O’Shaughnessy D. Speech Communications. Human and Machine

IEEE Press, 2000. — 560 p.Speech commW1ication is an interdisciplinary subject. Although much of the research material for the book comes from engineering literature (e.g., IEEE journals), a wide variety of sources is employed (especially for Chapters 3-5). The book is directed primarily at an engineering audience le.g., to a final-year undergraduate or graduate course in electrical engineering or to those in speech research), but it should also be accessible to linguists, phoneticians, psychologists, audiologists, computer scientists, and systems engineers.
Linguists view speech in tenns of language description; they characterize languages via phonemes and intonation and note differences and similarities in how different languages convey messages. Phoneticians examine the relationships between phonemes, their articulation, and their acoustic properties. Psychologists deal in perception, noting the relationships between acoustic properties of signals (e.g., speech) and what people hear and understand. Audiologists deal with disorders of the hearing system and often use sounds other than speech in their work. Speech communication is treated as a programming problem by computer scientists; using artificial intel1igence techniques, they seek to simulate the human actions of producing and understanding speech signals or to represent speech in an efficient fashion for transmission. Systems engineers have similar objectives and use similar tools as the computer scientists, but tend to be more concerned with questions of efficiency and practicality, whereas scientists seek more to understand the communication process. By simultaneously including both technical detail and basic explanations about speech communication, this book should address the concerns of all researchers involved in speech, as well as interest the student wishing an introduction to the field.
Since the book covers a wide field, it should be useful as a reference text. In it, readers will find an introduction to virtually any subject of relevance to speech communication. By examining its up-to-date reference list, readers can quickly locate references for further details. To facilitate such reference searches, the Hst contains major information sources that should be available in most scientific or engineering libraries, such as textbooks, journals, and a few conference proceedings. Relatively inaccessible sources, such as technical reports and references with abstracts only, have been avoided where possible.
It is always difficult to write a book for an audience to include both students (from both engineering and other disciplines) and professionals. I have attempted to satisfy both g-roups with this book, but of necessity I have had to compromise. For students, the book is highly tutorial, does not presume any prior knowledge about speech communication, presents topics in a sequence that is relatively easy to follow, and provides problems at the end of most chapters. For researchers, the coverage of the book is complete in virtually all aspects of speech communication, the list of references is large and up-to-date [the most important references, and those of tutorial nature. are noted by an asterisk (*)], and new developments are put into the perspective of older, classical results. As a result, students may find the text at times too detailed, while professionals m3) wish for more detail in areas that interest them. I believe, however, that I have struck a good compromise: researchers, by using the reference list, should be able to locate further detail quickly, while one-semester students may skip over sections noted with a double dagger (‡).
Finally, this second edition has more details on major developments in speech processing between 1987 and 1999, for example. among others, on CELP, HMMs, and neural networks. A related major development of the 1990s has been the massive interest in the Internet (World Wide Web). In response, I have included a large list of relevant Web sites for speech information. I wi1l update this information periodically at the Web site for this book, http://www. inrs-telecom.uquebec.ca/users/ spchwww/EngIish/persons/dougolbook.htmlReview of Mathematics for Speech Processing
Speech Production and Acoustic Phonetics
Hearing
Speech Analysis
Coding of Speech Signals
Speech Enhancement
Speech Synthesis
Automatic Speech Recognition
Speaker Recognition