Media Conversion between Music and Color Combination considering time-series changing of musical impression

Abstract

We propose a new method to realize a media conversion between music and color combination based on impression. In this research, we focus on time-series changing of musical impression.In order to determine time-series changing of musical impressions, an inputted music is divided into some sub-sequences according to change positions of impression. In the subjective assessment, the impression for each sub-sequence is evaluated, and we define an algorithm to derive the most appropriate color combination to the impression of sub-sequence based on "Affective mapping table between Music and Color Combination" which has been obtained from our previous research.From the experimental results using five musical sources, the average of concordance rate between music and the obtained color combinations using our algorithm is about 82%. In this research, it is clarified that the accuracy of media conversion between music and color combination is improved by our algorithm which consider the time-series changing of musical impression.