Scribd
Upload
56 views

SPFirst-L08 EV0

ece2011 at monash university signal processing first lecture notes if you are studying at monash uni electrical engineering than this will help

Uploaded by

kewancam
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
56 views

SPFirst-L08 EV0

ece2011 at monash university signal processing first lecture notes if you are studying at monash uni electrical engineering than this will help

Uploaded by

kewancam
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 26

Signal

Processing
First

Copyright Monash University 2009

Lecture8
Sampling
and
Aliasing

READING ASSIGNMENTS
ThisLecture:
Chap4,Sections41and42

OtherReading:
Recitation:StrobeDemo(Sect43)
NextLecture:Chap.4Sects.44and45

Copyright Monash University 2009

LECTURE OBJECTIVES
SAMPLINGcancauseALIASING
SamplingTheorem
SamplingRate>2(HighestFrequency)

Spectrumfordigitalsignals,x[n]
NormalizedFrequency

2f
2
Ts
fs

ALIASING

Copyright Monash University 2009

SYSTEMS Process Signals


x(t)

SYSTEM

y(t)

SIGNALPROCESSINGGOALS:
Changex(t)intoy(t)
Forexample,moreBASS

Improvex(t),e.g.,imagedeblurring
ExtractInformationfromx(t)
Copyright Monash University 2009

System IMPLEMENTATION
ANALOG/ELECTRONIC:
Circuits: resistors, capacitors, op-amps
x(t)

ELECTRONICS

y(t)

DIGITAL/MICROPROCESSOR
Convertx(t)tonumbers storedinmemory
x(t)

A-to-D

Copyright Monash University 2009

x[n]

COMPUTER

y[n]

D-to-A

y(t)

SAMPLING x(t)
SAMPLINGPROCESS

ContinuoustoDiscretetime
Convertx(t)tonumbers x[n]
nisaninteger;x[n]isasequenceofvalues
Thinkofnasthestorageaddressinmemory

UNIFORMSAMPLINGatt=nTs
IDEAL:x[n]=x(nTs)
x(t)

Copyright Monash University 2009

C-to-D

x[n]

SAMPLING RATE, fs
SAMPLINGRATE(fs)
fs =1/Ts
NUMBERofSAMPLESPERSECOND

Ts=125microsec fs =8000samples/sec
UNITSAREHERTZ:8000Hz

UNIFORMSAMPLINGatt=nTs=n/fs
IDEAL:x[n]=x(nTs)=x(n/fs)
x(t)

Copyright Monash University 2009

C-to-D

x[n] = x(nTs )
7

f 100Hz

f s 2 kHz

f s 500Hz

Copyright Monash University 2009

SAMPLING THEOREM
HOWOFTEN?
DEPENDSonFREQUENCYofSINUSOID
ANSWEREDbySHANNON/NYQUISTTheorem
ALSODEPENDSonRECONSTRUCTION

Copyright Monash University 2009

Reconstruction? Which One?


Given the samples, draw a sinusoid through the values

x[n ] cos(0.4 n )

When n is an integer
cos(0.4 n ) cos(2.4 n )

We choose to reconstruct the smallest frequency


.
Copyright Monash University 2009

10

STORING DIGITAL SOUND


x[n] isaSAMPLEDSINUSOID
Alistofnumbersstoredinmemory

EXAMPLE:audioCD
CDrateis44,100samplespersecond
16bitsamples
Stereouses2channels

Numberofbytesfor1minuteis
2X(16/8)X60X44100=10.584Mbytes
Copyright Monash University 2009

