Abstract:
Aiming at objective evaluation of L2 (second language) learners' proficiency, we statistically analyzed mispronunciation characteristics of Japanese learners' English speech based on phonetic knowledge. Through automatic forced alignment using a speech recognition tool kit, phone deletion, substitution and insertion were detected by substituting the pronunciation dictionary reflecting learners’ phonetic error characteristics. For phonetic knowledge, typical mispronunciations, vowel insertion, vowel change and consonant omission and change were taken into account. Mispronunciation analysis using 50 Japanese speakers showed consistent error characteristics along their subjective naturalness scores. The correlation sore 0.595 was obtained between subjective evaluation scores and objective measures by a multiple linear regression analysis, using the mispronunciation statistics as parameters.
