| 研究実績の概要 |
In practical perspective, (1) the repetitive attempts pitch display module has been implemented; (2) the voice activity detection algorithm has been tested and adjusted to assure correct catch of the beginning of speech; (3) the components of StudyIntonation CAPT environment have been extended and assessed to support tonal language pronunciation training (Vietnamese) with a possibility to pitch graph segmentation for differentiation between phrasal intonation and specific tone mastering; (4) pilot versions of the beginner course for mora-timed language (Japanese) have been designed.
For theory, (1) longitudinal and microgenetic analysis of L2 pronunciation development has been conducted based on Vygotskian sociocultural theory concepts, thus, providing rationale for scaffolding learners through their zone of proximal development; (2) corpus comprising 1050 speech records labelled with orthographic transcript, pitch readings, and similarity metrics have been constructed for quantitative evaluation of learners' progress with using dynamical modeling and cross-recurrence quantification analysis; (3) accent recognition possibilities for further personalization of CAPT feedback have been studied.
|
