Engineering/Signal Processing undoubtedly the most accessible book on digital signal processing vailable to the beginner. Using intuitive explanations and well-chosen +s, this book gives you the tools to develop a fundamental under- 3 of DSP theory. hor covers the essential mathematics by explaining the meaning and ince of the key DSP equations. Comprehensive in scope, and gentle ach, the book will help you achieve a thorough grasp of the basics ve gradually to more sophisticated DSP concepts and applications. ak begins with a complete explanation of the often misunderstood f periodic sampling. The introduction to the important discrete transform, and its fast Fourier transform (FFT) implementation, is :t lucid and illuminating explanation available anywhere. You will d extensive information on both finite impulse response (FIR) and impulse response (IIR) digital filters, as well as coverage of the of signal averaging, In addition, the book demystfies the abstruse f the Convolution theorem and complex signals. The practical uses us binary number formats are also carefully described and compared a collection of tricks-of-the-trade used by professionals to make DSP ms more efficient will help you apply DSP concepts successfully G. Lyons is a Systems Engineer with the Systems Integration Group + Ine, He has been involved with the design and testing of digital rocessing systems for the past fifteen years. He is the author of us articles on the topic and is a member of the IEEE and Eta Kappa electrical engineering honor society. ‘A Prentice Hall PTR Publication eee" I hi ign by Avande L. Peters by Mind of the Machine ISBN 0-201-63467-8 id J2UBIS jeySiq Suipueyssapuf, 3 Q a o yn ra 3 ga IDR Gm AGO)Understanding Digital Signal Processing ETS @ LTACTIGET. BIBLIOTECA UPV_/ EHU_UBURUTEGIA A Wobomawls de duo's Maar cod“ume” Understanding Digital Signal Processing Richard G. Lyons PRENTICE HALL PTR "A Prenice Hall PTR Pubestion Upper Sade iver, Ni 07458 wwewiphet.comSe Many of the designations used by manufacturets and sellers to distinguish their products are claimed as trademarks. Where ‘those designations appear inthis book and Addison-Wesley was aware ofa trademark claim, the designations have been printed ‘with initial capital eters, ‘The publisher offers discounts on this book when ordered in quantity for special sles, For more information, please contact ‘Corporate Sales Depertment Prentice Hall PTR 1 Lake Street Upper Saddle River, NJ.O7458 Library of Congress Cataloging-in-Publication Data Lyon, Rear G, 198 ‘Unsertnding dpa signal processing / RichardG. Lys. Includes blgraphi references and index. ISBN oan 1. Sigal procesting-Digtal tchiques. 1. Tie RSME 7 e@aaeraden oasis or Copyright 2001 By Prentice Hall PTR Al rights reserved. No part of this publication may be repro- ced stored in a retrieval system, or transmitted, in any form oF bby any means, electronic, mechanical, photocopying, recording, of otherwise, without the prior written permission of the pub: lisher. Printed in the United States of America, Published simul- taneously in Canada. 0201-53467.8 8 9-MA~01 ighth Printing, April 2001 DIEIDINCTAYSIIOINS | pammamannames I dedicate this book to my two daughters Julie and Meredith, I wish 1 could go with you; to my mother Ruth for making me finish my home- ‘work; to my father Grady who didn’t know what he started when he built that workbench in the basement; to my brother Ray for improving us all; to my brother Ken who succeeded where I failed; to my sister Nancy for running interference for us;to John Lennon for not giving up; to Dr. Laura Schlessinger for keeping us honest; to my advisor Glenn Caldwell and to the Iron Riders Motorcycle Club (Niles, CA) who keep me alive. Finally to Sigi Pardula, without your keeping the place running this book wouldn't existCONTENTS 5 3 Preface DISCRETE SEQUENCES AND SYSTEMS 1A. Discrete Sequences and Their Notation 12. Signal Amplitude, Magnitude, Power 13 Signal Processing Operational Symbols 14 Introduction to Discrete Linear Thme-Invariant Systems 13 Discrete Linear Systems 16 Timedavariant Systems 17 The Commutative Property of Linear Time-Invaiant Systems 18 Analyzing Linear Time-Invariant Systems PERIODIC SAMPLING. 21 Allasing: Signal Ambiquity inthe Frequency Domain. 22. Sampling Low-Pass Signals 23. Sampling Bandpass Signals 244 Spectral Inversion in Bandpass Sampling ‘THE DISCRETE FOURIER TRANSFORM. 31. Understanding the DPT Equation .. 