Multi-scale Modelling of Segmentation

Abstract

While listening to music, people often unwittingly break down musical pieces into constituent chunks such as verses and choruses. Music segmentation studies have suggested that some consensus regarding boundary perception exists, despite individual differences. However, neither the effects of experimental task (i.e., real-time vs. annotated segmentation), nor of musicianship on boundary perception are clear. Our study assesses musicianship effects and differences between segmentation tasks. We conducted a real-time experiment to collect segmentations by musicians and nonmusicians from nine musical pieces. In a second experiment on non-real-time segmentation, musicians indicated boundaries and their strength for six examples. Kernel density estimation was used to develop multi-scale segmentation models. Contrary to previous research, no relationship was found between boundary strength and boundary indication density, although this might be contingent on stimuli and other factors. In line with other studies, no musicianship effects were found: our results showed high agreement between groups and similar inter-subject correlations. Also consistent with previous work, time scales between one and two seconds were optimal for combining boundary indications. In addition, we found effects of task on number of indications, and a time lag between tasks dependent on beat length. Also, the optimal time scale for combining responses increased when the pulse clarity or event density decreased. Implications for future segmentation studies are raised concerning the selection of time scales for modelling boundary density, and time alignment between models.

Repetition and Aesthetic Judgment in Post-tonal Music for Large Ensemble and Orchestra

“…The analysis did not reveal any differences in response behavior as a function of musical training with regards to the perception of repetitions, neither as measured by the repetition strength factor, nor for the time series based on the synchronous responses. Note that a comparable independence of musical training and response behavior has been observed in the segmentation literature (Hartmann et al, 2016;Popescu et al, 2021). With regards to Factor 1 (aesthetic preference), however, the regression model indicated a (comparatively weak) interaction effect of excerpt category and musical training, which was based on an association between preference scores for modernist excerpts and the level of musical training of participants, demonstrating a slight rise in preference for post-tonal music by listeners with previous musical training.…”

The definition of a musician in music psychology: A literature review and the six-year rule

“… a Exclusive use of “musician” term only; includes vocal and non-vocal performers. One study reported a minimum of six years of music training, but this has been corrected to four years following personal communication with the corresponding author (Hartmann, Lartillot, & Toiviainen, 2016). b Any mention of the word “trained”; includes Makam-trained (which is an Arabic music system), and trained instrumentalist/singers.…”

Western listeners detect boundary hierarchy in Indian music: a segmentation study

“…Musical stimuli include both improvised performances 18,19 , and score-based, synthesised realisations of composed items [20][21][22][23] . Although segmentation studies based on Western music frequently compare the performance of musically trained and untrained listeners 1,10,23,[26][27][28] , among cross-cultural studies only Mungan et al 23 compare (Turkish) musicians and non-musicians, finding higher convergence of the musicians with the expert listeners. Another variable is the number of times the listeners hear and segment the piece; Mungan et al report that in three segmentation trials, listeners were already relatively accurate in the first trial, with little subsequent change 23 .…”