Towards an Educational Music Processor for Folk and Popular Musics

Abstract

This paper describes an educational musical processor that takes spectrographic data of a sonic object and turns them into a series of meaningful layers associated with different musical knowledge representations, in such a way that it can be understood, reproduced, played, compared, and taught by everyone across cultures, regardless of their musical backgrounds. Any music audio file can be used as input. Within the scope of this paper, the authors focus on processing of musical audio files onto common graphic platform of physical sound properties, in Hertz, Decibels, and milliseconds, so that culturally dependent musical units such as notes, beats, measures, phrases, chords, and sections can be viewed in separate layers. Syntactic techniques, such as frequency of occurrences, and adjacency are applied to musical units, such as pitches and musical chords. They are key pitches in context and key chords in context. The results are then mapped onto circles of fifths which reveal distinct patterns of each song, each section of one song, of each artist, each genre, and each culture. Semi-automatic generation of layers of annotations on top of the spectrogram helps teachers to quickly discover and/or compare distinctive features of a song, while preparing lessons. Learners of all levels can choose the most prominent patterns of the song to learn. This can also advance methods for preservation for further studies of sonic objects in the future.