Skip to Main content Skip to Navigation
Conference papers

Uncovering Audio Patterns in Music with Nonnegative Tucker Decomposition for Structural Segmentation

Axel Marmoret 1 Jérémy Cohen 1 Nancy Bertin 1 Frédéric Bimbot 1
1 PANAMA - Parcimonie et Nouveaux Algorithmes pour le Signal et la Modélisation Audio
Inria Rennes – Bretagne Atlantique , IRISA-D5 - SIGNAUX ET IMAGES NUMÉRIQUES, ROBOTIQUE
Abstract : Recent work has proposed the use of tensor decomposition to model repetitions and to separate tracks in loop-based electronic music. The present work investigates further on the ability of Nonnegative Tucker Decompositon (NTD) to uncover musical patterns and structure in pop songs in their audio form. Exploiting the fact that NTD tends to express the content of bars as linear combinations of a few patterns, we illustrate the ability of the decomposition to capture and single out repeated motifs in the corresponding compressed space, which can be interpreted from a musical viewpoint. The resulting features also turn out to be efficient for structural segmentation, leading to experimental results on the RWC Pop data set which are potentially challenging state-of-the-art approaches that rely on extensive example-based learning schemes.
Complete list of metadata

Cited literature [24 references]  Display  Hide  Download

https://hal.archives-ouvertes.fr/hal-02928733
Contributor : Axel Marmoret <>
Submitted on : Wednesday, September 2, 2020 - 6:06:05 PM
Last modification on : Tuesday, October 6, 2020 - 1:14:02 PM
Long-term archiving on: : Wednesday, December 2, 2020 - 5:13:43 PM

File

Uncovering audio patterns in m...
Files produced by the author(s)

Identifiers

  • HAL Id : hal-02928733, version 1

Citation

Axel Marmoret, Jérémy Cohen, Nancy Bertin, Frédéric Bimbot. Uncovering Audio Patterns in Music with Nonnegative Tucker Decomposition for Structural Segmentation. ISMIR 2020 - 21st International Society for Music Information Retrieval, Oct 2020, Montréal (Online), Canada. pp.1-7. ⟨hal-02928733⟩

Share

Metrics

Record views

243

Files downloads

250