Quantification and Display of Emotions in Music

In this project, we quantify emotion in music, and portray our findings in a visualization system. The work can be divided into three main stages: processing music files in MIDI format to extract the desired musical properties, mapping these properties to emotions, and producing an appropriate and effective visualization of these emotional quantities. The MIDI processing stage provides continuous data on musical content by performing file segmentation followed by property extraction. A mathematical model we develop then maps the extracted properties to quantitative emotional values. The use of vectors and matrices in the structure of the model uniquely suits the problem’s demand for a mapping between lists of quantitative data. The numerical values used in the model are derived directly from cumulative documented results of experimental studies. The final visualization presents the numerical results in a Processing sketch, where emotions are encoded with color, and fluid dynamics infuse it with dynamism.