1.Understanding Music through Modern Technology

    a)  Recognizing Music

An important aspect of Music Information Retrieval (MIR) is to make computers distinguish music from other types of ‘noise’, thus aiding the listener in tagging and thereby being able to remember and re-find a sound. For instance, computer programs are developing that, referring to a large database of songs or works, can recognize the structures of a given piece of music in other sonic versions. This raises interesting questions as to how music is experienced and remembered, both on the individual and on the cultural level. (See also Grund (2005) “Music Information Retrieval, Memory and Culture: Some Philosophical Remarks.”) The answering of these questions will become more and more daunting as the years go by and virtually all of the world’s music becomes stored – and thus retrievable – in gargantuan databases.

These questions are closely connected with the project of understanding the role music has played in human evolution. The claim that music is biologically important in our species and the formulation of evolutionary hypotheses for the origin of music depends not only on localization of brain centers dedicated to music processing, but also on demonstration of universal features that distinguish musical sounds in contrast to other sounds. If such features cannot be demonstrated it is very difficult to claim that music biologically is different from other acoustic communication and has had its own evolutionary rationale during human history. Presently, there is no consensus whether there are brain centers dedicated to music processing, since the brain regions crucial for music processing are also involved in speech processing (interesting research has been done, showing the importance of music in infant learning of language – see Trevarthen (2002)).

One project will try to investigate whether there are any universal characteristics distinguishing musical from non-musical sounds. In Western culture, all sounds can be used in a musical context, but for some of the sound types exactly context may be all-important in order to classify them as music. Therefore, the project will aim to expose diverse collection of test subjects to a large matrix of brief sounds, asking for an immediate classification of the sounds as ‘musical’ or ‘non-musical’. For a subset of the participants, the test will be performed while recording brain activity using PET scan or magnetoencephalography to correlate the behavioural response with response in the relevant brain centers. The SDU researchers do not have facilities for these kinds of measurements, but will collaborate either with the Dr. Albert Gjedde, PET center at Aarhus University or with Dr. David Poeppel, Department of Linguistics, University of Maryland for magnetoencephalography.

(Jakob Christensen-Dalsgaard, Cynthia M. Grund)

(Click here to return to Projects page.)