The IRISA Text-To-Speech System for the Blizzard Challenge 2017

Abstract : This paper describes the implementation of the IRISA unit selection-based TTS system for our participation to the Blizzard Challenge 2017. We describe the process followed to build the voice from given data and the architecture of our system. It uses a selection cost which integrates notably a DNN-based prosodic prediction and also a specific score to deal with narrative/direct speech parts. Unit selection is based on a Viterbi-based algorithm with preselection filters used to reduce the search space. A penalty is introduced in the concatenation cost to block some concatenations based on their phonological class. Moreover, a fuzzy function is used to relax this penalty based on the concatenation quality with respect to the cost distribution. Integrating a lot of constraints, this system achieves average results compared to others.
Type de document :
Communication dans un congrès
Blizzard Challenge, Aug 2017, Stockholm, Sweden
Liste complète des métadonnées

Littérature citée [24 références]  Voir  Masquer  Télécharger
Contributeur : Damien Lolive <>
Soumis le : mercredi 13 décembre 2017 - 09:53:42
Dernière modification le : vendredi 11 janvier 2019 - 14:27:06


Fichiers produits par l'(les) auteur(s)


  • HAL Id : hal-01662361, version 1


Damien Lolive, Pierre Alain, Nelly Barbot, Jonathan Chevelu, Gwénolé Lecorvé, et al.. The IRISA Text-To-Speech System for the Blizzard Challenge 2017. Blizzard Challenge, Aug 2017, Stockholm, Sweden. 〈hal-01662361〉



Consultations de la notice


Téléchargements de fichiers