Estimating the Structural Segmentation of Popular Music Pieces Under Regularity Constraints

Abstract

Music structure estimation has recently emerged as a central topic within the field of Music Information Retrieval. Indeed, as music is a highly structured information stream, knowledge of how a music piece is organized represents a key challenge to enhance the management and exploitation of large music collections. This article focuses on the benefits that can be expected from a regularity constraint on the structural segmentation of popular music pieces. Specifically here, we study how a constraint which favors structural segments of comparable size provides a better conditioning of the boundary estimation process. Firstly, we propose a formulation of the structural segmentation task as an optimization process which separates the contribution from the audio features and the one from the constraint. We illustrate how the corresponding cost function can be minimized using a Viterbi algorithm. We present briefly its implementation and results in three systems designed for and submitted to the MIREX 2010, 2011 and 2012 evaluation campaigns. Then, we explore the benefits of the regularity constraint as an efficient mean for combining the outputs of a selection of systems presented at MIREX between 2010 and 2015, yielding a level of performance competitive to that of the state-of-the-art on the "MIREX10" dataset (100 J-Pop songs from the RWC database).

Audio-Based Music Structure Analysis: Current Trends, Open Challenges, and Applications

“…Additionally, Sargent et al (2017) propose a technique that specifically employs the regularity principle by constraining the sizes of the final segmentation. Moreover, a method that employs the homogeneity, repetition, and regularity principles using a Bayesian approach yields stateoftheart results when combining the two subtasks of segmentation and labeling (Shibata et al, 2019).…”

Barwise Music Structure Analysis with the Correlation Block-Matching Segmentation Algorithm

“…The boundary retrieval problem can be approached as an optimization problem (Jensen, 2006;Sargent et al, 2016). In particular, by associating a score u(S) to each potential segment S, the optimal segmentation Z * is the segmentation maximizing 5 the sum of all its segment scores:…”

Barwise Compression Schemes for Audio-Based Music Structure Analysis

“…Segmentation is obtained via the dynamic programming algorithm presented in [7], and inspired from [27]. The principle of dynamic programming is to solve a combinatorial optimization problem by dividing it in several atomic problems, easier to solve, and which solutions can be stiched together to form the global solution.…”

“…To account for regularity constraints as in [27], the present algorithm adds a regularity penalty p(n) to the plain convolution score, which is a function of the size n (in bars) of the segment. This function p(n) is equal to 0 if n = 8, 1 4 if n = 4, 1 2 if n ≡ 0 (mod 2), and finally 1 otherwise.…”

“…This function p(n) is equal to 0 if n = 8, 1 4 if n = 4, 1 2 if n ≡ 0 (mod 2), and finally 1 otherwise. This penalty function is musically motivated for pop music, as segments are more likely to be of size 4 or 8 bars (especially in RWC-Pop with MIREX10 annotations [27]), than of odd bar size, as shown in Figure 5.…”