A CONTEXT-BASED EMOTION-ANALYZER FOR TEACHING TONALITY IN MUSIC COURSES

摘要

This paper presents a context-based emotion-analyzerrndedicated for supporting to learn tonality in music courses.rnThis emotion-analyzer realizes a new music retrievalrnenvironment to find and visualize music items withrnconsidering genre dependent perceptual preference ofrnmusic. This system generates emotive annotations forrnmusic by analyzing tonality along with a timeline. Thernsystem helps users to identify music tonality from thernviewpoint of emotions. Tonality is a musical system thatrnis constructed by sound elements, such as harmonies andrnmelodies. Change of tonality causes change of impression.rnThis system realizes an automatic time-duration selectorrnthat detects repetitions and bridges by analyzing physicalrnand structural music features, such as pitch and tonality.rnThis system enables users to submit emotive keywords asrna query for retrieving music according to the impressivernchanges in each musical piece. This paper shows arnprototype system that searches MIDI music files byrnanalyzing them automatically. This paper also showsrnseveral experimental results for clarifying the feasibilityrnof the system.