Low-Rank Time-Frequency Synthesis - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2014

Low-Rank Time-Frequency Synthesis

Cédric Févotte
Matthieu Kowalski

Résumé

Many single-channel signal decomposition techniques rely on a low-rank factor-ization of a time-frequency transform. In particular, nonnegative matrix factoriza-tion (NMF) of the spectrogram – the (power) magnitude of the short-time Fourier transform (STFT) – has been considered in many audio applications. In this set-ting, NMF with the Itakura-Saito divergence was shown to underly a generative Gaussian composite model (GCM) of the STFT, a step forward from more empiri-cal approaches based on ad-hoc transform and divergence specifications. Still, the GCM is not yet a generative model of the raw signal itself, but only of its STFT. The work presented in this paper fills in this ultimate gap by proposing a novel signal synthesis model with low-rank time-frequency structure. In particular, our new approach opens doors to multi-resolution representations, that were not pos-sible in the traditional NMF setting. We describe two expectation-maximization algorithms for estimation in the new model and report audio signal processing results with music decomposition and speech enhancement.
Fichier principal
Vignette du fichier
nips2014.pdf (822.28 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-01103496 , version 1 (14-01-2015)

Identifiants

  • HAL Id : hal-01103496 , version 1

Citer

Cédric Févotte, Matthieu Kowalski. Low-Rank Time-Frequency Synthesis. Advances in Neural Information Processing Systems (NIPS), Dec 2014, Montréal, France. ⟨hal-01103496⟩
211 Consultations
127 Téléchargements

Partager

Gmail Facebook X LinkedIn More