How potential users of music search and retrieval systems describe the semantic quality of music

Abstract

A large-scale study was set up aiming at the clarification of the influence of demographic and musical background on the semantic description of music. Our model for rating high-level music qualities distinguishes between affective/ emotive, structural and kinaesthetic descriptors. The focus was on the understanding of the most important attributes of music in view of the development of efficient search and retrieval systems. We emphasized who the users of such systems are and how they describe their favorite music. Particular interest went to inter-subjective similarities among listeners. The results from our study suggest that gender, age, musical expertise, active musicianship, broadness of taste and familiarity with the music have an influence on the semantic description of music.