Generating speech database for target vocabulary involves concatenating training text segments with target vocabulary words according to phonetic description

摘要

The method involves converting the words of a target vocabulary into a phonetic description so that individual words are represented by phonemes, and concatenating segments of a spoken training text independent of the target vocabulary and consisting of one or more phonemes with words of the target vocabulary according to the phonetic description. Independent claims are also included for the following: an arrangement for generating a speech database for a target vocabulary for training a speech recognition system.