In the last several years many books have been devoted to the study of digital signal processing. This is partly because today’s state-of-the-art technology, such as signal processing chips, permits a constantly increasing number of useful digital signal processing applications to be made inexpensively and quickly.
This is a textbook on one-dimensional digital signal processing, written in the best tradition of Oppenheim and Shafer [1] and Rabiner and Gold [2]. As such, it does not differ too much from Oppenheim and Shafer’s book. (That book has some introductory chapters on two-dimensional digital signal processing.)
One valuable feature of the book as a whole is that most of the different problems are treated by many different approaches, which permits the reader to appreciate as many distinct mathematical tools in use in the field as can be covered in a reasonable amount of space. Other global features are the algorithmic approach to most of the theoretical formulations and the extensive use of helpful examples for practically every topic discussed. These features and the extremely clear presentation help bring the hard field of signal processing down to earth.
The book contains the following 11 chapters:
:9N(1) Digital-Signal Processing
(2) Discrete Time Signals and Systems
(3) The z-Transform
(4) Fourier Analysis of Discrete-Time Signals and Systems
(5) Fast Algorithms for the Discrete Fourier Transform
(6) The Approximation Problem for Digital Filters (with two appendices: 6A--Design of Analog Low Pass Filter Prototypes and 6B--The Chebyshev Window Function)
(7) Least-Square Filter Design (plus appendix 7A--Quadratic Forms)
(8) Internal Descriptions for Digital Filters (and appendix 8A--Spectral Representation and Functions of a Matrix)
(9) Finite Length Register Effects in Fixed Point Digital Filters
(10) Digital Processing Structures (with appendix 10A--Orthogonal Matrices and All-Pass Filters)
(11) Spectral Estimation.
The first chapter is a brief but consistent answer to the newcomer’s legitimate question about digital signal processing: what for. The second, third, and fourth chapters are classical ones but also contain the above-mentioned global features. Although, as stated in the first few pages, computational efficiency is now of primary interest, chapter 5 is a little superficial as compared to the rest of the book.
Chapter 6 is a good introduction to the approximation problem for digital filters. This chapter and the related problems will help the alert reader to gain skill in the design of practical digital filters. However, the addition of some design problems would probably make this important chapter more practical.
Chapter 7 is concerned with some statistical representations of discrete-time signals and with the optimization problem for digital filters, specifically the least-square method. At the end of the chapter there is a brief introduction to adaptive filtration. The eighth chapter deals with some methods of digital filter representation that are now gaining increasing popularity, namely, signal flowgraphs and state variable descriptions.
Chapter 9 constitutes the keystone of the book. Here is an excellent and very deep treatment of the subject. No other known book gives such deep insight into almost all fixed point effects that arise in the design and use of the fixed point digital filters that constitute most of the today’s real-time hardware implementations. Included are quantization noise, limit cycles and overflow oscillations and methods to overcome them, roundoff noise, scaling rules and geometrical representation of scaling, minimum roundoff noise design, optimal structures, coefficient sensitivity, and deadband effect.
Chapter 10 is a careful introduction to some hardware structures for digital filters from a global theoretical viewpoint. Preference is given to the distributed arithmetic implementation, block processing digital filters, multirate digital filters, and a good introduction to orthogonal filter structures.
The subject of the last chapter is one of the most important in the area of signal processing, namely, the spectral estimation. Included are periodograms, autoregressive models (with special emphasis on the Levinson-Durbin and Burg algorithms), and ARMA spectra.
The book is self-contained with regard to most of the mathematical tools used. Although mostly digital filter oriented, this is an excellent introductory textbook for the whole field of digital signal processing and as such can be used as an alternative to Oppenheim and Shafer’s text. In fact, this is probably the first textbook to assume this responsibility.
