Computer Generation and Perception Evaluation of Music-Emotion Associations

Authors

Abstract

Music is intertwined with human emotions as an artistic form
with expressive qualities.We present a pilot study of music-emotion associations
based on a generative system, which produces parameter-based
music to represent four emotions: Happiness, Sadness, Calm, and Anger.
To study the perceptual relevance of each parameter, we performed a
series of user tests where participants explored multiple combinations
of musical parameters to reach a representation for each emotion. Results
were compared with the ones from previous studies and empirical
experiments proposed by other authors, which gave us a starting point
to evaluate each association and discover new possible connections. Although
most of the associations were conrmed, a few discrepancies were
found, such as the user preference for low pitch in Anger over the expected
high pitch. These ndings provide better insight and validation of
the relationship between music and emotions, and thus a starting point
to explore novel representations.

Keywords

algorithmic composition, generative music, music-emotion associations, emotion perception

Conference

14th International Symposium on Computer Music Multidisciplinary Research (CMMR), October 2019

PDF File