32. DFT Symmetry 33 DFT Linearity 34 DFT Magnitudes 35. DFT Frequency Axis 36 DFT Shifting Theorem 37° Inverse DFT 38 DFT Leakage 39. Windows 340 DFT Scalloping Loss 311 DFT Resolution, Zero Staffing, and Frequency Domain Sampling 3A2_ DFT Processing Gain 313 The DFT of Rectangular Functions : 3.4 The DFT Frequency Response to a Complex Input 315. The DFT Frequency Response to a Real Cosine Input 3.16 The DFT Single-Bin Frequency Response toa Real Cosine Input ng 123 125Vil Unctentancing Dotat Signal Pocesing Contents 4. THE FAST FOURIER TRANSFORM 129 9 DIGITAL DATA FORMATS AND THEIR EFFECTS 38 41. Relationship ofthe FFT tothe DFT 130 8.1 Faed-Point Binary Formats 349 42 Hints on Using FFT in Practice 131 82 Binary Number Precision and Dynamic Range 356 43. FFT Software Programs 6 83 _Eifects of Finite Fixed-Point Binary Word Length 357 44 Derivation of the Radix-2 FFT Algorithm 136 94 Floating-Point Binary Formats us 45. FFT Input/Output Data index Bit Reversal Ms 95 lock Hosting Pint Binary Format: set 46 Radix? FFT Butterfly Structures 46 , 10 DIGITAL SIGNAL PROCESSING TRICKS 385 5 FINITE IMPULSE RESPONSE FILTERS 187 101 Frequency Translation without Multiplication 385 5 An Introduction to Finite Impulse Response FIR Filters 158, 10.2 High-Speed Vector-Magnitude Approximation 300 52 Convolution in FIR Filters 163 103 Data Windowing Tricks 406 53. Low-Pass FIR Filter Design ea 104 Fast Maaiplication of Complex Numbers att 54 Bandpass FIR iter Design fa 105 Efficiently Performing the FFT of Rel Sequences a2 55 _Highpass FIR Filter Design 393 106 Calculating the Inverse FFT Using the Forward FFT 925 56 Remez Exchange FIR Fier Design Method 194 107 Fast FFT Averaging 429 57 Hal-Dond FIR Fiters a 108 Simplified FIR Filter Structure 20 58 Phase Response of FIR Filters 19 109 Accurate A/D Converter Testing Technique 432 59. AGenerie Description of Diseete Convolution 204 10.10 Fast FIR Filtering Using the FFT 435 10.1 Caleulation of Sines and Cosines of Consecutive Angles =... 486 6 INFINITE IMPULSE RESPONSE FILTERS oa 10.12 Generating Normally Distributed Random Data 438 61 An Introduction to Infinite Impulse Response Filters 20 [APPENDIX A. THE ARITHMETIC OF COMPLEX NUMBERS “3 62. The Laplace Transform 223 63. The =-Transform 20 [A_ Graphical Representation of Real and Complex Numbers ..... 443, (64 Impulse Invariance IIR Fier Design Method 256 ‘82 _Anthmetc Representation of Complex Numbers 44 {65 Bilinear Transform UR Filter Design Method oa ‘A3__Anthmetic Operations of Complex Numbers or 66 Optimized IR Filer Design Method oe ‘A Some Practical Implications of Using Complex Numbers... 458, 67 Ptalls in Building tI Digital Filters. 286 68 Cascade and Parallel Combinations of Digital Filters 290 [APPtHDIX CLOSsD On OF Bee OMETc Sealey 455 69° A Brief Comparison of HR and FIR Filters 292 APPENDIX C, COMPLEX SIGNALS AND NEGATIVE FREQUENCY. ...458 7 ADVANCED SAMPLING TECHNIQUES 27 CA. Development of imaginary Numbers 460 71 Quadrature Sempling a7 C2 Representing Real Signals Using Complex Phasors 462 7.2 Quadrature Sampling with Digital Mixing. 301 C3. Representing Real Signals Using Negative Frequencies 467 73 Digital Resampling 303 C4 Complex Signals and Quadrature Mixing mn 8 SIGNAL AVERAGING sis [APPENDIX D. MEAN, VARIANCE, AND STANDARD DEVIATION .....76 81. Coherent Averaging 320 D1 Statistical Measures 476 82 _ Incoherent Averaging sz D2_ Standard Deviation, ot RMS, ofa Continuous Sinewave .... 480 83 Averaging Multiple Fast Fourier Transforms 330 D3 TheMean and Variance of Random Functions 431 84 Ritering Aspects of Time-Domain Averaging 340 D4 The Normal Probability Density Function 334 85 Exponential Averaging 3aiS Understanating Digital Signal Processing APPENDIX E, DECIBELS (a8 AND dBm) $86, ——— EX Using Lops Demin Rate Signa Power =. 86 GUE . 52. Some Useful Decibel Numbers an E3. Absolute Power Using Decibels ‘3 APPENDIX F, DIGITAL FILTER TERMINOLOGY ; 4 Learning Digital signal Processing Index 5 ee a7 Learning the fundamentals, and how to speak the language, of digital sig nal processing does not require profound analytical skills or an extensive background in mathematics. All you need is a little experience with ele- mentary algebra, knowledge of what a sinewave is, this book, and enthu- siasm, This may sound hard to believe, particularly if you've just flipped ‘through the pages ofthis book and seen figures and equations that appear rather complicated. The content here, you say, looks suspiciously like the material in technical journals and textbooks, material that is dificult 10 understand, Well, this is not just another book on digital signal processing, ‘This book's goal is to gently provide explanation followed by illustra- tion, rot so that you may understand the material, but that you must understand the material" Remember the first time you saw two people playing chess? The game probably appeared to be mysterious and con- fusing. As you now know, no individual chess move is complicated. Given a little patience, the various chess moves are easy to learn. The ‘game's complexity comes from deciding what combinations of moves to ‘make and when to make them. So it is with understanding