Pronunciation evaluation technology based on computer aided Chinese learning system
Hu Ju · Computer Engineering and Applications Journal · 2014
A method for automatic pronunciation evaluation is proposed, applying to the learners with Mandarin as a second language learning Chinese. Bilingual language speech database is introduced and the evaluation is divided into two parts, which are the manner of articulation and pronunciation. This paper extracts the parameters of each part and makes use of the improved method based on comparison of speech similarity, an algorithm named Bilingual Dynamic Time Warping. It gets the final result by means of score mechanism. Experiments verificate the validity of the system.