Lecture 2
Lecture 2
Instructor: Dr. Phan Van Ca Lecture #2: Review of Signals and the Fourier Transform
Overview
Information in communication systems is transferred through the use of EM waves At each point in the system, we observe signals. These signals can be described mathematically using both the time and the frequency domains. While the time domain is more familiar to most students, often the frequency domain is more intuitive for understanding certain signal characteristics At the receiver we observe both the desired waveform as well as undesired waveforms such as noise and interference. Reading
Sections 2.1-2.3
Review important properties of signals Review a key mathematical tool for analyzing communication systems the Fourier Transform
Motivation
Course Objectives
Design digital communication systems, given constraints on data rate, bandwidth, power, fidelity, and complexity; Analyze the performance of a digital communication link when additive noise is present in terms of the signal-to-noise ratio and bit error rate; Compute the power and bandwidth requirements of modern communication systems, including those employing ASK, PSK, FSK, and QAM modulation formats; Design a scalar quantizer for a given source with a required fidelity and determine the resulting data rate; Determine the auto-correlation function of a line code and determine its power spectral density; Determine the power spectral density of bandpass digital modulation formats.
Have finite time duration (finite energy!) Occupy finite frequency spectrum Are continuous Have finite peak value Are real-valued All real-world signals will have these properties, although sometimes we use mathematical models which violate these conditions.
Mathematical Representations
discontinuity
mathematical
Ends after T seconds Has finite frequency content continuous Real wave form
Mathematical Representations
Thus, when we analyze communication systems we use mathematical models. These models allow for convenient analysis but are not completely accurate concerning the real world. Fortunately, they provide close enough approximation that the conclusions reached using the models are still valid.
Energy: E = w2 (t )dt
0< E <
1 T /2 2 = lim w (t )dt T T T / 2
For periodic signals, power can be computed by integrating over one period A signal w(t) is a Power Signal if 0 < P <
Decibels
and energy levels vary over orders of magnitude It is the ratio of two powers that is important
Complex Numbers
A complex number is a number composed of two real numbers, one which represents the real part and one which represents the imaginary part (originally created for defining roots of a polynomial)
z = x + jy
1
Real part
Imaginary part
Complex Plane
Imaginary part y z
Real part x
z = x2 + y 2
1
Why do we care?
Further we may always represent a signal as the sum (or integral) of sinusoidal signals Sinusoids are conveniently represented using complex numbers because of phase
x(t ) = A cos (c t + ) = A cos cos (c t ) A sin sin (c t )
(t ) = Ae j x
Using phasors
(t )e jct x(t ) = Re x
(t ) is complex, the true signal x(t) is always real. Note that while x The complex nature is simply a convenient mathematical construct to readily handle phase components Further note that (t ) = Ae j x
= A cos + jA sin This form will be particularly convenient when we analyze bandpass communication systems Note that while physical signals are always real, the imaginary part (t ) will have physical meaning as we will see later in the of x course
Phase
A cos (c t + )
Eulers Identities
e jx = cos( x ) + j sin( x )
e jx e jx sin( x ) = 2j
e jx + e jx cos( x ) = 2
This is termed a phasor representation of sinusoidal signals. We will use this representation extensively.
Note:
e jx = 1
Frequency
For a sinusoidal function the frequency is the inverse of the time it takes to complete one cycle (i.e., the period)
fo = 10
3sin(2 f o t + /9)
fo = 30
Frequency
Any signal can be expressed as the weighted sum of sinusoids of varying frequencies and phases. Many physical objects respond to EM waves based on the frequency of the wave. Thus, we are interested in the frequency of signals. Fourier Theory allows us to view the frequency content of a signal by decomposing the signal into an infinite sum (or an integral) of sinusoids. The Fourier Transform tells us how much of each frequency is needed.
