### 基本信息

- 原书名：Digital Signal Processing: system Analysis and Design, Second Edition
- 原出版社： Cambridge University Press

- 作者：
**(巴西)Paulo S.R.Diniz****Eduardo A.B.da Silva****Sergio L. Netto** - 丛书名：
**经典原版书库** - 出版社：机械工业出版社
- ISBN：
**9787111382539** - 上架时间：2012-6-11
- 出版日期：2012 年6月
- 开本：32开
- 页码：889
- 版次：2-1
- 所属分类：通信 > 通信技术理论与基础

教材

### 【插图】

### 内容简介

通信书籍

《数字信号处理：系统分析与设计(英文版·第2版)》将理论与实际有机融合，涵盖了数字信号处理(DSP)分析和设计的所有重要内容，提供了数字信号处理这一前沿技术领域难得的设计理念和方法。《数字信号处理：系统分析与设计(英文版·第2版)》不仅可作为高等院校电子、通信等专业本科生或研究生教材，还可作为工程技术人员DSP设计方面的参考用书。

第2版在上一版的基础上，扩充了滤波器组和小波分析的内容，新增了随机信号处理、谱估计和求解差分方程的内容，数学推导中给出了易于读者理解的步骤。第2版还提供了120个范例、20个案例研究、约400道练习题。此外，第2版还有一大新特点——每章末增加了一节“Do-it-Yourself”，使读者通过MATLAB实验获得解决实际信号处理问题的亲身体验。

### 作译者

EduardoA．B．daSilva巴西里约热内卢联邦大学(UFRJ)电子与计算机工程系和COPPE／UFRJ研究生院副教授。

Sergio L．Netto 巴西里约热内卢联邦大学(UFRJ)电子与计算机工程系和COPPE／UFRJ研究生院副教授。

### 目录

Preface

Introduction

1 Discrete-time signals and systems

1.1 Introduction

1.2 Discrete-time signals

1.3 Discrete-time systems

1.3.1 Linearity

1.3.2 Time invariance

1.3.3 Causality

1.3.4 Impulse response and convolution sums

1.3.5Stability

1.4 Difference equations and time-domain response

1.4.1 Recursive x nonrecursive systems

1.5 Solving difference equations

1.5.1 Computing impulse responses

1.6 Sampling of continuous-time signals

1.6.1 Basic principles

1.6.2 Sampling theorem

1.7 Random signals

### 前言

For many years, former students asked why the original text was not transformed into a book, as it presented a very distinct view that they considered worth publishing. Among the numerous reasons not to attempt such task, we could mention that there were already a good number of well-written texts on the subject; also, after many years of teaching and researching on this topic, it seemed more interesting to follow other paths than the painful one of writing a book; finally, the original text was written in Portuguese and a mere translation of it into English would be a very tedious task.

In later years, the second and third authors, who had attended the signal processing courses using the original material, were continuously giving new ideas on how to proceed. That was when we decided to go through the task of completing and updating the original text, turning it into a modem textbook. The book then took on its first-edition form, updating the original text, and including a large amount of new material written for other courses taught by the three authors up to 2002.

This second edition barely resembles the original lecture notes for several reasons. The original material was heavily concentrated on filter design and realization, whereas the present version includes a large amount of material on discrete-time systems, discrete transforms, spectral estimation, multirate systems, filter banks, and wavelets.

This book is mainly written for use as a textbook on a digital signal processing course for undergraduate students who have had previous exposure to basic linear systems, or to serve as a textbook on a graduate-level course where the most advanced topics of some chapters are covered. This reflects the structure we have at the Federal University of Rio de Janeiro,as well as at a number of other universities we have contact with. The second edition has a special feature designed for readers to test their learning by hands-on experience through so-called Do-it-yourself sections, with the aid of MATLAB. A Do-it-yourself section is included in all chapters of the book. The book also includes, at the end of most chapters, a brief section aimed at giving a start to the reader on how to use MATLAB as a tool for the analysis and design of digital signal processing systems. As in the first edition, we decided that having explanations about MATLAB inserted in the main text would in some cases distract the readers, making them lose focus on the subject.

