Sampling of Continuous-Time Signals

Quote of the Day

Optimist: "The glass is half full."
Pessimist: "The glass is half empty."
Engineer: "That glass is twice as large as it needs to be."

Content and Figures are from Discrete-Time Signal Processing, 2e by Oppenheim, Shafer, and Buck, ©1999-2000 Prentice Hall Inc.
 Signal Types
• Analog signals: continuous in time and amplitude
– Example: voltage, current, temperature,…
• Digital signals: discrete both in time and amplitude
– Example: attendance of this class, digitizes analog signals,…
• Discrete-time signal: discrete in time, continuous in amplitude
– Example:hourly change of temperature in Austin
• Theory for digital signals would be too complicated
– Requires inclusion of nonlinearities into theory
• Theory is based on discrete-time continuous-amplitude signals
– Most convenient to develop theory
– Good enough approximation to practice with some care
• In practice we mostly process digital signals on processors
– Need to take into account finite precision effects
• Our text book is about the theory hence its title
– Discrete-Time Signal Processing

Copyright (C) 2005 Güne 351M Digital Signal Processing 2

r Arslan
 Periodic (Uniform) Sampling
• Sampling is a continuous to discrete-time conversion

-3 -2 -1 0 1 2 3 4

• Most common sampling is periodic

xn  x c nT     n  

• T is the sampling period in second

• fs = 1/T is the sampling frequency in Hz
• Sampling frequency in radian-per-second s=2fs rad/sec
• Use [.] for discrete-time and (.) for continuous time signals
• This is the ideal case not the practical but close enough
– In practice it is implement with an analog-to-digital converters
– We get digital signals that are quantized in amplitude and time
Copyright (C) 2005 Güne 351M Digital Signal Processing 3
r Arslan
 Periodic Sampling
• Sampling is, in general, not reversible
• Given a sampled signal one could fit infinite continuous signals
through the samples

0.5

-0.5

-1
0 20 40 60 80 100

• Fundamental issue in digital signal processing

– If we loose information during sampling we cannot recover it
• Under certain conditions an analog signal can be sampled without
loss so that it can be reconstructed perfectly

Copyright (C) 2005 Güne 351M Digital Signal Processing 4

r Arslan
 Sampling Demo
• In this movie the video camera is sampling at a fixed rate of
30 frames/second.
• Observe how the rotating phasor aliases to different speeds
as it spins faster.

pt   e  j2 fot
fo
 j2  n
pn  pnT   pn / fs   e fs

• Demo from DSP First: A Multimedia Approach by McClellan, Schafer, Yoder

Copyright (C) 2005 Güne 351M Digital Signal Processing 5
r Arslan
 Representation of Sampling
• Mathematically convenient to represent in two stages
– Impulse train modulator
– Conversion of impulse train to a sequence

s(t)

Convert impulse
xc(t) x train to discrete- x[n]=xc(nT)
time sequence

xc(t) x[n]
s(t)

t n
-3T-2T-T 0 T 2T3T4T -3 -2 -1 0 1 2 3 4

Copyright (C) 2005 Güne 351M Digital Signal Processing 6

r Arslan
 Continuous-Time Fourier Transform
• Continuous-Time Fourier transform pair is defined as

X c j    c
x t e  jt
dt


1
x c t    X c j e jt
d
2  

• We write xc(t) as a weighted sum of complex exponentials

• Remember some Fourier Transform properties
– Time Convolution (frequency domain multiplication)
x(t)  y(t)  X( j)Y( j)
– Frequency Convolution (time domain multiplication)
x(t)y(t)  X( j)  Y( j)
– Modulation (Frequency shift)
x(t)e jot  Xj   o 

Copyright (C) 2005 Güne 351M Digital Signal Processing 7

r Arslan
 Frequency Domain Representation of Sampling
• Modulate (multiply) continuous-time signal with pulse train:
 
x s t   x c t st    x t t  nT 
c s(t)   t  nT 
n   n  

• Let’s take the Fourier Transform of xs(t) and s(t)

1 2 
X s j   X c j   Sj  Sj      k s 
2 T k  

• Fourier transform of pulse train is again a pulse train

• Note that multiplication in time is convolution in frequency
• We represent frequency with  = 2f hence s = 2fs

1 
X s j    X c j  k s 
T k  

Copyright (C) 2005 Güne 351M Digital Signal Processing 8

r Arslan
 Frequency Domain Representation of Sampling
• Convolution with pulse creates replicas at pulse location:
1 
X s j    X c j  k s 
T k  
• This tells us that the impulse train modulator
– Creates images of the Fourier transform of the input signal
– Images are periodic with sampling frequency
– If s< N sampling maybe irreversible due to aliasing of images

X c j 
-N N

X s j  s>2N

3s -2s s -N N s 2s 3s

X s j  s<2N

3s -2s s -N N s 2s 3s

Copyright (C) 2005 Güne 351M Digital Signal Processing 9
r Arslan
 Nyquist Sampling Theorem
• Let xc(t) be a bandlimited signal with

X c ( j)  0 for   N

• Then xc(t) is uniquely determined by its samples x[n]= x c(nT)

if 2
s   2fs  2N
T

 N is generally known as the Nyquist Frequency

• The minimum sampling rate that must be exceeded is known
as the Nyquist Rate Low pass filter

X s j  s>2N

3s -2s s -N N s 2s 3s

X s j  s<2N

3s -2s s -N N s 2s 3s

Copyright (C) 2005 Güne 351M Digital Signal Processing 10
r Arslan
 Demo

Aliasing and Folding Demo

Samplemania from John Hopkins University

Copyright (C) 2005 Güne 351M Digital Signal Processing 11

r Arslan