IJMSTA - Vol. 2 - Issue 1 - January 2020
Analyzing relationships between color, emotion and music using Bayes' rule in Bach's Well-Tempered Clavier Book 1
Authors: Renee Timmers
Abstract - A probabilistic approach to the perception of emotion and color in music is proposed and the application of Bayes' rule to predict previously collected data is investigated. Specifically, performances of Bach's Well-Tempered Clavier Book I were analyzed in terms of mode, tempo and intensity. Estimates of probabilistic relationships between features and emotion dimensions were used to predict listeners' associations with the music in terms of emotion and color. Predictions were particularly successful for emotion perception, although color was also reliably predicted for 14 out of 24 Preludes. If color was predicted directly from emotion perception, reliable prediction increased to 18 out of 24 Preludes.
Keywords: Bayesian probability, Emotion, Music perception, Synesthesia