digital signal processing. First we learn the fundamental rules and processes and, then, practice using them in combination. If learning digital signal processing is so easy, then why does the sub- ject have the reputation of being difficult to understand? The answer lies partially in how the material is typically presented in the literature. Its difficult to convey technical information, with its mathematical subtleties, in written form. Its one thing to write equations, but i's another matter altogether to explain what those equations really mean from a practical stancipoint, and that's the goal of this book. ‘Too often, written explanation of digital signal processing theory appears in one of two forms: either mathematical miracles occur and you are simply given a short and sweet equation without further explanation, or you are engulfed in a flood of complex variable equations and phrases Here we have the opportunity of expounding more claely what has already bee sid {Rene Descartes) xx LUnderstanaing Digital Signal Processing such as “itis obvious that,” “such that W(f) 2 0 f," and “with judicious application of the homogeneity property.” Authors usually do provide the needed information, but, too often, the reader must grab a pick and shovel, put on a miner’s helmet, and try to dig the information out of a ‘mountain of mathematical expressions. (This book presents the resuls of several fruitful mining expeditions.) How many times have you followed the derivation of an equation, after which the author states that he or she is going to illustrate that equation with a physical example—and this tums out to be another equation? Although mathematics is necessary to describe digital signal processing, I've tried to avoid overwhelming the reader because a recipe for technical writing that’s too rich in equations is hard for the beginner to digest? ‘The intent of this book is expressed ina popular quote from E.B. White in the introduction of his Elements of Style (New York: Macmillan Publishing, 1959) ‘Will Strunk) fel thet the reader was in serious trouble most of the time, aman floundering in a swamp, and that was the duty ofan ‘one attempting to write English to dean the swamp quickly and get bis man up on dey ground, or at leat throw him a rope. ve attempted to avoid the traditional instructor-student relationship, but, rather, to make reading this book like talking to a friend while walk- ing in the park. I've used just enough mathematics to develop a funda- mental understanding of the theory, and, then, illustrate that theory with examples, The Journey Learning digital signal processing is not something you accomplish; it’s a journey you take. When you gain an understanding of some topic, ques- tions arise that cause you to investigate some other facet of digital signal ‘processing, Armed with more knowledge, you're likely to begin exploring, further aspects of digital signal processing much like those shown in the following diagram. This book is your tour guide during the first steps of your journey. ‘You don't need a computer to learn the material in this book, but it ‘would help. Digital signal processing software allows the beginner to ver- "We need elucidation of the obvious more than investigation of the abscute” (Oliver Wendell Holmes) How canbe tect sare fees coca eal be erergee? (window tunetions ) How can OFT How con gta Str rmassuremamtaccumcy | foauncy responses bo ceimpeved? ‘morons? ow cane spect osarees ‘Saal be anayaoa? iy ate Sarcauses GET wadonieg WOK? ‘oaage? ro cy passband rl in ‘ital there? teoucton tats avoragages era? rie be roscnd “omercs srs | ow can special Convolution Figure Ft ify signal processing theory through trial and error* In particular, soft- ‘ware routines that plot signal data, perform the fast Fourier transform, and analyze digital filters would be very useful ‘As you go through the material in this Book, don’t be discouraged if your understanding comes slowly. As the Greek mathematician Menaechmus * "One must een by doing the thing for though you tink you know it you have no cer tainty sti you try Gophodes xixv ——— Understanding Oigital signal Processing curtly remarked to Alexander the Great, when asked for @ quick explana- tion of mathematics, “There is no royal road to mathematics.” Menaechmus ‘was confident in telling Alexander thatthe only way to learn mathematics is through careful study. The same applies to digital signal processing. Also, don’t worry if you have to read some of the material twice. While the con- ‘cepts in this book are not as complicated as quantum physics, as mysteri- ‘ous as the lyrics of the song “Louie Louie,” or as puzzling as the assembly instructions of a metal shed, they do get alittle involved. They deserve your attention and thought, So go slow and read the