Filterbank design for end-to-end speech separation - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2020

Filterbank design for end-to-end speech separation

Résumé

Single-channel speech separation has recently made great progress thanks to learned filterbanks as used in ConvTasNet. In parallel, parameterized filterbanks have been proposed for speaker recognition where only center frequencies and bandwidths are learned. In this work, we extend real-valued learned and parameterized filterbanks into complex-valued analytic filterbanks and define a set of corresponding representations and masking strategies. We evaluate these fil-terbanks on a newly released noisy speech separation dataset (WHAM). The results show that the proposed analytic learned filterbank consistently outperforms the real-valued filterbank of ConvTasNet. Also, we validate the use of parameterized filterbanks and show that complex-valued representations and masks are beneficial in all conditions. Finally, we show that the STFT achieves its best performance for 2 ms windows.
Fichier principal
Vignette du fichier
main.pdf (301.31 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-02355623 , version 1 (08-11-2019)
hal-02355623 , version 2 (10-02-2020)

Identifiants

  • HAL Id : hal-02355623 , version 2

Citer

Manuel Pariente, Samuele Cornell, Antoine Deleforge, Emmanuel Vincent. Filterbank design for end-to-end speech separation. ICASSP 2020 - 45th International Conference on Acoustics, Speech, and Signal Processing, May 2020, Barcelona, Spain. ⟨hal-02355623v2⟩
306 Consultations
571 Téléchargements

Partager

Gmail Facebook X LinkedIn More