Scale Transform in Rhythmic Similarity of Music

Abstract

PostprintThis is the accepted version of a paper published in IEEE Transactions on Audio, Speech, and Language Processing. This paper has been peer-reviewed but does not include the final publisher proof-corrections or journal pagination. Citation for the original published paper (version of record):Holzapfel, A., Stylianou, Y. (2011) Scale transform in rhythmic similarity of music.

An Intelligent Interface for Drum Pattern Variation and Comparative Evaluation of Algorithms

“…Apart from approaches that compute similarity between rhythmic patterns on a symbolic representation, there exist methods that consider other aspects of rhythmic similarity. Holzapfel and Stylianou [7] use the scale transform to develop a tempo invariant rhythmic similarity measure for music. Jensen et al [8] as well as Gruhne and Dittmar [5] use logarithmic autocorrelation functions calculated on different forms of onset density functions to obtain tempo invariant rhythmic features.…”

Adaptive nonlinear correlation with a binary mask invariant to rotation and scale

“…For a given signal, the frequency content via the Fourier transform can be determined; analogously, a scale transform is needed to indicate the scale content in the signal [3]. Applications of scale concept have been made in speech analysis [4], processing of biological signals [5], machine vibration analysis [6] and other areas [7,8]. The scale transform was also applied in multi-dimensional signal processing and it is used for image filtering and denoising [9].…”

In Search of Automatic Rhythm Analysis Methods for Turkish and Indian Art Music

“…Within these tempo classes, the full range of the tempo octave can be utilized by a composition (Holzapfel & Stylianou, 2011).…”

“…An important step to arrive at this understanding is the correct recognition of the length of the measure cycle in beats. In MIR, related tasks are referred to as time signature recognition (Pikrakis et al, 2004), meter tracking (Klapuri et al, 2006), and rhythmic similarity (Holzapfel & Stylianou, 2011). Meter tracking aims at tracking pulsation at several levels simultaneously, and the length of a measure cycle can be derived from the relation between beat and measure pulsation time-spans.…”