Introducing Speech and Language Processing

by John Coleman

• ISBN13: 9780521823654
• ISBN10: 052182365X
• Edition: CD
• Format: Hardcover
• Copyright: 2005-04-11
• Publisher: Cambridge University Press
• More Book Details

Summary

This major new textbook provides a clearly-written, concise and accessible introduction to speech and language processing. Assuming knowledge of only the very basics of linguistics and written specifically for students with no technical background, it is the perfect starting point for anyone beginning to study the discipline. Students are introduced to topics such as digital signal processing, speech analysis and synthesis, finite-state machines, automatic speech recognition, parsing and probabilistic grammars, and are shown from a very elementary level how to work with two programming languages, C and Prolog. The accompanying CD-ROM contains all the software described in the book, along with a C compiler, Prolog interpreter and sound file editor, thus providing a self-contained, one-stop resource for the learner. Setting a firm grounding in speech and language processing and an invaluable foundation for further study, Introducing Speech and Language Processing is set to become the leading introduction to the field.

Table of Contents

Acknowledgments and copyright notices	x
CD-Rom and companion website	x
Introduction	1	(12)
About this book	2	(1)
Purpose of this book	2	(1)
Some reasons to use this book	3	(2)
What's in the book (and what's not)	5	(3)
Computational set-up needed for this book	8	(1)
Computational skills that are necessary in order to use the book	9	(1)
Free software suggestions	10	(1)
Book structure	10	(3)
Sounds and numbers	13	(34)
Preparatory assignments	14	(7)
Solutions	21	(3)
Sampling	24	(1)
Quantization	24	(3)
The sampling theorem	27	(2)
Generating a signal	29	(2)
Numeric data types	31	(3)
The program	34	(1)
Structure of a loop	35	(2)
Structure of an array	37	(1)
Calculating the cosine values	38	(1)
Structure of the program	39	(1)
Writing the signal to a file	40	(7)
Chapter summary	43	(1)
Further Exercises	43	(3)
Further reading	46	(1)
Digital filters and resonators	47	(24)
Operations on sequences of numbers	48	(1)
A program for calculating RMS amplitude	48	(2)
Filtering	50	(2)
A program for calculating running means of 4	52	(2)
Smoothing over a longer time-window	54	(1)
Avoiding the need for long windows	54	(7)
IIR filters in C	61	(1)
Structure of the Klatt formant synthesizer	62	(9)
Chapter summary	68	(1)
Exercises	68	(1)
Further reading	69	(2)
Frequency analysis and linear predictive coding	71	(40)
Spectral analysis	72	(1)
Spectral analysis in C	72	(7)
Cepstral analysis	79	(1)
Computation of the cepstrum in C	80	(3)
Pitch tracking using cepstral analysis	83	(3)
Voicing detection	86	(4)
fo estimation by the autocorrelation method	90	(5)
Linear predictive coding	95	(5)
C programs for LPC analysis and resynthesis	100	(6)
Trying it out	106	(1)
Applications of LPC	106	(5)
Chapter Summary	109	(1)
Further exercises	109	(1)
Further reading	110	(1)
Finite-state machines	111	(46)
Some simple examples	112	(1)
A more serious example	113	(3)
Deterministic and non-deterministic automata	116	(2)
Implementation in Prolog	118	(11)
Prolog's processing strategy and the treatment of variables	129	(3)
Generating strings	132	(2)
Three possibly useful applications of that idea	134	(1)
Another approach to describing finite-state machines	135	(2)
Self-loops	137	(2)
Finite-state transducers (FSTs)	139	(5)
Using finite-state transducers to relate speech to phonemes	144	(5)
Finite-state phonology	149	(4)
Finite-state syntactic processing	153	(4)
Chapter summary	156	(1)
Further exercises	156	(1)
Further reading	156	(1)
Introduction to speech recognition techniques	157	(28)
Architectures for speech recognition	158	(8)
The pattern-recognition approach	166	(2)
Dynamic time warping	168	(9)
Applications	177	(4)
Sources of variability in speech	181	(4)
Chapter summary	182	(1)
Further reading	183	(2)
Probabilistic finite-state models	185	(36)
Introduction	186	(1)
Indeterminacy: n-gram models for part-of-speech tagging	187	(3)
Some probability theory for language modelling	190	(2)
Markov models	192	(6)
Trigram models	198	(4)
Incompleteness of the training corpus	202	(7)
Part-of-speech model calculations	209	(1)
Using HMMs for speech recognition	210	(3)
Chomsky's objections to Markov models and some rejoinders	213	(8)
Chapter summary	219	(1)
Further reading	219	(2)
Parsing	221	(30)
Introduction	222	(1)
A demo	222	(1)
`Intuitive' parsing	223	(2)
Recursive descent parsing	225	(7)
The simplest parsing program	232	(1)
Difference lists	233	(3)
Generating a parse tree	236	(2)
Syllabification	238	(4)
Other parsing algorithms	242	(1)
Chart parsing	242	(3)
Depth-first vs. breadth-first search	245	(1)
Deterministic parsing, Marcus parsing and minimal commitment parsing	246	(3)
Parallel parsing	249	(2)
Chapter summary	249	(1)
Further reading	250	(1)
Using probabilistic grammars	251	(24)
Motivations	252	(4)
Probabilistic context-free grammars	256	(2)
Estimation of rule probabilities	258	(3)
A practical example	261	(6)
A limitation of probabilistic context-free grammars	267	(1)
Tree adjoining grammars	268	(3)
Data-oriented parsing	271	(4)
Chapter Summary	272	(1)
Conclusion and suggestions for further reading	272	(3)
Appendix: The American Standard Code for Information Interchange (ASCII)	275	(2)
Glossary	277	(16)
References	293	(6)
Index	299