Interactive Speech Technology: Human Factors Issues In The Application Of Speech Input/Output To Computers: Human Factors Issues In The Application Of Speech Input/Output To Computers PDF

Foreword

This text deals with two important technologies in human-computer interaction: computer generation of synthetic speech and computer recognition of human speech. These technologies are quite different and the ergonomics problems in implementation are also different. Nonetheless, synthetic speech and speech recognition are usually dealt with in the same context of speech technology.

Speech technology provides what many think of as the most 'natural' and most efficient mode of human-computer interaction: talk to a computer and get a spoken response back. Indeed, Chapanis and his co-workers confirmed 15 years ago that speech is more efficient than other modes of interaction (Chapanis et at., 1977). However, speech is still difficult to use with computers. In particular, computer speech recognition has not yet been perfected, and to cope with imperfections, users may have to perform unnatural tasks, such as leaving long pauses between spoken words. The question then arises if maybe 'natural' keying is better than 'unnatural' speaking. The other problem has been the limitation in number of recognized words that a computer can understand (its library, or lexicon, of words). This limitation may be less severe than originally thought, since it has been demonstrated that humans can effectively solve problems using a limited vocabulary of about 200 words. In the past few years speech technology has rapidly improved, and we may now be on the verge of solving some of the major technical problems. Such development would have significant economic implications since many viable applications could be implemented. Below, we will briefly review some of the issues addressed in this book.

Authors: Chris Baber, J Noyes