Public systems with speech input have been more and more introduced in the last few years. By these systems a new form of information collection and information dissemination will be provided for the general public. People are up to now well acquainted with public announcement systems, which do not have a special information input or where they have to give their input information via touch tone telephones. At least the speech output side is often rather common to the unexperienced user. People use many such announcement services, like entertainment, time announcement, weather forecast etc. In most cases of such systems there is no speech input. To introduce now the more advanced and more general speech input is unusual for most users and therefore they have their first problems how to speak and interact with such a speech driven system. They expect a speech recognizer which is perfectly working and often they suppose the recognizer would be able to understand continuous speech. For the near future we will have two sides of the problem, how to guarantee a minimum recognition capability necessary for establishing a certain service and on the other side how to prepare the occasional user for such a system.
Cite as: Mangold, H. (1989) Assessment of speech recognizers in public information and ordering systems. Proc. Speech Input/Output Assessment and Speech Databases, Vol.1, 37-58
@inproceedings{mangold89_sioa,
author={Helmut Mangold},
title={{Assessment of speech recognizers in public information and ordering systems}},
year=1989,
booktitle={Proc. Speech Input/Output Assessment and Speech Databases},
pages={Vol.1, 37-58}
}