A new approach to pronunciation assessment of English sentences

The extent of pronunciation and intonation is the key factor in assessing the pronunciation quality of English sentences, and the curve of the speech intensity reflects the variation of intonation. Therefore, this curve can be used as a character in the speech quality evaluation. A new standard-based pitch contour algorithm is proposed in this paper. Since people's pronunciation differs in duration and intensity, this algorithm builds a DTW(Dynamic Time Warping) model by aligning the evaluated sentences with the standard sentences in duration and intensity, then outputs the distance between the standard sentences and the evaluated sentences as fractions. To evaluate the performance of this method, we used a test set of recordings of 5 sentences read by 10 speakers. Experiment showed that the result of this approach is quite objective.