An approach to synthesis of a phonetically representative english text of minimal length

摘要

To assess foreign language pronunciation one must listen to a speaker, which is a very time consuming task. To reduce the efforts it is proposed to construct a text of minimal length preserving English corpus natural phoneme distribution. Such a text being uttered will give a comprehensive evaluation of English pronunciation. The approach uses International Phonetic Alphabet (IPA) for the internal representation of English texts. The purpose of the work is the development of software for the synthesis of phonetically representative text for the English language. The paper deals with the English language constructs. A system for the synthesis of phonetically representative text for the English language has been developed. The Kolmogorov-Smirnov test shows that the obtained short texts have the phoneme/di-phone/tri-phone frequencies close to their frequencies in initial corpora.