A distinctive feature of this book is to present a wide range of topics in digital signal processing design and analysis in a concise and complete form, while allowing the reader to fully develop practical systems. Although this book is primarily intended as an undergrad-uate and graduate textbook, its origins on training courses for industry warrant its potential usefulness to engineers working in the development of signal processing systems. In fact,our objective is to equip the readers with the tools that enable them to understand why and how to use digital signal processing systems; to show them how to approximate a desired transfer function characteristic using polynomials and ratios of polynomials; to teach them why an appropriate mapping of a transfer function into a suitable structure is important for practical applications; and to show how to analyze, represent, and explore the trade-off between the time and frequency representations of deterministic and stochastic signals.For all that, each chapter includes a number of examples and end-of-chapter problems to be solved. These are aimed at assimilating the concepts, as well as complementing the text. In particular, the second edition includes many new examples and exercises to be solved.

Chapters I and 2 review the basic concepts of discrete-time signal processing and z transforms. Although many readers may be familiar with these subjects, they could benefit from reading these chapters, getting used to the notation and the authors' way of presenting the subject. In Chapter I we review the concepts of discrete-time systems, including the representation of discrete-time signals and systems, as well as their time-domain responses.Most important, we present the sampling theorem, which sets the conditions for the discrete-time systems to solve practical problems related to our real continuous-time world. The basic concepts of random signals are also introduced in this chapter, followed by the Do-it-yourself section aiding the reader to test their progress in discrete-time signals and systems.Chapter 2 is concerned with the z and Fourier transforms, which are useful mathematical tools for representation of discrete-time signals and systems. The basic properties of the z and Fourier transforms are discussed, including a stability test in the z transform domain.The chapter also shows how the analysis of random signals can benefit from the z-domain formulation.

Chapter 3 discusses discrete transforms, with special emphasis given to the discrete Fourier transform (DFT), which is an invaluable tool in the frequency analysis of discrete-time signals. The DFT allows a discrete representation of discrete-time signals in the frequency domain. Since the sequence representation is natural for digital computers, the DFT is a very powerful tool, because it enables us to manipulate frequency-domain infor-mation in the same way as we can manipulate the original sequences. The importance of the DFT is further increased by the fact that computationally efficient algorithms, the so-called fast Fourier transforms (FFTs), are available to compute the DFT. This chapter also presents real coefficient transforms, such as cosine and sine transforms, which are widely used in modem audio and video coding, as well as in a number of other applications. A discussion about orthogonality in transforms is also included. This section also includes a discussion on the several forms of representing the signals, in order to aid the reader with the available choices.

Chapter 4 addresses the basic structures for mapping a transfer function into a digital filter.It is also devoted to some basic analysis methods and properties of digital filter structures.The chapter also introduces some simple and useful building blocks widely utilized in some designs and applications.

Chapter 5 introduces several approximation methods for filters with finite-duration impulse response (FIR), starting with the simpler frequency sampling method and the widely used windows method. This method also provides insight to the windowing strategy used in several signal processing applications. Other approximation methods included are the maximally flat filters and those based on the weighted least-squares (WLS) method. This chapter also presents the Chebyshev approximation based on a multivariable optimization algorithm called the Remez exchange method. This approach leads to linear-phase transfer functions with minimum order given a prescribed set of frequency response specifications.This chapter also discusses the WLS-Chebyshev method which leads to transfer functions where the maximum and the total energy of the approximation error are prescribed. This approximation method is not widely discussed in the open literature but appears to be very useful for a number of applications.

Chapter 6 discusses the approximation procedures for filters with infinite-duration impulse response (IIR). We start with the classical continuous-time transfer-function approximations, namely the Butterworth, Chebyshev, and elliptic approximations, that can generate discrete-time transfer functions by using appropriate transformations. Two transformation methods are then presented: the impulse-invariance and the bilinear trans-formation methods. The chapter also includes a section on frequency transformations in the discrete-time domain. The simultaneous magnitude and phase approximation of IIR digital filters using optimization techniques is also included, providing a tool to design transfer functions satisfying more general specifications. The chapter closes by addressing the issue of time-domain approximations.