Magnitude tells us the amount of each frequency The phase tells us how much cos vs sin
One of the most common mathematical tools for analyzing a signal or waveform is the Fourier Transform. The Fourier Transform provides us with information concerning the frequency content. This is useful for:
Determining bandwidth Demodulating frequency modulated signals Understanding how objects or systems will respond to a signal (Transfer Function) Equalization
We denote a Fourier Transform pair by: w( t ) W ( f ) The Fourier Transform always exists if w(t) is an Energy Signal We sum complex sinusoids of
w( t ) = F 1{W ( f )} = W ( f )e j 2ft df
We correlate the signal with a complex sinusoid of frequency f to determine how much of that frequency is present.
different frequencies f, weighting them by the amount of each frequency contained by the signal.
w ( t ) cos ( 2 f ) dt
we would lose any signal information related to sin(t) If we only used sin(t) in the transform, we would lose any signal information related to cos(t) Thus we must have both: ejt = cos(t)+ jsin(t)
The Fourier Transform may be thought of as a tool for looking at a signal from a different perspective
Consider how different a chair might look when viewed from different angles
W( f )=
T 2
sin( x ) = e jx e jx
T 2
1 e
j 2 ft
dt
e j f T e j f T = j2 f
) 2j
Time - Frequency
Time Waveform 1.4
5 4.5
1.2
Time Waveform
4 3.5 3 2.5
Spectrum As we expand the duration of the pulse, the signal changes less rapidly. Thus the signal has more low frequency content. Time expands Frequency compresses
2 1.5 1
0.4
0.2
0.5
5 t (sec) -->
10
15
0.5
1.5
2 Frequency
2.5
3.5
1.2
3.5 3 2.5
2 1.5 1
0.4
0.2
0.5 0
5 t (sec) -->
10
15
0.5
1.5
2 Frequency
2.5
3.5
We could continue to find transform pairs according to the definition, but this is inefficient In general, we compile a table of known transform pairs We also compile a table of simple rules for modifying transform pairs. Using the known pairs and transform properties we can find most transforms needed.
T [sinc( fT ) ]
T [sinc( fT ) ]
1 1 ( f )+ 2 j 2 f 1 j f
2
t tri T
u (t )
sgn ( t )
1
( t to )
(f)
e j 2 fto
1 f rect 2W 2W
e j ( f f o )
sinc(2Wt )
e jot + cos(2 ft + )
1 j 1 e ( f f o ) + e j ( f + f o ) 2 2
Note: Think of a constant as a sinusoid with an infinite period (f = 0). Does the transform make sense?
( t / to ) 2
to e
( f to ) 2
Linearity Time Delay Scale Change Duality Modulation Convolution Differentiation Integration
Function
Fourier Transform
a1w1 (t ) + a2 w2 (t )
a1W1 ( f ) + a2W2 ( f )
1 f W a a
( f ) e j 2 f
w( f )
1 j e W ( f f c ) + e j W ( f + f c ) 2
w(t )e j 2 fct
j 2 f c t Re g ( t ) e
(f
fc )
1 G ( f f c ) + G* ( f f c ) 2
dn w(t ) dt n
( j 2 f ) n W ( f )
( j 2 f ) 1W ( f ) + 1 W (0) ( f ) 2
w( s )ds
w1 (t ) * w2 (t ) w1 (t ) w2 (t )
W1 ( f )W2 ( f ) W1 ( f ) *W2 ( f )
( j 2 )
n
Multiplication by t
n RM Buehrer
d nW ( f ) df n
Summary
Today we have reviewed important concepts from signal and system theory that are particularly useful to communication system analysis Most importantly we will extensive use the Fourier Transform to analyze signals and systems in the frequency domain You should review
Next class we will examine useful functions termed singularity functions, review the concepts of Energy and Power Spectral Density, and the application of Fourier Theory to linear systems
Appendix
Additional Examples
e at , = 0,
t 0 t < 0
W(f)=
( a + j 2 f )t at j 2 ft = e dt u t e e dt () 0
e = a + j2 f
( a + j 2 f )t
1 = a + j2 f
W( f ) =
1 a 2 + ( 2 f ) 2
Example #2 (cont.)
Time Waveform
1 0.9 0.8 0.7 0.6 a = 0.5 0.5 0.4 0.3 0.2 0.1 0 a=1
Spectrum
2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 a=1 a = 0.5
5 Time (s)
10
0.2
0.4
0.6
1.4
1.6
1.8
As time waveform decreases more slowly, the more low frequency content in the wave.
1 1 j 2 ft w (t ) = ( f fc ) e df + ( f + f c ) e j 2 ft df 2 2
j 2 f c t
+ e 2
j 2 f c t
= c o s ( 2 f c t )