TOPICS IN DIGITAL SIGNAL PROCESSING
C. S. BURRUS and T. W. PARKS: DFTIFFT AND CONVOLUTION
ALGORITHMS
Rice University
T. W. PARKS and C. S. BUR RUS: DIGITAL FILTER DESIGN
Cornell University and Rice University
J. TREICHLER, R. JOHNSON, JR. and M . LARIMORE: THEORY AND
DESIGN OF ADAPTIVE FILTERS
Cornell University and Applied Signal Technology, Inc.
W. KOHN: DIGITAL CONTROL (in preparation)
Rice University
Digital Filter Design
T. W. Parks
School of Electrical Engineering
Cornell University
Ithaca, New York 14853
C. S. Burrus
Department of Electrical and Computer Engineering
Rice University
Houston, Texas 77251
A WILEY-INTERSCIENCE PUBLICATION
JOHN WlLEY & SONS, Inc.
New York Chichester Brisbane Toronto Singapore
Copyright 0 1987 by Texas Instruments Incorporated
Published by John Wiley & Sons, Inc.
All rights reserved. Published simultaneously in Canada.
Reproduction or translation of any part of this work
. beyond that permitted by Section 107 or 108 of the
1976 United States Copyright Act without the permission
of the copyright owner is unlawful. Requests for
permission or further information should be addressed to
the Permissions Department, John Wiley & Sons, Inc.
Library of Congress Cataloging-in-Publication Data:
Parks, T. W.
Digital filter design.
"A Wiley-Interscience publication."
Includes bibliographies and index.
1. Electric filters, Digital-Design and
construction. L Burrus, C. S. 11. Title.
TK7872.F5P37
ISBN 0-471 -82896-3
1987
621.3815'324
Printed in the United States of America
86-32500
To our parents
William and Mildred Parks
and
Aleta Huffman
Preface
This digital filter design book is addressed to the mathematician, scientist, or
engineer who has an understanding of continuous-time signals and who has
been introduced to discrete-time signal analysis.
The main topic of this book is the frequency-domain analysis and design
of linear, constant coefficient, digital filters. The book is divided into two
major parts: finite-duration impulse-response (FIR) filters and infinite-duration
impulse-response (IIR) filters. Each part consists of a complete, self-contained
treatment of the corresponding filter type. All aspects of each filter type are
discussed. Each part begins with a discussion of filter properties, which leads
into material on design of the filter to meet frequency-domain specifications.
This aspect of filter design is called the approximation problem and makes up a
major portion of the book. Each of the two parts concludes with a chapter on
realization pro-
implementation of the filter with fixed-point arithmetic-the
blem. The chapters on implementation both include a detailed design example
that presents a step-by-step design and implementation of a typical filter. The
design examples begin with the frequency-domain specifications for the filter
and conclude with a listing of the assembly language program for implementing
the filter on a signal-processing chip (the TMS32010 from Texas Instruments).
The book begins with an introductory chapter that reviews the concepts of
frequency-domain analysis of discrete-time systems and states the major
problems in digital filter design. The final chapter summarizes the main results
in the book with a discussion of the unique characteristics of the FIR and IIR
filter types. An appendix with listings of ten FORTRAN programs for filter
design is included.
This book may be used in several ways. For some applications one might
viii
Preface
turn to the appendix, run the appropriate design program to get the coefficients
of a filter that meets given frequency-domain specifications, then turn to the
listing in the design example, insert the coefficients in the listing, and run the
program on a TMS32010. If all goes smoothly in this process, one may not need
to read and completely understand the theory in the book. However, if, as often
happens, the problem one is faced with is not exactly covered by the programs in
the appendix, then with some reading of the theory, one can probably modify
the appropriate design program or write a special program to obtain the
appropriate filter coefficients. Even if the coefficients can be obtained from a
program in the appendix, the implementation in the design example may not be
exactly what one wants. For example, the filter may take too much time to
execute or may require too much memory or may have undesirable quantization
effects. Again, some time spent in reading the theory in the chapters on
implementation should allow the reader to develop an appropriate implem-
entation of the desired digital filter.
This book would not have been written without the support and encourage-
ment of Texas Instruments, Inc. We would especially like to thank Mike Hames,
who has always been ready with a smile and a helping hand when all of us
realized just how much work is involved in writing a book. Maridene Lemmon
has continued to patiently correct and improve our writing styles and has
carefully read through countless revisions of the manuscript. The engineers at TI
have read early versions of the text and helped correct our errors.
We would like to thank Professor H. W. Schiissler who helped us begin to
understand the issues in digital filter implementation when he was on leave at
Rice University. Some of our examples are taken from his notes. We would like
also to thank Cole Erskine for working out the two detailed design examples
and for providing the necessary TMS32010 code. Jim Kaiser and Dick Roberts
provided us with very thorough reviews of the manuscript and made several
good suggestions, which we have incorporated in the text.
We appreciate the long hours of reading put in by our graduate students
Doug Jones and Henrik Sorenson, who have made many good suggestions for
improving the book. Thanks also is given to the students at Rice University in
our digital signal-processing courses who have helped us develop this book over
the years.
Contents
Part One
INTRODUCTION
1. Introduction to Digital Filters
1.1. Properties of Discrete-Time Systems
1.2. Linear, Stationary, Discrete-Time Systems
1.3. Frequency Response and Transfer Functions
1.4. Digital Filter Design
1.4.1 The Approximation Problem
1.4.2 The Realization Problem
1.5. Properties of FIR and IIR Filters
References
Part Two
FINITE IMPULSE-RESPONSE (FIR) FILTERS
2. Properties of Finite Impulse-Response Filters
2.1. Frequency-Domain Description of FIR Filters
2.2. Linear-Phase FIR Filters
2.2.1. Four Types of Linear-Phase FIR Filters
2.2.2. Calculation of FIR Filter Frequency Response
2.2.3. Zero Locations for Linear-Phase FIR Filters
References