Chapter 7 introduces the basic concepts of classical estimation theory. It starts by describ-ing the nonparametric spectral estimation methods based on a periodogram, followed by the minimum-variance spectral estimator. The chapter continues with a discussion on model-ing theory, addressing the rational transfer function models and presenting the Yule-Walker equations. Several parametric spectral estimation methods are also presented, namely: the linear prediction method; the covariance method; the autocorrelation method; the Levinson-Durbin algorithm; and Burg's method. The chapter also discusses the Wiener filter as an extension of the linear prediction method.

Chapter 8 deals with basic principles of discrete-time systems with multiple sampling rates. In this chapter we emphasize the basic properties of multirate systems, thoroughly addressing the decimation and interpolation operations, giving examples of their use for efficient digital filter design. The chapter discusses many key properties of multirate systems, such as inverse operations and noble identities, and introduces some analytical tools, such as polyphase decomposition and the commutator models. In addition, we dis-cuss the concepts of overlapped block filtering, which can be very useful in some fast implementations of digital signal processing building blocks. The chapter also includes some discussion on how decimators and interpolators affect the properties of random signals.

Chapter 9 discusses some properties pertaining to the internal structure of filter banks, fol-lowed by the concept and construction of perfect reconstruction filter banks. The chapter also includes some analysis tools and classifications for the filter banks and transmultiplexers. This chapter presents several design techniques for multirate filter banks, including several forms of two-band filter banks, cosine-modulated filter banks, and lapped transforms.

Chapter 10 introduces the concepts of time-frequency analysis and the discrete wavelet transform. It also presents the multiresolution representation of signals through wavelet transforms and discusses the design of wavelet transforms using filter banks. In addition,some design techniques to generate orthogonal (as well as biorthogonal) bases for signal representation are presented. Several properties of wavelets required for their classification,design, and implementation are discussed in this chapter.

Chapter 11 provides a brief introduction to the binary number representations most widely used in the implementation of digital signal processing systems. The chapter also explains how the basic elements utilized in these systems work and discusses a particular, and yet instructive, type of implementation based on distributed arithmetic. Chapter 11 also includes the models that account for quantization effects in digital filters. We discuss several approaches to analyze and deal with the effects of representing signals and filter coefficients with finite wordlength. In particular, we study the effects of quantization noise in products,signal scaling that limits the internal signal dynamic range, coefficient quantization in the designed transfer function, and the nonlinear oscillations which may occur in recursive realizations. These analyses are used to indicate the filter realizations that lead to practical finite-precision implementations of digital filters.

In Chapter 12 we present some techniques to reduce the computational complexity of FIR filters with demanding specifications or specialized requirements. The first structure discussed is the lattice form, which finds application in a number of areas, including the design of filter banks. Several useful implementation forms of FIR filters, such as polyphase, frequency-domain, recursive running sum, and modified-sinc forms, are pre-sented to be employed as building blocks in several design methods. In particular, we introduce the prefilter and interpolation methods which are mainly useful in designing narrowband lowpass and highpass filters. In addition, we present the frequency-response masking approach, for designing filters with narrow transition bands satisfying more general specifications, and the quadrature method, for narrow bandpass and bandstop filters.

Chapter 13 presents a number of efficient realizations for IIR filters. For these filters, a number of realizations considered efficient from the finite-precision effects point of view are presented and their salient features are discussed in detail. These realizations will equip the reader with a number of choices for the design of good IIR filters. Several families of structures are considered in this chapter, namely: parallel and cascade designs using direct-form second-order sections; parallel and cascade designs using section-optimal and limit-cycle-free state-space sections; lattice filters; and several forms of wave digital filters.In addition, this chapter includes a discussion on doubly complementary filters and their use in the implementation of quadrature mirror filter banks.

This book contains enough material for an undergraduate course on digital signal pro-cessing and a first-year graduate course. There are many alternative ways to compose these courses; in the following we describe some recommendations that have been employed successfully in signal processing courses.

An undergraduate course in discrete-time systems or digital signal processing at junior level. This should include most parts of Chapters 1,2, 3, 4, and the nonparametric methods of Chapter 7. It could also include the noniterative approximation methods of Chapters 5 and 6, namely the frequency sampling and window methods described in Chapter 5, the analog-based approximation methods, and also the continuous-time to discrete-time transformation methods for IIR filtering of Chapter 6.