Using Quantized Melody Contour to Discover Significant Repeating Figures in Classic Music

Abstract

To develop a content-based music database, it is very important to extract music features from the raw data of music object and organize them as the music index for music retrieval. A repeating pattern is a series of notes which appears more than once in a music object. The musicologists believe repetition is a universal characteristic in music forms. Most of the repeating patterns are themes or easy to remember tone for people. Hence, the themes and other repeating patterns are important music features that can be used for both music data analysis and content-based music retrieval. In this paper, we use the idea of interval between two adjacent notes to form quantized melody contour for representing music objects. This representation differs from existing method that uses note to form a melody string. A figure is defined as a melody contour of a musical segment. A sequence pattern is a melody segment that has the same figure with other melody segment. From musicologist’s view, figures are more meaningful than repeating patterns. Hence, we also propose a method to find significant repeating figures. These figures could cover most of the repeating patterns got by noted melody string. In addition, the number of repeating figures is less than the number of repeating patterns. As a result, the searching cost including execution time and memory space can be dramatically decreased.