9/9/99 (T.F.

Weiss)

Lecture #1: Introduction to signals

Signals and systems This subject deals with mathematical methods used to describe signals and to analyze and synthesize systems.

Motivation: To describe signals, both man-made and naturally occurring.

Signals are variables that carry information Systems process input signals to produce output signals.

Outline: Classication of signals Building-block signals complex exponentials, impulses Conclusions

Today SIGNALS; Next time SYSTEMS.

Demonstration of dierent types of signals

EKG Microphone CD Oscillator Pulse generator AM/FM generator Switch box

Classication of signals Identity of the independent variable Time is often the independent variable for signals. For example, the electrical activity of the heart recorded with electrodes on the surface of the chest the electrocardiogram (ECG or EKG).
ECG amplitude (mV) 1 0.5 0 -0.5

Oscilloscope Speaker Audio Amplifier

8 10 12 Time (secs)

14

16

18

20

ECG amplitude (mV)

Generic time The term time is often used generically to represent the independent variable of a signal. The independent variable may be a spatial variable as in an image. Here color information is specied as a function of position.

Dimensionality of the independent variable The independent variable can be 1-D (time t in the EKG signal) or 2-D (space x, y in the image), 3-D, or N -D.

1 0.5 0 -0.5

8 10 12 Time (secs)

14

16

18

20

In 6.003 we shall consider largely 1-D signals, but signals in many applications (e.g., radio astronomy, medical imaging, seismometry) have multiple dimensions.
5 6

Continuous time (CT) and discrete time (DT) signals CT signals take on real or complex values as a function of an independent variable that ranges over the real numbers and are denoted as x(t). DT signals take on real or complex values as a function of an independent variable that ranges over the integers and are denoted as x[n]. Note the subtle use of parentheses and square brackets to distinguish between CT and DT signals.

For example, consider the image shown on the left and its DT representation shown on the right
1 2 3 1 2 3 n N

x(t)

x[n]
M

-4

-2

-4

-2

n
7

The image on the left consists of 302 435 picture elements (pixels) each of which is represented by a triplet of numbers {R,G,B} that encode the color. Thus, the signal is represented by c[n, m] where m and n are the independent variables that specify pixel location and c is a color vector specied by a triplet of hues {R,G,B} (red, green, and blue).
8

Real and complex signals Signals can be real, imaginary, or complex. An important class of signals are the complex exponentials:

Real and complex signals, contd For both exponential CT (x(t) = est) and DT (x[n] = z n) signals, x is a complex quantity. To plot x, we can choose to plot either its magnitude and angle or its real and imaginary parts whichever is more convenient for the analysis.
1 0.5

the CT signal x(t) = est where is s is a complex number, the DT signal x[n] = z n where z is a complex number. For example, suppose s = j/8 and z = ej/8, then the real parts are: {x(t)} = Q. Why do we deal with complex signals? A. They are often analytically simpler to deal with than real signals. {x[n]} = {ejt/8} = cos(t/8), {ejn/8} = cos[n/8].

x(t)

0 -0.5 -1 -20 -15 -10 1 0.5 0 -0.5 -1 -20 -15 -10 -5 -5

10

15

20

x[n]

10

15

20

10

Periodic and aperiodic signals Periodic signals have the property that x(t + T ) = x(t) for all t. The smallest value of T that satises the denition is called the period. Shown below are an aperiodic signal (left) and a periodic signal (right).

Causal and anti-causal signals A causal signal is zero for t < 0 and an anti-causal signal is zero for t > 0

x(t)
0 t 0

x(t)
t

x(t)

x(t)

Right- and left-sided signals A right-sided signal is zero for t < T and a left-sided signal is zero for t > T where T can be positive or negative.

x(t)
T t

x(t)
T t
12

11

Bounded and unbounded signals Even and odd signals Even signals xe(t) and odd signals xo(t) are dened as

Causal

Anti-causal

x(t)
Unbounded Unbounded

x(t)

xe(t) = xe(t) and xo(t) = xo(t).

xe(t) t

xo(t) t

Bounded t

Bounded t

13

14

Any signal is a sum of unique odd and even signals. Using x(t) = xe(t) + xo(t) and x(t) = xe(t) xo(t), yields xe(t) = 1 1 (x(t) + x(t)) and xo(t) = (x(t) x(t)). 2 2
1

Building-block signals We will represent signals as sums of building-block signals. Important families of building-block signals are the eternal, complex exponentials and the unit impulse functions. Eternal, complex exponentials These signals have the form x(t) = Xest for all t and x[n] = Xz n for all n, where X, s, and z are complex numbers. We illustrate the richness of this class of functions for CT signals; DT signals are similarly rich. In general s is complex and can be written as s = + j,

x(t)
1/2

xe(t) t

x(t)

t
1/2

xo(t) t
-1/2
15

where and are the real and imaginary parts of s.

16

Eternal, complex exponentials real s If s = is real and X is real then x(t) = Xet, and we get the family of real exponential functions. Eternal, complex exponentials imaginary s If s = j is imaginary and X is real then x(t) = Xejt = X(cos t + j sin t), and we get the family of sinusoidal functions. Eternal, complex exponentials complex s If s = + j is complex and X is real then x(t) = Xe(+j)t = Xet(cos t + j sin t), and we get the family of damped sinusoidal functions.
17

For x(t) = Xest, {x(t)} = Xet cos t is plotted for dierent values of s superimposed on the complex s-plane.

Complex s-plane

18

For x(t) = Xest, {x(t)} = Xet sin t is plotted for dierent values of s superimposed on the complex s-plane.

Eternal complex exponentials why are they important?

