Change search
ReferencesLink to record
Permanent link

Direct link
Inferring metrical structure in music using particle filters
KTH, School of Computer Science and Communication (CSC), Media Technology and Interaction Design, MID. (Sound and Music Computing)ORCID iD: 0000-0003-1679-6018
2015 (English)In: IEEE Transactions on Audio, Speech and Language Processing, Vol. 23, no 5, 817-827 p.Article in journal (Refereed) Published
Abstract [en]

In this paper, we propose a new state-of-the-art particle filter (PF) system to infer the metrical structure of musical audio signals. The new inference method is designed to overcome the problem of PFs in multi-modal probability distributions, which arise due to tempo and phase ambiguities in musical rhythm representations. We compare the new method with a hidden Markov model (HMM) system and several other PF schemes in terms of performance, speed and scalability on several audio datasets. We demonstrate that using the proposed system the computational complexity can be reduced drastically in comparison to the HMM while maintaining the same order of beat tracking accuracy. Therefore, for the first time, the proposed system allows fast meter inference in a high-dimensional state space, spanned by the three components of tempo, type of rhythm, and position in a metric cycle.

Place, publisher, year, edition, pages
IEEE Press, 2015. Vol. 23, no 5, 817-827 p.
Keyword [en]
approximate inference; Bayesian modeling; beat tracking; downbeat tracking; particle filters (PFs)
National Category
Media Engineering
Research subject
Computer Science; Information and Communication Technology; Speech and Music Communication
URN: urn:nbn:se:kth:diva-193770DOI: 10.1109/TASLP.2015.2409737ISI: 000352281500001ScopusID: 2-s2.0-84954458463OAI: diva2:1040350

QC 20161027

Available from: 2016-10-27 Created: 2016-10-10 Last updated: 2016-11-22Bibliographically approved

Open Access in DiVA

fulltext(1150 kB)8 downloads
File information
File name FULLTEXT01.pdfFile size 1150 kBChecksum SHA-512
Type fulltextMimetype application/pdf

Other links

Publisher's full textScopus

Search in DiVA

By author/editor
Holzapfel, André
By organisation
Media Technology and Interaction Design, MID
Media Engineering

Search outside of DiVA

GoogleGoogle Scholar
Total: 8 downloads
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

Altmetric score

Total: 1 hits
ReferencesLink to record
Permanent link

Direct link