Subject: Signals & Systems for Speech & Hearing (further info) From: Stuart Rosen <stuart(at)PHONETICS.UCL.AC.UK> Date: Fri, 14 Jun 1996 17:57:52 +0100
Because a number of people have requested further information about this book, I have taken the liberty of sending it to the list as a whole. Signals and Systems for Speech and Hearing Stuart Rosen & Peter Howell, University College London, England An introductory textbook from Academic Press (1991) ISBN 0-12-597230-X ISBN 0-12-597231-8 pbk (I have been told that although the softcover version is no longer available, the hardback is available for the price of the softback.) Ordering information can be obtained from: http://www.apnet.com/textbook/sbe/ling.htm For those who are in the position of teaching basic aspects of hearing and speech science to students without much technical training (psychologists, audiologists, biologists, speech therapists, etc.), or who themselves need some brushing up on the basics, may I suggest they consider this book. In essence, it addresses the problem of presenting highly technical material to those who lack advanced technical training. This is done with minimal mathematics but over 300 figures (all drawn specifically for the book), and integrated closely with the text (no figure numbers are used). Exercises are provided at the end of most chapters. Further details follow. But please note: the mathematical/technical level is intentionally kept extremely low, so you will not learn how to calculate a Fourier transform from this book! But you should gain an intuitive understanding of what a Fourier transform does, and perhaps more importantly, why you would want to perform such a calculation in the first place. EXCERPTS FROM THE PREFACE. Many people working in the speech and hearing sciences come from non-technical backgrounds, making it difficult for them to master the essential technical underpinnings of the area. Perhaps the main barrier to acquiring knowledge in this area is that available textbooks nearly always assume an audience consisting of students of engineering. It is relatively difficult for the non-engineer, even with sufficient mathematical expertise, to get much out of these books. For readers without a knowledge of calculus, it is well nigh impossible. The problem of patchy understanding of basic technical concepts is acknowledged, at least implicitly, in most books dealing with speech and hearing. Nearly all begin with an introductory chapter purporting to cover the topics explored in this book - from relatively straightforward ones like the nature of sine waves and the calculation of decibels to the much more complex ideas of Fourier analysis and synthesis. Only those who already know this material are likely to appreciate its significance, and they do not need to read such a chapter. It seemed to us that only an entire book could cover in a sensible and thorough way the main concepts required. Furthermore, although we realized that there are some students who could handle a highly mathematical text, there are many who would have no mathematical training beyond algebra and trigonometry. Therefore, we decided to try to tell the entire story in words and pictures, with only minimal mathematics. We have attempted to provide the reader with a thorough introduction to the concepts of signals and systems analysis that play a role in the speech and hearing sciences. Few equations are used, and we have tried to maintain an informal, friendly and informative style throughout. Readers who like their technical material straight will, we hope, forgive us our little jokes. Because much of the story is told through figures, we have gone to great lengths to provide clear and truthful figures that show what the text says they do! We hope the reader will come away with a strong visual understanding of the concepts involved. This book can be used at many levels, from the student who hasn't heard of a spectrum before, to the experienced worker who has only a fuzzy understanding of the notion of an impulse response. We have tried to keep the underlying conceptual structure of signals and systems analysis explicit, in the hope that even some readers with advanced technical training might find clarification of the basic principles. ANNOTATED TABLE OF CONTENTS. Chapter I - Introduction - The relevance of studying the properties of signals and systems for work in the speech and hearing sciences; examples of signals and systems; input and output signals. Chapter II - Signals in the real world - the similarity of acoustic, mechanical and electrical signals; transduction. Chapter III - Introduction to signals - The specification and construction of sinusoids; the concepts of frequency, period, amplitude and phase; periodic and aperiodic sounds; measures of amplitude - peak- to-peak, rms; dB scales. Chapter IV - Introduction to systems - Linearity, additivity, homogeneity and time-invariance. Saturating nonlinearities. Application to middle ear and basilar membrane vibration, and to tape recorders. Chapter V - A preview - If we know what a linear time-invariant system does to sinusoids, we know what it will do to any signal. Chapter VI - The frequency response of systems - Amplitude responses as the ratio of output levels to input levels at particular frequencies. Low-pass, high-pass, band-pass and band-stop filters. The use of logarithmic axes (dB vs log Hz). The amplitude response of a cascade of systems. Application to middle ear vibrations and vocal tracts of varying shapes. Formants as resonances. Phase responses, linear and otherwise. Chapter VII - The frequency characterization of signals - Fourier analysis and synthesis. The amplitude and phase spectra of periodic signals: sinusoids, sawtooths, square waves, triangle waves and pulse trains. The spectra of aperiodic signals: transients and noise. Chapter VIII - Signals through systems - Determining the output of systems to specified signals in the frequency domain, for both amplitude and phase. A sawtooth through ideal and realistic low-pass filters. Noise through a filter. Distortion. Chapter IX - The time characterization of systems - The notion of an impulse, and the impulse response. The relationship between the frequency response of a system and its impulse response. Determining the frequency response of a set of headphones in 3 ways: via the impulse response, a sinusoid swept in frequency, and white noise. Chapter X - The relationship between the time and frequency domains - Signals short in time tend to be wide in spectrum, and vice versa. The trade-off between temporal resolution and frequency resolution in band- pass filters. The relationship between a system's impulse response and its frequency response. Chapter XI - The sound spectrograph: Practical short-term spectral analysis - Determining dynamic spectral changes; the notion of a filter bank; rectification and smoothing; short-term spectra; the use of wide- and narrow-band filters; details of practical spectrographs. Chapter XII - Applications to hearing - The measurement of frequency responses, and notions of linearity and nonlinearity in the peripheral auditory system: head and pinna, the ear canal resonance, middle ear vibration, basilar membrane motion; the auditory system as a filter bank; measuring the auditory filter behaviorally. Chapter XIII - Applications to speech production - The source-filter theory of speech production and its application to vowels, diphthongs and fricatives. Chapter XIV - Digital signals and systems - The differences between analog and digital systems; sampling and quantization; the sampling theorem; digital-to-analog and analog-to-digital conversion; aliasing.