Project Description:


This project aims at the development of an "evidence-based" methodology to study emotive/affective and qualitative description of music. Establishing causal relationships between musical (audio-)stimuli and verbal/motoric responses of the listener forms a crucial point in our approach as this knowledge can lead to useful theories and practical applications. Formerly developed methods for cognitive information processing will now be used in the research on emotions, more specifically: (i)extraction of musical content from audio using computational models and neural nets , (ii) quantitative description of musical responses on audio stimuli, (iii) correlation techniques based on multivariate regression models. This methodology will moreover be extended with (i) motorial continuous annotation of perceived quantitative properties, and (ii) more advanced analysis od data.

Research objectives

The project aims at the development and realisation of:

  • a taxonomy of musical qualification for a broad range of cognitive and emotive, affective expressive descriptions
  • measuring devices for registration of musical quantitative responses and elaboration of new ways for musical annotation
  • correlation analysis of the annotations with stimulus, in order to investigate if quantitative descriptions of musical audio can be extracted directly from an analysis of the audio.
  • a methodology and epistemology for musical semantics with connatative sense, which agrees with the existing cognitive epistemology concerning music, and which forms the foundation of empirical aesthetics.


Promotors: Prof. Dr. M. Leman (IPEM : Dept. of Musicology - Ghent University)
Researchers: Liesbeth De Voogdt (200409-200509)
Micheline Lesaffre (200510-200708)
Financial Support:

Ghent University - Fund for Scientific Research (FWO) - Flanders (Belgium)