English sentence pronunciation evaluation using rhythm and intonation

Rhythm and intonation are important factors in the English sentence pronunciation evaluation. In this paper, the Mel Frequency Cepstrum Coefficient (MFCC) feature and Hidden Markov Model (HMM) algorithm are used to establish a model for speech recognition. Then it makes an evaluation of English sentence pronunciation focusing on rhythm and intonation, and gives feedbacks and recommendations about pronunciation problems to the users based on the expert knowledge. Verified by experiments, the model has a certain precision and reliability. It has applied to a pronunciation evaluation system which can improve users´ pronunciation.