if distinct inputs result in distinct outputs.

If a system is invertible, then there exists an inverse system which
converts output of the original system to the original input.
Examples:
x(t)
System
y (t ) = 4 x(t )
y(t)
Inverse
System
w(t)=x(t)
x[k ]
y (t ) =
w[n] = y[n] y[n 1]
w(t ) =
y[n] =
k =
w(t ) =
1
y (t )
4
y (t ) = x 4 (t )
MG2011
x(t )dt
dy (t )
dt
Not invertible
61
System Properties
Causality
A system is causal if the output at any time depends only on values of

the input at the present time and in the past
Examples:
Capacitor voltage in series RC circuit (casual)
y (t ) = 2 x(t + 4)
Non-causal
y[n] = x[n]
Non-causal (why?)
y (t ) = 2 x(t 4) cos(t + 1)
Causal (why?)
Systems of practical importance are usually casual

However, with pre-recorded data available we do not constrain
ourselves to causal systems (or if independent variable is not time, any
example??)
M
1
y[n] =
MG2011
(For n<0, system requires future

inputs)
x[n k ]
2M + 1
k = M
Averaging system in a block of data
62
System Properties
Stability
A system is stable if small inputs lead to responses that do not diverge

More formally, a system is stable if it results in a bounded output for any
bounded input, i.e. bounded-input/bounded-output (BIBO).
If |x(t)| < k1, then |y(t)| < k2.
Example:
t
y (t ) = x(t )dt
y[n] = 100 x[n]
stable
Ball at the base of valley
Ball at the top of hill

x(t)
y(t)
y(t)
x(t)
MG2011
Averaging system:
Interest system:
M
1
y[n] =
x[n k ]
2M + 1 k = M
y[n] = 1.01 y[n 1] + x[n]

stable
unstable (say x[n]=[n], y[n]
grows without bound
63
System Properties
Time-Invariance
A system is time-invariant if the behavior and characteristics of the

system are fixed over time
More formally: A system is time-invariant if a delay (or a time-shift)
in the input signal causes the same amount of delay (or time-shift)
in the output signal, i.e.:
x(t) = x1(t-t0) y(t) = y1(t-t0)
x[n] = x1[n-n0] y[n] = y1[n-n0]
Examples:
y[n] = nx[n]
y (t ) = x(2t )
y (t ) = sin x(t )
x1[n] y1[n] = nx1[n]
x2 [n] = x1[n n0 ] y2 [n] = nx1[n n0 ]

y1[n n0 ] = (n n0 ) x1[n n0 ] y2 [n]
(Ex. 1.15)
Not TIV
MG2011
Not TIV
(explicit operation
on time)
(Ex. 1.16)
TIV
(Ex. 1.14)
When showing a system is not TIV, try to

find counter examples
64
System Properties
Linearity
A system is linear if it possesses superposition property,

i.e., weighted sum of inputs lead to weighted sum of
responses of the system to those inputs
In other words, a system is linear if it satisfies the
properties:
It is additivity:
x(t) = x1(t) + x2(t) y(t) = y1(t) + y2(t)
And it is homogeneity (or scaling): x(t) = a x1(t) y(t) = a y1(t), for
a any complex constant.
The two properties can be combined into a single

property:
Superposition:
x(t) = a x1(t) + b x2(t) y(t) = a y1(t) + b y2(t)
x[n] = a x1[n] + b x2[n] y[n] = a y1[n] + b y2[n]
MG2011
How
do you check linearity

of a given system?
65
System Properties
Linearity
Examples:
y (t ) = x (t )
2
y[n] = 2 x[n] + 3
y[n] = Re{x[n]}
x1[n] = 2, x2 [n] = 3
x1[n] = r[n] + js[n] y1[n] = r[n]

nonlinear
x2 [n] = ax1[n] = j (r[n] + js[n])
x1[n] y1[n] = 2.2 + 3 = 7

x2 [n] y2 [n] = 2.3 + 3 = 9
for a = j
x2 [n] = s[n] + jr[n]

x2 [n] y2 [n] = s[n] ay1[n]
x1[n] + x2 [n] = 2 + 3 = 5
x1[n] + x2 [n] y1+ 2 [n] = 2.5 + 3 = 13
y1[n] + y2 [n] = 7 + 9 = 16 y1+ 2 [n]
nonlinear
nonlinear
y[n] = 2 x[n 1]
linear
MG2011
66
Superposition in LTI Systems
For an LTI system:

given response y(t) of the system to an input signal x(t)
it is possible to figure out response of the system to any signal
x1(t) that can be obtained by scaling or time-shifting the input
signal x(t), i.e.:
x1(t) = a0 x(t-t0) + a1 x(t-t1) + a2 x(t-t2) +
y1(t) = a0 y(t-t0) + a1 y(t-t1) + a2 y(t-t2) +
Very useful property since it becomes possible to solve a

