Mode and Register Categorizations for Predicting Mood in Classical Piano Music2.

Le résumé fourni par la source

Previous work has shown that differentiating positive and negative mood in low-arousal music is difficult, and probably the biggest current bottleneck in automatic music emotion recognition. Researchers generally agree that the mode of the music (i.e., whether Major or Minor) is the primary factor in determining whether the mood is positive or negative. This paper considers the accuracy of Major and Minor as mood predictors for a selection of 300 low-arousal classical piano excerpts. Our results indicate that Major-Minor was 90% accurate in predicting positive and negative mood on these excerpts. For the residual 10% fail cases (Major but negative, or Minor but positive), we considered more detailed categories of mode and register, and found that Natural Minor excerpts accounted for more than half of the fail cases. In addition, among the specific modes, there were significant differences between Pure Major, Mostly Major, Natural Minor, Harmonic Minor, and Highly Chromatic (in descending order of valence). These results indicate the strong predictive power of Major and Minor for differentiating positive and negative mood, and give strong motivation for including mode features in music emotion recognition systems.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.