Abstract
This study aims at developing a new system that outputs a visual representation of prosody. It employs an automatic speech-recognition system and analyzes the segmental and prosodic features of a speech to provide an easy-to-understand visual feedback to Japanese learners. The problem of delayed F0 fall is corrected by assigning the F0 value to the mora that immediately precedes the segment in which the pitch drops. However, the inaccuracy in representing sentence-final intonation remains unsolved.
