Digital Signal Processing

CSE - 610
Lecture # 1: Introduction

Dr. Muhammad Shehzad Hanif

[email protected]

Department of Mechatronics and Control Engineering

University of Engineering and Technology, Lahore
 What is a signal?
 A signal conveys information about the state or
behavior of a physical system
 It is a measured quantity that varies with time (or
position)

 Examples:
 Voltage: Represented as a function over time  1D signal
 Image signal: Represented as an intensity function of two spatial
variables  2D signal
 Video signal: A sequence of images spanning over a period of time
3D signal

2
 Signal Processing

Signal processing is concerned with the

representation, transformation, and manipulation of
signals and the information they contain.

3
 Types of signals

 Continuous-Time (CT) or Analog signal:

Example: Voltage, Current, Speech signal, etc.

 Discrete-Time (DT) signal:

Example: Daily stock market price, Daily average
temperature, Sampled continuous signals

4
 Discrete-Time (DT) signal: Example
 Stock market data

 Population

5
 Discrete-Time (DT) signal: Example
 Temperature measurements

6
 Discrete-Time (DT) signal: Example
 Sampled continuous time signal

Speech signal and its sampled version

7
 Discrete Time Signal

Discrete time signal x[n]

8
 Discrete Time System

9
 Discrete Time Signal Processing
(DTSP)
 Discrete time processing of continuous signals

10
 Digital Signal Processing (DSP)
 Digital signals processing (DSP) is derived from
DTSP

11
 Discrete Time Signal

12
 Digital Signal = Discrete Time
Discrete Amplitude Signal

13
 Sampling of Signal: Ambiguity

14
 DSP Scheme
 Converting analog signal into a digital signal
 Perform signal processing operations in the digital
form
 Convert back the digital signal to analog one when
necessary
Analog Analog
Input Output

Analog DSP Analog

ADC DAC Filter
Filter Processor

15
 Why Process The Signals Digitally?
 Digital data storage and transmission is more effective than in
the analog form

 Flexibility: Processing function can be modified or adjusted

 Can implement very complex processing functions

 Speed of digital operations tends to grow rapidly with the years

of technical progress

 A very high accuracy and reliability is possible

 Dynamic range can be increased

 Simultaneous (Parallel) processing

16
 DSP Market
DSP Revenues

25 34.7% CAGR 50%

20 40%
Revenue ($B)

% Change
15 30%

10 20%

5 10%

0 0%
98 99 00 01 02 03 04
Year
Revenue ($B) % Change

17
Source: http://bwrc.eecs.berkeley.edu/People/Grad_Students/czhong/documents/Opportunity%20in%20DSP%20ver%204.doc
 Why Do DSP Processors Need to Do
Well?
Most DSP tasks require:
 Repetitive numeric calculations
 Attention to numeric fidelity
 Fixed- vs. floating-point
 Standards
 High memory bandwidth
 Streaming data
 Real-time processing
 Processors must perform these tasks efficiently while
minimizing:
 Cost
 Power consumption
 Memory use
 Development time

18
 Benchmark
Implementation of Complex Block FIR Filter
DSP vs. High Performance CPU
(lower is better)

19
 Example DSP Applications
 Digital cell phones  Satellite communications
 Automated inspection  Seismic analysis
 Vehicle collision  Secure communications
avoidance  Tapeless answering
 Voice -over-Internet machines
 Motor control  Sonar
 Consumer audio  Cordless phones
 Voice mail  Digital cameras
 Navigation equipment  Modems (POTS, ISDN,
 Audio production cable, ...)
 Videoconferencing  Noise cancellation
 Toys, games consoles  Medical ultrasound
 Music synthesis, effects  Patient monitoring
 Radar
20
And more to come...
 Original speech signal
Speech Processing
Lowpass filter

Downsample
Highpass filter

21
 Echo Cancellation
 Echo-path in switched telephony network

 Echo-path in hands-free communication

22
 Echo Cancellation (2)
 Solution: Recreate the echo signal and subtract it
from the sent signal

23
 Equalization
 Selectively enhance/attenuate
some parts of the frequency
spectrum
 Applications
 Coding & compression
 Room simulation
 Echo or chorus effects

24
 Time Series Prediction
 Predict the next value of a time series from its
past samples:

x[n  1]  f ( x[0], x[1], , x[n])

 Solution: Prediction filter

25
 Speech Transmission

26
 Image Processing

27
 Image Processing

28
 Signal Interpretation
 The objective of the processing is not to obtain an
output signal but to obtain a characterization of
the input signal
 Example: Speaker Identification
Database of
Attributes

Signal Attribute
Attributes
Interpretation Matching

Complete System = Signal Processing + Pattern Recognition

Speaker
29
 Course Outline
 Introduction: Continuous Time (CT) and Discrete Time (DT)
signals and systems
 Basics of signal conditioning & processing
 Signal representation
 Sampling of CT signals (A/D and D/A conversion, Multi-rate signal
processing)
 Linear Time Invariant (LTI) systems
 Convolution
 Fourier Transform
 Z- Transforms
 Analysis of LTI systems
 Filter design techniques and implementation (FIR and IIR filter
structures)
 Discrete Fourier Transform (DFT)
 Computation via FFT
 Applications of DFT
30
 Course Outline
 Pre-requisite: Introductory knowledge of signals and systems

 Course meeting times

Lectures: 1 session/week (3 hours on Wednesday 2pm - 5pm)

 Course teacher
Dr. Muhammad Shehzad Hanif
Mechatronics Lab, 1st Floor, Research Center,
University of Engineering and Technology, Lahore, Pakistan.
Email: [email protected]

 Office hours
Monday - Saturday 0800 – 1500 HRS

31
 Group for Course

 Subscribe to digital signal processing group by sending an

email to [email protected]

 Group website is
http://groups.yahoo.com/group/cse_610_fall_2010/

 Lectures slides, assignments (computer/written), solution

to problems, research papers, projects, and
announcements will be uploaded to group repository and
will be notified to all group members through email

32
 Grading & Evaluation
Activities Percentage
Quizzes 20%
Mid Term 30%
Final 40%
Homework 5%
Computer Assignments 5%
 A minimum of 75% attendance is necessary in order to avoid
drop-out from the course
 Homework: Problem sets will be assigned every one to two
weeks, to be turned in at the beginning of the class when they
are due
 Computer assignments: A component of the homework will be
computer assignments using MATLAB. You are required to send
your MATLAB codes and results to [email protected]
when they are due
33
 Text and Reference Books
 Text book
Discrete-Time Signal Processing
by Alan V. Oppenheim, Ronald W. Schafer & John R. Buck. 2nd
Edition, Pearson Education - Prentice Hall, 1999
(Local edition is available)
 Reference books
 Digital Signal Processing: Principles, Algorithms and
Applications
by J. G. Proakis and D. G. Manolakis. 3rd Edition, Prentice-
Hall, 1995
 Digital Signal Processing, A Computer-Based Approach
by S. K. Mitra. 2nd Edition, McGraw-Hill, 2001 (contains
many MATLAB examples)
34
 References

 Chapter # 1, Discrete-Time Signal Processing

by Alan V. Oppenheim, Ronald W. Schafer &
John R. Buck. 2nd Edition, Pearson Education -
Prentice Hall, 1999

50