Blinchikoff H.J., Zverev A.I. Filtering in the Time and Frequency Domains

SciTech Publishing, 2001. — 518 p.Filtering, a fundamental signal process in electronics, includes such diverse operations as channeling, demodulating, equalizing, detecting, decoding, phasesplitting, integrating, and differentiating. Consequently, it is increasingly important for system-oriented engineers as well as for equipment designers to know the capabilities and limitations of filters in order to ensure effective system design and synthesis.
Books on filtering are written primarily for the filter designer and are therefore too detailed for the broader audience. The non-specialist wants a treatment of filtering that deals with pertinent questions related to insertion loss, group delay, delay equalization, trade-offs between attenuation and group delay, trade-offs between attenuation and impulse response overshoots, and the feasibility of specified responses in both time and frequency domains.
Accordingly, the first goal of this book is to provide helpful information to the filter designer and to the electronics engineer who does not specialize in filter design, so that each can perform his task more efficiently and knowledgeably. The second goal is to show that the area of filtering includes devices and concepts that historically have not been treated within this discipline.
Much of the material presented here was given in course form to our colleagues during the years 1969-1975. Most participants were engineers interested in the general aspects of filtering. That this course was well received encouraged us to write this book, a text for electrical engineering students as well as for practicing engineers in industrial applied science schools. Our intent is to establish relationship between industrial thinking and university teaching.
Thus in this book the systems engineer can find what the filter can do for him. Knowledge of the response trade-offs and the required effort for filter design and synthesis is very important in the decision-making process associated with electronic systems. To aid him in this task we provide the necessary denormalization equations for the design curves in Handbook of Filter Synthesis by A. Zverev (Wiley, 1967). This allows the significant responses of the various filter types to be easily determined. The two books complement each other and offer more complete picture of the filtering possibilities.
To establish the appropriate mathematical relationships we use four basic transforms-Fourier, Laplace, Hilbert, and z. The first two relate the frequency-domain and time-domain characterizations while the Hilbert transform relates attenuation and phase (group delay) for minimum-phase networks. The z-transform is the tool for analyzing discrete-time systems.
Although the word "filter" usually suggests a device designed to exhibit specified magnitude response, the widespread use of pulsed systems (such as radar, television, and communication) and the newest signal processing schemes have greatly expanded the concept of a filter.
The pulsed systems verified that the frequency responses were not powerful enough to adequately describe the important system behavior that occurred in the time domain. As a result, the impulse response emerged as the more important system characterization. However, the frequency-domain approach to analysis and design has such justifiably deep roots that it is still difficult to appreciate that in many cases optimum system performance can be obtained only through time domain.
The time-domain approach and the extension of the filtering concept exemplified by the matched filter, whose output signal-to-noise ratio is optimized when its impulse response is the mirror image of the transmitted pulse. This true filter in every sense of the word, yet frequency need not be mentioned classify and design it. The filter is signal selective rather than frequency selective. No attempt is made to preserve the shape of the input signal at the output; rather, effort is directed toward signal detection.
Filtering aspects are discussed in terms of lumped-constant networks because these components and their analysis are familiar. We generally avoid explicit discussion of the hardware and concentrate instead on the function of hardware, while also including many applications and implementations. The solid state engineer can use this material for designing and analyzing active filters, whose basic element is the operational amplifier. By using lumped equivalent circuits for his cavities and distributed elements, the microwave engineer transfer the data presented here to his own expertise. The designer of digital circuits, surface wave devices, and optical systems will also find this information beneficial.Time-Domain Analysis
Frequency-Domain Analysis
Linear System Responses
Frequency Transformations
All-Pass Functions
Finite-Q Elements and Predistortion
Optimum Linear Filtering
Time-Domain Operations
Digital Filtering
Answers to Problems