We are working on digital tools for extracting semantic meaning from raw data, and for applying emotion to generative music. Some of the tools aim to find phrase boundaries, measure onset density and detect emotion by fuzzy logic. Another set of tools will generate melody based using probabilistic methods. The second stage will be expanding and expose these tools to emotion rendering where appropriate. The toolbox is being developed in max/msp/jitter and will eventually be available under the creative commons scheme.
This tool determines the boundaries of those score attributes that support emotion expression in music. When an emotion is requested in the two-dimensional emotion field (or the unnamed position between several emotions), the corresponding score attribute boundaries are displayed as a set of ranges.
The score attributes are divided into score features and performance instructions. The distinction into these groups is because musical emotion has two different sources, namely from the score and that are provided by the performer at the moment of performance.
The score features are tempo, note rate, melodic skipiness, pitch dissonance and duration contrast. Overall liveliness and formal intricacy are also included as score features, even though they are on a higher level and therefore more open to interpretation. Performance instructions are dynamics range and amounts of rubato, attack, articulation, phrasing and timbre.
Eight basic emotions are provided in the two-dimensional field, placed on its corners and sides. Any position between these points will provide the corresponding, interpolated attributes.
Research into emotion expression in music has been very active since Hevner’s first study in 1935. The findings have provided the framework for the present tool, and in particular the the work with the KTH rule system. It is particularly noteworthy that the tool is independent of musical style, and can be applied to the far larger part of western musical styles, as well as very many non-western. The emotional quailties in music are probably somewhat less important in contemporary academic music in the modernistic tradition, which tend to focus on aspects of musical material and its textural possibilities. However, music’s capacity to express emotion is probably its strongest feature, and the reason for its widespread and increasing use in a wide range of human activities.
Download Anders Friberg’s overview of digital audio emotions.