Music preferences based on audio features, and its relation to personality

Abstract

Recent studies have summarized reported music preferences by genre into four broadly defined categories, which relate to various personality characteristics. Other research has indicated that genre classification is ambiguous and inconsistent. This ambiguity suggests that research relating personality to music preferences based on genre could benefit from a more objective definition of music. This problem is addressed by investigating how music preferences linked to objective audio features relate to personality. Participants (N = 354; 165 males) completed a personality measure and rated their preference to 120 music clips from various music (sub-)genres (e.g., classical, rock). Principle Components Analysis revealed a nine-component model that accounted for 61% of the variance in music preference ratings. Audio features computationally derived from the music clips were subsequently analysed to discriminate between music contained within each of these factors. In addition, participants’ estimated music preference scores to these factors were related to personality facets. Aggregated results showed, for example, that Excitement-Seeking was positively related to music with a greater number of percussive events while negatively related to music with fewer percussive events. Results are discussed in terms of how objective features can provide greater insight into how music preferences relate to personality.