material twice if you have to; you'll be glad you did. If you show persistence, to quote a phrase from ‘Susan B. Anthony, “Failure is impossible” Coming Attractions Chapter 1 of this book begins by establishing the notation used through- out the remainder of our study. In that chapter, we introduce the concept of discrete signal sequences, show how they relate to continuous signals, and illustrate how those sequences can be depicted in both the time and frequency domains. In addition, Chapter 1 defines the operational sym- bols we'll use to build our signal processing system block diagrams. We conclude that chapter with a brief introduction to the idea of linear sys- tems and see why linearity enables us to use a number of powerful math- ‘ematical tools in our analysis. Chapter 2 introduces the most frequently misunderstood process in digital signal processing, periodic sampling. Although it’s straightfor- ward to grasp the concept of sampling a continuous signal, there are mathematical subtleties in the process that require thoughtful attention. Beginning gradually with simple examples of low-pass sampling and progressing to the interesting subject of bandpass sampling, Chapter 2 explains and quantifies the frequency-domain ambiguity (aliasing) asso- ciated with these important topics. The discussion there highlights the ‘power and pitfalls of periodic sampling. Chapter 3 is devoted to one of the foremost topics in digital signal pro- cessing, the discrete Fourier transform (DFT). Coverage begins with detailed examples illustrating the important properties of the DFT and how to interpret DFT spectral results, progresses to the topic of windows used to reduce DFT leakage, and discusses the processing gain afforded by the DFT. The chapter concludes with a detailed discussion of the vari ous forms of the transform of rectangular functions that the beginner is likely to encounter in the literature. That last topic is included there to clarify and illustrate the DFT of both real and complex sinusoids. Pretoce Chapter 4 covers the innovation that made the most profound impact on the feld of digital signal processing, the fast Fourier transform (FFT). There we show the relationship of the popular radix-2 FFT to the DFT, quantify the powerful processing advantages gained by using the FFT, demonstrate why the FFT functions as it does, and present various FFT implementation structures. Chapter 4 also includes a lst of recommenda- tions tohelp us when we use the FFT in practice. Chapter 5 ushers in the subject of digital filtering, Beginning with @ simple low-pass finite impulse response (FIR) filter example, we care- fully progress through the analysis of that filter's frequency-domain ‘magnitude and phase response. Next we learn how window functions affect and can be used to design FIR filters. The methods for converting low-pass FIR filter designs to bandpass and highpass digital filters are presented, and the popular Remez Exchange (Parks McClellan) FIR fil- ter design technique is introduced and illustrated by example. In that chapter we acquaint the reader with, and take the mystery out of, the process called convolution. Proceeding through several simple convolu- tion exemples, we conclude Chapter 5 with a discussion of the powerful convolution theorem and show why it's so useful as a qualitative tool in understanding digital signal processing, Chapter 6 introduces a second class of digital filters, infinite impulse response (IIR) filters. In discussing several methods for the design of UR filters, the reader is introduced to the powerful digital signal pro- cessing analysis tool called the z-transform. Because the z-transform is so closely related to the continuous Laplace transform, Chapter 6 starts by genily guiding the reader from the origin, through the properties, and onto the utility of the Laplace transform in preparation for learn- ing the z-transform, We'll see how TIR filters are designed and imple- mentee, and why their performance is so different from FIR filters. To indicate under what conditions these filters should be used, the chap- ter concludes with a qualitative comparison of the key properties of FIR and IK filters. Chapter 7 discusses two important advanced sampling techniques prominent in digital signal processing, quadrature sampling and digital resampling. In the chapter we discover why quadrature sampling is so useful when signal phase must be analyzed and preserved, and how this special sampling process can circumvent some of the limitations of tradi- tional periodic sampling techniques. Our introduction to digital resam- pling shows how we can, and when we should, change the effective sample rate of discrete data after the data has already been digitized. We've delayed the discussion of digital resampling to this chapter