Complex s-plane

Almost any signal of practical interest can be represented as a superposition (sum) of eternal complex exponentials.

The output of a linear, time-invariant (LTI) system (to be dened next time) is simple to compute if the input is a sum of eternal complex exponentials.

Eternal complex exponentials are the eigenfunctions or characteristic (unforced, homogeneous) responses of LTI systems.
19 20

Building-block signals Unit impulse denition The unit impulse (t), aka the Dirac delta function, is not a function in the ordinary sense. It is dened by the integral relation

f (t)(t) dt = f (0),

Unit impulse narrow pulse approximation To obtain an intuitive feeling for the unit impulse, it is often helpful to imagine a set of rectangular pulses where each pulse has width and height 1/ so that its area is 1.

and is called a generalized function. The unit impulse is not dened in terms of its values, but is dened by how it acts inside an integral when multiplied by a smooth function f (t). To see that the area of the unit impulse is 1, choose f (t) = 1 in the denition. We represent the unit impulse schematically as shown below; the number next to the impulse is its area.
Unit impulse

p (t)

(t) 1

t
The unit impulse is the quintessential tall and narrow pulse!

t
21 22

Unit impulse intuiting the denition To obtain some intuition about the meaning of the integral definition of the impulse, we will use a tall rectangular pulse of unit area as an approximation to the unit impulse.

1 p (t) f (t) f (0) f (t)p (t) f (0) t

Unit impulse the shape does not matter There is nothing special about the rectangular pulse approximation to the unit impulse. A triangular pulse approximation is just as good. As far as our denition is concerned both the rectangular and triangular pulse are equally good approximations. Both act as impulses.

1 t
Area = 1

f (t)p (t) dt

f (0)

= f (0).
23 24

Unit impulse the values do not matter The values of the approximating functions do not matter either. The function on the left has unit area and takes on the arbitrary value A for t = 0. The function on the right, which we shall encounter frequently in later lectures, has the property that it has non-zero values at most of its values, all but a countably innite number of points, but still acts as a unit impulse.

Two-minute miniquiz problem Problem 1-1 Interpret and sketch the generalized function x(t) where x(t) = et/4(t + 4).

A
Area = 1

sin(t/ ) t

A 2 t 2 t

What all these approximations have in common is that as gets small the area of each function occupies an increasingly narrow time interval centered on t = 0.
25 25-1

Two-minute miniquiz solution Problem 1-1 To determine the meaning of x(t) we place it in an integral

et/4(t + 4) dt.
x(t)

Problem 1-1 The result can also be seen graphically. The left panel shows both et/4 and (t + 4), and the right panel shows their product.

Let = t + 4 so that

x(t)

et/4(t + 4) dt =

e( 4)/4( ) d.

et/4 (t + 4)
1

e1(t + 4) e1
0

From the denition of the unit impulse,

f (t)(t) dt = f (0),
-4

-4

x(t) = e1(t + 4).

Impulses in space are also useful.

Unit impulse what do we need it for? The unit impulse is a valuable idealization and is used widely in science and engineering. Impulses in time are useful idealizations.

Impulse of mass density in space represents a point mass. Impulse of charge density in space represents a point charge. Impulse of light intensity in space represents a point of light.

Impulse of current in time delivers a unit charge instantaneously to a network.

Impulse of force in time delivers an instantaneous momentum to a mechanical system.

We can imagine impulses in space and time.

Impulse of light intensity in space and time represents a brief ash of light at a point in space.
26 27

Unit step Integration of the unit impulse yields the unit step function
t

Unit impulse as the derivative of the unit step As an example of the method for dealing with generalized functions consider the generalized function d u(t). dt Since u(t) is discontinuous, its derivative does not exist as an ordinary function, but it does as a generalized function. To see what x(t) means, put it in an integral with a smooth testing function d f (t) u(t) dt, y(t) = dt and apply the usual integration-by-parts theorem x(t) = y(t) = f (t)u(t)|

u(t) = which is dened as u(t) =

( ) d,

0 if t < 0 1 if t 0.

Unit impulse
(t)
1 0

Unit step
u(t)
1

u(t)

d f (t) dt, dt

to obtain y(t) = f ()
28

d 0

dt

f (t) dt = f (0).
29

Successive integration of the unit impulse Successive integration of the unit impulse yields a family of functions. Unit impulse as the derivative of the unit step, contd The result is that d f (t) u(t) dt = f (0), dt which, from the denition of the unit impulse, implies that d u(t). dt That is, the unit impulse is the derivative of the unit step in a generalized function sense. (t) =

Integration on t
Unit impulse Unit step Unit ramp Unit parabola

(t)

u(t)

tu(t)

t2 u(t) 2!

tn1 u(t) (n 1)!

Later we will talk about the successive derivatives of (t), but these are too horrible to contemplate in the rst lecture.
30 31

Building-block signals can be combined to make a rich population of signals Eternal complex exponentials and unit steps can be combined to produce causal and anti-causal decaying exponentials.

Unit steps and unit ramps can be combined to produce pulse signals.

u(t) u(t 1)

tu(t) 2(t 1)u(t 1) + (t 2)u(t 2)

x(t) = et cos(t)u(t)

x(t) = e+t cos(t)u(t)

1 1
t

1 1 2
t

32

33

Conclusions

We are awash in a sea of signals. Signal categories identity of independent variable, dimensionality, CT & DT, real & complex, periodic & aperiodic, causal & anti-causal, bounded & unbounded, even & odd, etc.

Building block signals eternal complex exponentials and impulse functions are a rich class of signals that can be superimposed to represent virtually any signal of physical interest.
34