The GIPSA-Lab Text-To-Speech System for the Blizzard Challenge 2023 - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2023

The GIPSA-Lab Text-To-Speech System for the Blizzard Challenge 2023

Résumé

This paper describes the GIPSA-Lab submission to the Blizzard Challenge 2023. The Text-To-Speech system trained for this challenge is a Transformer-based non-autoregressive encoderdecoder architecture based on FastSpeech2. Updates of the FastSpeech2 framework were provided to specifically train the model on orthographic inputs, which is our main focus for this edition of the challenge. This model was trained with both orthographic and phonetic transcriptions of the same dataset. An additional phonetic prediction layer was added to the model. This additional layer enables to train the text encoder on phonetic prediction alone, without the need for audio recordings.
Fichier principal
Vignette du fichier
lenglet23_blizzard.pdf (709.61 Ko) Télécharger le fichier
Origine : Fichiers éditeurs autorisés sur une archive ouverte

Dates et versions

hal-04269935 , version 1 (06-11-2023)

Identifiants

Citer

Martin Lenglet, Olivier Perrotin, Gérard Bailly. The GIPSA-Lab Text-To-Speech System for the Blizzard Challenge 2023. 18th Blizzard Challenge Workshop, Aug 2023, Grenoble, France. pp.34-39, ⟨10.21437/Blizzard.2023-3⟩. ⟨hal-04269935⟩
58 Consultations
28 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More