Culture specific approaches in music content analysis

Abstract

The extraction of culturally meaningful features from audio recordings of music different from the western repertoires that are the most studied requires new signal analysis techniques and machine learning methods. The fact that many music traditions have fundamental differences from western ones, such as different musical instruments, tuning systems, performance styles, or musical forms, imply that at the level of feature analysis, most of the descriptors and extraction methodologies being used to analyze western music are not appropriate, or at least they have to be developed much further. Culture specific issues have profound research implications, offering new research problems and requiring new approaches. In this article we will present some initial results in the audio content analysis of the classical music traditions of India (Hindustani and Carnatic) and Turkey (Ottoman) especially for the issues of melodic and rhythmic description. This research is being carried out within a large project entitled CompMusic that aims to promote and develop multicultural perspectives in music computing research. In this project we want work on culture specific music problems with the goal to find new computational methodologies of interest for a wide variety of music information processing problems.