StyleWaveGAN: Style-based synthesis of drum sounds using generative adversarial networks for higher audio quality - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2022

StyleWaveGAN: Style-based synthesis of drum sounds using generative adversarial networks for higher audio quality

Résumé

In this paper we introduce StyleWaveGAN, a stylebased drum sound generator that is a variation of StyleGAN, a state-of-the-art image generator. By conditioning StyleWaveGAN on the type of drum, we are able to synthesize waveforms faster than real-time on a GPU directly in CD quality up to a duration of 1.5s while retaining some control over the generation. We also introduce an alternative to the progressive growing of GANs and experimented on the effect of dataset balancing for generative tasks. The experiments are carried out on an augmented subset of a publicly available dataset comprised of different drums and cymbals. We evaluate against two recent drum generators, WaveGAN and NeuroDrum, demonstrating significantly improved generation quality using two quality measures: first the Frechet Audio Distance and second a perceptual test.

Domaines

Son [cs.SD]
Fichier principal
Vignette du fichier
eusipco_22.pdf (444.86 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-03693960 , version 1 (13-06-2022)

Identifiants

  • HAL Id : hal-03693960 , version 1

Citer

Antoine Lavault, Axel Roebel, Matthieu Voiry. StyleWaveGAN: Style-based synthesis of drum sounds using generative adversarial networks for higher audio quality. 30th European Signal Processing Conference (EUSIPCO 2022), Aug 2022, Belgrade, Serbia. ⟨hal-03693960⟩
75 Consultations
183 Téléchargements

Partager

Gmail Facebook X LinkedIn More