Computing Reviews
Today's Issue Hot Topics Search Browse Recommended My Account Log In
Review Help
Search
Speech input and output assessment: multilingual methods and standards
Fourcin A. (ed), Harland G. (ed), Barry W. (ed), Hazan V. (ed), Halsted Press, New York, NY, 1989. Type: Book (9789780470214398)
Date Reviewed: Feb 1 1990

This book describes an interesting topic: how to assess speech synthesis and speech recognition systems, that is, what components to look at and what methodologies to use in assessing their performance. The text is largely unreadable and of little interest.

First, the typography is perfectly awful. Much of the text is spread out over the 41-2--inch lines with only a few words per line. The result is a large amount of empty space on each line. A typical six-word line consists of 35-16- inches of text and 13-16- inches of space between words, with the spacing haphazardly chosen as 6-16-, 5-16-, 2-16-, 3-16-, and 3-16- inches between pairs of words. A normal human being used to conventional typography will find such a presentation most disturbing and will give up in a hurry.

This book represents the initial report of “Esprit Project 1541,” entitled “Speech Assessment Methodologies.” (Esprit is a research program initiated some years ago by the European Economic Community (EEC) to encourage international cooperation between industrial and academic research groups in various countries of the EEC.) Several research projects have been supported under the Esprit umbrella, including this speech assessment work. Each chapter of this book is a report by one of the cooperating organizations and deals with different aspects of speech assessment, including speech recognition, speech synthesis, existing speech databases, phonetic transcription methods, database requirements, and speech workstation technology. The reports were apparently prepared early in 1988.

Even allowing for the fact that the text does not deal with the guts of the problem--the actual state of the art in speech work--but only with some relatively arcane questions of the assessment of speech handling systems, it is disturbing to find that it contains little that is new or interesting. The reader would do much better to read a shorter, more concentrated, and properly focused piece written by a single expert (see, for example, Klatt [1]) instead of this lengthy catalog of systems, components, and possibilities.

A few specialists in speech performance assessment might find something of interest here. The rest of us should, and no doubt will, stay away.

Reviewer:  Gerard Salton Review #: CR113917
1) Klatt, D. H.Review of text-to-speech conversion for English. J. Acoustical Soc. Am. 82 (1987), 737–793.
Bookmark and Share
 
Speech Recognition And Synthesis (I.2.7 ... )
 
 
Text Analysis (I.2.7 ... )
 
 
User/ Machine Systems (H.1.2 )
 
Would you recommend this review?
yes
no
Other reviews under "Speech Recognition And Synthesis": Date
On-line recognition of spoken words from a large vocabulary
Kohonen T. (ed), Riittinen H., Reuhkala E., Haltsonen S. Information Sciences 33(1-2): 3-30, 1984. Type: Article
Oct 1 1985
Connected spoken word recognition algorithms by constant time delay DP, O (n) DP and augmented continuous DP matching
Nakagawa S. Information Sciences 33(1-2): 63-85, 1984. Type: Article
Jun 1 1985
The phonetic basis for computer speech processing
Ladefoged P., Prentice Hall International (UK) Ltd., Hertfordshire, UK, 1985. Type: Book (9789780131638419)
Dec 1 1987
more...

E-Mail This Printer-Friendly
Send Your Comments
Contact Us
Reproduction in whole or in part without permission is prohibited.   Copyright 1999-2024 ThinkLoud®
Terms of Use
| Privacy Policy