Speech Transcription
Automated Speech Transcription is the process of automatically providing a textual representation from audio input (Chinchor 2000).
Many terms are used to refer to this technology, including both speech recognition
and voice recognition. This is the "oldest" human language technology that has
been evaluated by DARPA in order
to encourage advancement of the technology, with evaluations first begun in the 1970's
(Chinchor 2000). Recent evaluations include the Hub-4 Broadcast News evaluations and the
Hub-5 Large Vocabulary Conversational Speech Recognition evaluations.
Transcripts made by these systems
often tend to include certain types of errors. You may notice these to some extent even
in the closed captioning displayed on some televisions, even when these transcriptions
are created by humans. The most common errors include
insertion, deletion, or substitution
httpd://www.oocities.org/ejb_wd/Speech.html
© 2000-2006, Erica Jean Lindsey Brown, All rights reserved
This page has been accessed - 
- times since November 14, 2000