wider range of problems.
This property will be basis for many other techniques
that we will cover throughout the rest of the course.
MG2011
67
Superposition in LTI Systems
Exercise: Given response y(t) of an LTI system to the input signal x(t) below,
find response of that system to the input signals x1(t) and x2(t) shown below.
y(t)
x(t)
2
1
t
t
-1
1
x1(t)
2
1
x2(t) 4
3
2
t
-1
-1/2 1/2 1
MG2011
68
Audio
Physics Of Acoustics
Sound can be thought of in one of two ways:
In the time domain
In the frequency domain
Transformation is accomplished by a Fourier transform
Fourier:
Any waveform can be represented by a sum of sine waves
MG2011
69
Audio
cont
Loudness measured in dB (decibel): logarithm
+10 dB perceived as double loudness
loudness =10log I I
I0 threshold (just perceived) (10-12 Joule/s . m), OdB
+10 dB (double loudness) equals 10-fold energy
0
MG2011
70
Audio
cont
Frequency Pitch
Amplitude Loudness
MG2011
71
Audio
Analog to Digital: Theory of Sampling

From analog to digital:
0. Filter (low-pass): Remove frequencies larger than fmax
1. Sample: Measure signal value (1a: hold=store temporarily)
2. Quantize: Relate value to interval (A/D conv. quantizes, too)
3. Encode: Assign binary code (e.g., integer no. of interval)
Steps 1,2,3 standardized as "Pulse Code Modulation" PCM

Important factors:
Sampling rate: Number of sampled values per second
Quantization depth: Number of bits per digitized value
MG2011
72
Audio
Analog to Digital: Theory of Sampling
cont
Sampling rate determined by properties of recorded sound:

Nyquist: For lossless digitization, the sampling rate (frequency fs)
should be at least twice the maximum frequency considered, fmax
Mathematically precise: "any epsilon larger than...", i.e. fs > 2fmax
Attention: assumes 0 quantization error, unrealistic quantization
noise
Music typically extends from 20 Hz to 20 kHz
Speech 100 Hz to 10 kHz, major energy in band from 200Hz to 4kHz
Quantization depth determined by desired sound
quality (quant. noise): Typically 8 (256 levels) or
16 (65,536 levels)
Samples always "per channel": (e.g., 2x for stereo)
Logarithmic Quantization:
compensates for fact that quantization
error much more audible around 0 amplitude
MG2011
73
3
Audio
Audio Quality of Common Appliances
MG2011
74
3
Audio
Sampling Rates
MG2011
75
Audio
Common Coding Methods

Coding = representation of sampled values by integers / bits
PCM (Pulse Code Modulation)
integer value = (quantized) sampled value
simple but requires high number of bits
DPCM (Differential PCM)
integer value = difference between current value and predicted value
prediction based on previous values
requires less bits than PCM for same quality
DM (Delta Modulation)
as DPCM but only differences of 1 and -1 allowed
requires minimal number of bits but quality can be poor
ADPCM (Adaptive Differential PCM):
as DPCM but adapts predictor to signal characteristics
simplest prediction: "difference remains"
also adapts width of quantization steps to signal characteristics
better quality than DPCM with same storage requirements
MG2011
76
Audio
cont
Frequency Pitch
Amplitude Loudness
MG2011
77
3
Audio
Example
Phonem: Stairs
Diphon
Half syllable
Syllable
MG2011
78
3
Audio
Speech Synthesis in Frequency Domain
MG2011
79
3
Audio
Speech Input
MG2011
80
3
Audio
Speech Recognition
Sound pattern
Word model
MG2011
81
3
Audio
Sound Enhancement
Declicking
Noise Reduction
Echo suppression
MG2011
82
Audio
Declicking
Removing impulse distortions, also called declicking of
recorded sound, is performed in two steps. In a first step
impulse distortions - clicks - are detected within a signal,
which are going to be removed from the signal in the second
step.
MG2011
83
cont
Audio
Declicking
MG2011
84
3
Audio
Noise Reduction
MG2011
85
Audio
Echo Suppression
sonogram
MG2011
86
Baseline JPEG Encoder Block Diagram
MG2011
87
Baseline JPEG Decoder Block Diagram
MG2011
88
Baseline JPEG Pros and Cons

