Scribd
Upload
97 views

Digital Signal Processing: Chapter 4: Sampling and Reconstruction

This document provides an overview of sampling and reconstruction in digital signal processing. It discusses how continuous time signals can be processed in discrete time by sampling, and then reconstructed back to continuous time. The key aspects covered include: 1) Sampling rate conversion which allows interfacing between systems operating at different sampling rates, using techniques like downsampling/decimation to reduce the rate and upsampling/interpolation to increase it. 2) Downsampling by an integer factor M works by keeping every Mth sample and extending the sampling period. Upsampling inserts L-1 zeros between samples to increase the rate by L. 3) Frequency domain analysis shows downsampling compresses the spectrum and upsampling expands it

Uploaded by

aqsa
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
97 views

Digital Signal Processing: Chapter 4: Sampling and Reconstruction

This document provides an overview of sampling and reconstruction in digital signal processing. It discusses how continuous time signals can be processed in discrete time by sampling, and then reconstructed back to continuous time. The key aspects covered include: 1) Sampling rate conversion which allows interfacing between systems operating at different sampling rates, using techniques like downsampling/decimation to reduce the rate and upsampling/interpolation to increase it. 2) Downsampling by an integer factor M works by keeping every Mth sample and extending the sampling period. Upsampling inserts L-1 zeros between samples to increase the rate by L. 3) Frequency domain analysis shows downsampling compresses the spectrum and upsampling expands it

Uploaded by

aqsa
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/ 29

Digital Signal Processing

Chapter 4: Sampling and reconstruction

Dr. Muhammad Shahzad Younis

Electrical Engineering Department


SEECS, NUST
[email protected]
Discrete-time processing of continuous time
signals
Given a continuous time signal, process it in discrete time and
reconvert to a continuous time signal at output

Overall system properties depend on discrete time system used and


the sampling rate
Note in general C/D and D/C can use different sampling rates but for
now we assume them to be the same
Linear Time invariant Discrete-Time
Systems
Ex 4.3 : DT processing of CT signals

Heff (jW)
CT processing of DT Signals
Ex:4.7 :CT processing of DT Signals
Ex:4.7 :CT processing of DT Signals
Ex:4.8 :MA Filter with non-integer
delay(5th Order, M = 5)
Ex:4.8 :MA Filter with non-integer
delay(5th order, M = 5)
Ex:4.8 :MA Filter with non-integer
delay(6th Order, M = 5)
Multirate Signal Processing
Interoperability of systems designed to operate at
different sampling frequencies
Multirate signal processing involves design of sampling
rate converters
D.T. System 1
@ fs1 X
Compatibility
D.T. System 2
@ fs2
?

Sample rate
converter
fs1 fs2

Sampling rate decrease (downsampling/decimation)


Sampling rate increase (upsampling/interpolation)
Sampling Rate Conversion-Underlying
Principle
Intuitive Approach
starting from the discrete-time sequence x1[n], recover the
band-limited continuous time signal xc(t) with sampling
interval T1
resample xc(t) at the new desired rate T2, s.t. x2[n] = xc(nT2)

Non-practical approach
Non-ideal C/D and D/C converters, etc
Wastage of resources

Solution
consider sampling rate conversion directly in discrete-
domain
Resampling

13
Sampling Rate Reduction/ Downsampling
(Integer Factor)

M
x[n] xd[n] = x[nM]
sampling interval T sampling interval Td = MT

x[n]

M=2

xd[n]
Sampling Rate Reduction/ Downsamlpling
(Integer Factor)

M
x[n] xd[n] = x[nM]
sampling interval T sampling interval Td = MT

1
n
M p[n]

x[n] Remove n
X 1
Zeros xd[n]
n x[n]
1
n
M
Downsampling-Freq. Domain Analysis
Downsampling-Freq. Domain Analysis
p[n]
Remove
x[n] X M-1 zeros xd[n]
x[n]


X e jw

1 Removal of zeros =
Scaling of time axis
2p -wN wN 2p
Scaling of frequency axis
Pe jw

2p / M

X d e jw

2p -2p/M 2p/M 4p/M 2p 1/ M


X e
jw

1/ M -2p 0 2p

2p 0 2p/M 2p
Downsampling-Examples

2p
Ws 4W N
T
Decimation-Examples

Decimation
Sampling rate increase/Upsampling
Desired : xi[n] = xc[nTi], given x[n] = xc[nT], where Ti = T/L
We might not have samples existing at each nTi
Expanding a digital signal x[n] by a factor of L is to include L 1
zeros between its samples
Upsampling-Frequency Domain
Upsampling-Frequency Domain
xi[n] = xc[nTi]

Desired spectrum
Upsampling-Frequency Domain

Indeed this system satisfies the requirement and effectively fills in


(interpolates) the missing values and is thus named interpolator

Desired spectrum
Interpolation-Discrete Time

x[n]

xe[n]

xi[n]
Linear Interpolation
Linear Interpolation
Sampling Rate Conversion Non-
Integer Factor
Ex. 4.9: (L=2, M=3)

You might also like