11

DISCRETE-TIME SINUSOID
Changex(t)intox[n]DERIVATION

x (t ) A cos( t )
x[n ] x ( nTs ) A cos( nTs )

x[n ] A cos((Ts )n )
x[n ] A cos( n )
Ts f DEFINE DIGITAL FREQUENCY
s

Copyright Monash University 2009

12

DIGITAL FREQUENCY

VARIESfrom0 to2,asfvariesfrom0 to

fs thesamplingfrequency
UNITSareradians,not rad/sec
DIGITALFREQUENCYisNORMALIZED

2f
Ts
fs
Copyright Monash University 2009

13

SPECTRUM (DIGITAL)
1
2

f
2
fs
f s 1 kHz

X*

0.2

1
2

2x 0.1

x[n ] A cos(2 (100)( n / 1000) )

Copyright Monash University 2009

14

SPECTRUM (DIGITAL) ???


f
2

fs
f s 100 Hz

1
2

1
2

2x

x[n ] A cos(2 (100)( n / 100) )

x[n] is zero frequency???

Copyright Monash University 2009

15

The REST of the STORY


Spectrumofx[n]hasmorethanonelinefor
eachcomplexexponential
CalledALIASING
MANYSPECTRALLINES

SPECTRUMisPERIODICwithperiod=2
Because

n ) A cos((
2 )n )
A cos(

Wechoosetoreconstructthesmallest
0 2
or -
Copyright Monash University 2009

16

ALIASING DERIVATION
OtherFrequenciesgivethesame

x1 (t ) cos(400 t ) sampled at f s 1000 Hz


n
x1[n ] cos(400 1000
) cos(0.4 n )

x2 (t ) cos(2400 t ) sampled at f s 1000 Hz


n
x2 [n ] cos(2400 1000
) cos(2.4 n )

x2 [n ] cos(2.4 n ) cos(0.4 n 2 n ) cos(0.4 n )


x2 [n ] x1[n ]
Copyright Monash University 2009

2400 400 2 (1000)


17

ALIASING DERIVATION2

Otherfrequenciesgivethesame
If x (t) A cos( 2 ( f f s )t )

n )
and we want : x[n] Acos(

n
t
fs

2 ( f f s ) 2 f 2 f s
then :

fs
fs
fs
2 f
Ts
fs
Copyright Monash University 2009

2
18

ALIASING CONCLUSIONS
ADDINGfsor2fs orfs totheFREQofx(t)
givesexactlythesamex[n]
Thesamples,x[n]=x(n/fs)areEXACTLYTHE
SAMEVALUES

GIVENx[n],WECANTDISTINGUISHfoFROM
(fo+ fs)or(fo+2fs)

Copyright Monash University 2009

19

NORMALIZED FREQUENCY
DIGITALFREQUENCY

2f
2
Ts
fs

Copyright Monash University 2009

20

SPECTRUM for x[n]


PLOTversusNORMALIZEDFREQUENCY
INCLUDEALL SPECTRUMLINES
ALIASES
ADDMULTIPLESof2
SUBTRACTMULTIPLESof2

FOLDEDALIASES
(tobediscussedlater)
ALIASESofNEGATIVEFREQS
Copyright Monash University 2009

21

SPECTRUM (MORE LINES)


f
2
fs

1
2

1.8

f s 1 kHz

Copyright Monash University 2009

1
2

X
-

0.2

1
2

1
2

1.8

x[n ] A cos(2 (100)( n / 1000) )

22

SPECTRUM (Aliasing case fs<2fmax)


f
2
fs

1
2

X*

2
f s 80Hz

2.5

1
2

1
2

1.5 - 0.5

1
2

1
2

0.5 1.5

1
2

2.5

x[n ] A cos(2 (100)( n / 80) )

Copyright Monash University 2009

23

SAMPLING GUI (con2dis)

Copyright Monash University 2009

24

SPECTRUM (FOLDING CASE)


f
2
fs

2
f s 1 25Hz

1
2

1.6

Copyright Monash University 2009

1
2

0.4

1
2

1
2

X*

0.4

1.6

x[n ] A cos(2 (100)( n / 125) )

25

FOLDING DIAGRAM

f
x[n] A cos 2 n
fs

f s 2000 Hz

x(t ) A cos(2 f t )
Copyright Monash University 2009

26

You might also like