Disadvantages
Advantages
Single resolution
Memory Efficient
Single quality
Low complexity
Compression efficiency No target bit rate
Visual model utilization No lossless capability
No tiling
Robustness
No ROI
Blocking artifacts
Poor error resilience
MG2011
89
MG2011
90
MG2011
91
Noise reduction
Edge Enhancement
Zooming
MG2011
92
MG2011
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
93
Median Operation
61
621 59
1
651
57
601
63
1
59
551
49
53
62
59
65
561
60
63
56
58
1
571
55
58
57
55
45
C1,2=
MG2011
rank
59
9
8
7
6
5
4
3
2
1
65
63
62
60
59
58
57
56
55
94
9x9 Median
MG2011
95
Edge Detection
What is an edge
MG2011
A large change in image brightness of a short

spatial distance
Edge strength = (I(x,y)-I(x+dx,y))/dx
96
Roberts Operator
[I ( x, y ) I ( x + 1, y + 1)]2 + [I ( x, y + 1) I ( x + 1, y )]2
or
I ( x, y ) I ( x + 1, y + 1) + I ( x, y + 1) I ( x + 1, y )
1 0
0 -1
MG2011
0 1
+ -1 0
Does not return any information about the

orientation of the edge
97
Prewitt Operator
-1 -1 -1
P1= 0 0 0
1 1 1
P2
=
Edge Magnitude =
-1 0
-1 0
-1 0
1
1
1
P12 + P22
P1
Edge Direction = tan
P2
1
MG2011
98
Robinson Compass Masks
MG2011
-1 0 1
-2 0 2
-1 0 1
0 1 2
-1 0 1
-2 -1 0
1 2 1
0 0 0
-1 -2 -1
2 1 0
1 0 -1
0 -1 -2
1
2
1
0 -1 -2
-1 0 -1
2 1 0
-1 -2 -1
0 0 0
1 2 1
-2 -1 0
-1 0 1
0 1 2
0 -1
0 -2
1 -1
99
0 1 2
-1 0 1
-2 -1 0
1 2 1
0 0 0
-1 -2 -1
MG2011
100
2D Laplacian Operator
f ( x, y ) f ( x, y )
f ( x, y ) =
+
2
2
x
y
2
Convolution masks approximating a Laplacian

0 -1 0
-1 4 -1
0 -1 0
MG2011
1 -2 1
-2 4 -2
1 -2 1
-1 -1 -1
-1 8 -1
-1 -1 -1
101
0 -1 0
-1 4 -1
0 -1 0
Input
MG2011
Mask
Output
102
Chapter 4
Image Enhancement in the
Frequency Domain
MG2011
103
Chapter 4
Frequency Domain
MG2011
104
Chapter 3
Spatial Domain
MG2011
105
Chapter 3
Spatial Domain
MG2011
106

Chapter1 mg1

Hochgeladen von

Dokumentinformationen

Originaltitel

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

Chapter1 mg1

Hochgeladen von

Copyright:

Verfügbare Formate

TEL252E

Signals and Systems

TEL252E Signals and Systems

TEL252E Signals and Systems

TEL252E Signals and Systems

General Information (Cntd)

TEL252E Signals and Systems

TEL252E Signals and Systems

TEL252E Signals and Systems

Course Outline (Tentative)

Fundamental Concepts of Signals and Systems

Linear Time-Invariant (LTI) Systems

Fourier Transform & Properties

Discrete-Time Frequency Domain Methods

Response to complex exponentials

Convolution integral and sum

Stability analysis in z domain

TEL252E Signals and Systems

TEL252E Signals and Systems

TEL252E Signals and Systems

Signal is the variation of a physical phenomenon / quantity with respect

TEL252E Signals and Systems

Example 3:Image as a function of x-y coordinates

Fig. Stock exchange

TEL252E Signals and Systems

THE INDEPENDENT VARIABLES

TEL252E Signals and Systems

Most of the signals in the physical world are CT

TEL252E Signals and Systems

Examples of DT signals in nature:

TEL252E Signals and Systems

Many human-made DT Signals

Ex.#2 digital image

Courtesy of Jason Oppenheim.

Why DT? Can be processed by modern digital computers

TEL252E Signals and Systems

Continuous-Time vs. Discrete Time

Signals are classified as continuous-time (CT) signals and

In CT signals, the independent variable is continuous (See

DT signal is defined only for specified time instants!

TEL252E Signals and Systems

Continuous-Time vs. Discrete Time

The postfix (-time) is accepted as a convention, although some

To distinguish CT and DT signals, t is used to denote CT

Signals can be represented in mathematical form:

x(t) = et, x[n] = n/2

Discrete signals can also be represented as sequences:

TEL252E Signals and Systems

Continuous-Time vs. Discrete Time

It is meaningless to say 3/2th sample of a DT signal because it is not defined.

TEL252E Signals and Systems

Signal Energy and Power

In many applications, signals are directly related to physical quantities

Total energy of a CT signal x(t ) over

The time average of total energy is

Similarly, total energy of a DT signal x[n] over

Average power of x[n] over

TEL252E Signals and Systems

Signal Energy and Power

For infinite time intervals:

Energy: accumulation of absolute of the signal

Total energy in CT signal

Total energy in DT signal