Preface

This book was developed based on our teaching of undergraduate- and graduate-
level courses in digital signal processing over the past several years. In this book
we present the fundamentals of discrete-time signals, systems, and modern digital
processing as well as applications for students in electrical engineering, computer
engineering, and computer science. The book is suitable for either a one-semester
or a two-semester undergraduate-level course in discrete systems and digital signal
processing. It is also intended for use in a one-semester first-year graduate-level
course in digital signal processing.
It is assumed that the student has had undergraduate courses in advanced calcu-
lus (including ordinary differential equations) and linear systems for continuous-time
signals, including an introduction to the Laplace transform. Although the Fourier
series and Fourier transforms of periodic and aperiodic signals are described in Chap-
ter 4, we expect that many students may have had this material in a prior course.
Balanced coverage of both theory and practical applications is provided. A large
number of well-designed problems are provided to help the student in mastering the
subject matter. A solutions manual is available for download for instructors only.
Additionally, Microsoft PowerPoint slides of text figures are available for instructors
on the publisher'S website.
In the fourth edition of the book, we have added a new chapter on adaptive
filters and have substantially modified and updated the chapters on muItirate digital
signal processing and on sampling and reconstruction of signals. We have also added
new material on the discrete cosine transform.
In Chapter 1 we describe the operations involved in the analog-to-digital con-
version of analog signals. The process of sampling a sinusoid is described in some
detail and the problem of aliasing is explained. Signal quantization and digital-to-
analog conversion are also described in general terms, but the analysis is presented
in subsequent chapters.
Chapter 2 is devoted entirely to the characterization and analysis of linear time-
invariant (shift-invariant) discrete-time systems and discrete-time signals in the time
domain. The convolution sum is derived and systems are categorized according to
the duration of their impulse response as a finite-duration impulse response (FIR)
and as an infinite-duration impulse response (IIR). Linear time-invariant systems
characterized by difference equations are presented and the solution of difference
equations with initial conditions is obtained. The chapter concludes with a treatment
of discrete-time correlation.
xvii
xviii Preface

The z-transform is introduced in Chapter 3. Both the bilateral and the unilateral
z-transforms are presented, and methods for determining the inverse z-transform
are described. Use of the z-transform in the analysis of linear time-invariant systems
is illustrated, and important properties of systems, such as causality and stability, are
related to z-domain characteristics.
Chapter 4 treats the analysis of signals in the frequency domain. Fourier series
and the Fourier transform are presented for both continuous-time and discrete-time
signals.
In Chapter 5, linear time-invariant (LTl) discrete systems are characterized in the
frequency domain by their frequency response function and their response to periodic
and aperiodic signals is determined. A number of important types of discrete-time
systems are described, including resonators, notch filters, comb filters, all-pass filters,
and oscillators. The design of anum ber of simple FIR and IIR filters is also considered
In addition, the student is introduced to the concepts of minimum-phase, mixed-
phase, and maximum-phase systems and to the problem of deconvolution.
Chapter 6 provides a thorough treatment of sampling of continuous-time signals
and the reconstruction of the signals from their samples. Our coverage includes
the sampling and reconstruction of bandpass signals, the sampling of discrete-time
signals, and AID and DIA conversion. The chapter concludes with the treatment of
oversampling AID and D/A converters.
The DFf, its properties and its applications, are the topics covered in Chapter 7.
Two methods are described for using the DFf to perform linear filtering. The use of
the DFf to perform frequency analysis of signals is also described. The final topic
treated in this chapter is the discrete cosine transform.
Chapter 8 covers the efficient computation of the OFT. Included in this chapter
are descriptions of radix-2, radix-4, and split-radix fast Fourier transform (FFT)
algorithms, and applications of the FFTalgorithms to the computation of convolution
and correlation. The Goertzel algorithm and the chirp-z transform are introduced
as two methods for computing the DFf using linear filtering.
Chapter 9 treats the realization of IIR and FIR systems. This treatment includes
direct-form, cascade, parallel, lattice, and lattice-Iadderrealizations. The chapter also
examines quantization effects in a digital implementation of FIR and IIR systems.
Techniques for design of digital FIR and IIR filters are presented in Chapter 10.
The design techniques include both direct methods in discrete time and methods in-
volving the conversion of analog filters into digital filters by various transformations.
Chapter 11 treats sampling-rate conversion and its applications to multirate dig-
ital signal processing. In addition to describing decimation and interpolation by
integer and rational factors, we describe methods for sampling-rate conversion by
an arbitrary factor and implementations by polyphase filter structures. This chap-
ter also treats digital filter banks, two-channel quadrature mirror filters (QMF) and
M-channel QMF banks.
Linear prediction and optimum linear (Wiener) filters are treated in Chapter 12.
Also included in this chapter are descriptions of the Levinson-Durbin algorithm
and Schur algorithm for solving the normal equations, as well as the AR lattice and
ARMA lattice-ladder filters.
 Preface xix

Chapter 13 treats single-channel adaptive filters based on the LMS algorithm

and on recursive least squares (RLS) algorithms. Both direct form FIR and lattice
RLS algorithms and filter structures are described.
Power spectrum estimation is the main topic of Chapter 14. Our coverage in-
cludes a description of non parametric and model-based (parametric) methods. Also
described are eigen-decomposition-based methods, including MUSIC and ESPRIT.
A one-semester senior-level course for students who have had prior exposure to
discrete systems can use the material in Chapters 1 through 5 for a quick review and
then proceed to cover Chapters 6 through 10.
In a first-year graduate-level course in digital signal processing, the first six chap-
ters provide the student with a good review of discrete-time systems. The instructor
can move quickly through most of this material and then cover Chapters 7 through
11, followed by selected topics from Chapters 12 through 14.
Many examples throughout the book and approximately 500 homework prob-
lems are included throughout the book. Answers to selected problems appear in the
back of the book. Many of the homework problems can be solved numerically on a
computer, using a software package such as MATLAB®. Available for use as a self-
study companion to the textbook is a student manual: Student Manual for Digital
Signal Processing with MATLAB®. MATLAB is incorporated as the basic software
tool for this manual. The instructor may also wish to consider the use of other sup-
plementary books that contain computer-based exercises, such as Computer-Based
Exercisesfor Signal Processing Using MATLAB (Prentice Hall, 1994) by C. S. Burrus
et at.
The authors are indebted to their many faculty colleagues who have provided
valuable suggestions through reviews of previous editions of this book. These in-
clude W. E. Alexander, G. Arslan, Y. Bresler, I Deller, F. DePiero, V. Ingle, IS. Kang,
C. Keller, H. Lev-Ari, L. Merakos, W. Mikhael, P. Monticciolo, C. Nikias, M. Schet-
zen, E. Serpedin, T. M. Sullivan, H. Trussell, S. Wilson, and M. Zoltowski. We
are also indebted to R. Price for recommending the inclusion of split-radix FFT al-
gorithms and related suggestions. Finally, we wish to acknowledge the suggestions
and comments of many former graduate students, and especially those by A. L. Kok,
I Lin, E. Sozer, and S. Srinidhi, who assisted in the preparation of several illustrations
and the solutions manual.

JOHN G. PROAKIS
DIMITRIS G. MANOLAKIS