Infinite-state spectrum model for music signal analysis

12 years 8 months ago

Download www.jonathanleroux.org

This paper presents a nonparametric Bayesian extension of nonnegative matrix factorization (NMF) for music signal analysis. Instrument sounds often exhibit non-stationary spectral characteristics. We introduce inﬁnite-state spectral bases into NMF to represent time-varying spectra in polyphonic music signals. We describe our extension of NMF with inﬁnite-state spectral bases generated by the Dirichlet process in a statistical framework, derive an efﬁcient optimization algorithm based on collapsed variational inference, and validate the framework on audio data.

Masahiro Nakano, Jonathan Le Roux, Hirokazu Kameok

Real-time Traffic

ICASSP 2011 | Inﬁnite-state Spectral Bases | Music Signals | Non-stationary Spectral Characteristics | Signal Processing |

claim paper

» Music dereverberation using harmonic structure source model and Wiener filter

» Sensitivity Analysis for Direction of Arrival Estimation using a RootMUSIC Algorithm

» Operational performances of a MUSIC algorithm robust to outliers

» On the Modeling of Time Information for Automatic Genre Recognition Systems in Audio Signa...

» Modeling musical attributes to characterize ensemble recordings using rhythmic audio featu...

» Analysis of Musical Instrument Sounds by SourceFilterDecay Model

» IDivergencebased dereverberation method with auxiliary function approach

» Features and classifiers for the automatic classification of musical audio signals

Post Info
More Details (n/a)

Added	20 Aug 2011
Updated	20 Aug 2011
Type	Journal
Year	2011
Where	ICASSP
Authors	Masahiro Nakano, Jonathan Le Roux, Hirokazu Kameoka, Nobutaka Ono, Shigeki Sagayama

Comments (0)

Sciweavers

Infinite-state spectrum model for music signal analysis

ICASSP 2011 | Inﬁnite-state Spectral Bases | Music Signals | Non-stationary Spectral Characteristics | Signal Processing |

Explore & Download

Productivity Tools

Document Tools

Image Tools

Sciweavers