next up previous contents
Next: File Format Up: Speech Corpus Production Previous: Check List Annotation   Contents

Pronunciation Dictionary

The pronunciation dictionary lists the most likely pronunciation or citation form of all words that are contained in the speech corpus. A pronunciation dictionary cannot be considered as a `must' but the usability of the corpus - especially for ASR - will increase considerable if you provide this information together with the corpus. As in other parts of this cook book the choice of how to encode the pronunciation of your corpus can range from very simple and achievable with automatic procedures to very complex and time-consuming. In any case be prepared to reserve some man power for the creation of a dictionary, because in most cases there is manual work involved.


BITS Projekt-Account 2004-06-01