How to Build a Speech Recognition Application

Grounded in hours of human-computer experiments, and a multi-disciplinary approach to user interface design - this book is a rare combination of a careful ear for human language and dialogue, extensive engineering experience, and pragmatic knowledge of the strengths and limitations of current voice recognition technology.

The second edition has brought it bang up-to-date. It cuts through the hype that has always surrounded each successive generation of voice technology - focussing always on the building of robust useable interfaces which work with the user rather than against them.


Thoughts on the second edition   March 20, 2002
 5 out of 6 found this review helpful

This book is so well organized and articulated it's bound to be of value to anyone doing ASR application development. My own experience in voice response runs from end-user representative to application designer and I found every chapter enlightening. In this second edition of Bruce and David's ASR style guide, I believe the end users will find the new sections on voice portals and managing your voice talent of particular interest. And all users should take particular note of the expanded discussions of usability testing and performance reporting.

I found the first version of How to Build a Speech Recognition Application so useful that I actually took the time to compared the new edition, page for page, with the original. That was a relatively easy task, because the authors retained the original section numbering wherever possible. My comparison showed that the original guidelines have been substantially updated, based on continuing research and the hands-on experiences of both the authors and other acknowledged experts. In addition, I believe the new sections and expanded discussions of critical design considerations are going to prove valuable to both novice and seasoned developers.

In short, developing effective telephony dialogues is a complex, rapidly evolving and downright expensive task. Given that reality, every development team ought to have at least one copy of this landmark style guide.


No longer the only book on the block.   September 2, 2001
 1 out of 3 found this review helpful

When this book came out a couple of years ago it was the first and only book on designing speech recognition systems. It was very valuable then, but now more books are available that cover the same information and more, for a lot less money.


A must-have book for speech application developers   April 8, 2001
Developing speech applications is not easy to master. Even with VoiceXML becoming more widely adopted, there are a lot of intricacies that that a developer must understand. This book will provide you with a solid foundation to become an effective speech application developer.

The book did very well in presenting the limitations of the current speech recognition technology (dialog design, large vocabularies, promtp design, etc.) and made suggestions on how to overcome such problems in specific situations.


The "Strunk and White" for Speech Recognition   July 28, 1999
 5 out of 6 found this review helpful

Author, Bruce Balentine's goal with How to Build a Speech Recognition Application is to produce the Strunk and White of speech recognition. An electronic musician and composer, Balentine was a pioneer in the speech recognition field. The text's examination of the problems of navigation from human perception to machine recognition give comprehension to even the layman. The book is well organized and structured with the +,/,- system which allows a novice to follow. According to linguist, Dr. John White, the first chapter could stand alone as a treatise on the dynamics of the human speech interface with the machine.