The IRISA Text-To-Speech System for the Blizzard Challenge 2016

Pierre Alain 1 Jonathan Chevelu 1 David Guennec 1 Gwénolé Lecorvé 1 Damien Lolive 1
1 EXPRESSION - Expressiveness in Human Centered Data/Media
UBS - Université de Bretagne Sud, IRISA-D6 - MEDIA ET INTERACTIONS
Abstract : This paper describes the implementation of the IRISA unit selection-based TTS system for our participation in the Blizzard Challenge 2016. We describe the process followed to build the voices from given data and the architecture of our system. The search is based on a A* 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.
Complete list of metadatas

Cited literature [25 references]  Display  Hide  Download

https://hal.inria.fr/hal-01375897
Contributor : Damien Lolive <>
Submitted on : Monday, October 3, 2016 - 5:03:32 PM
Last modification on : Thursday, February 7, 2019 - 3:46:09 PM
Long-term archiving on : Friday, February 3, 2017 - 2:08:44 PM

File

IRISA_Blizzard2016.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-01375897, version 1

Citation

Pierre Alain, Jonathan Chevelu, David Guennec, Gwénolé Lecorvé, Damien Lolive. The IRISA Text-To-Speech System for the Blizzard Challenge 2016. Blizzard Challenge 2016 workshop, Sep 2016, Cupertino, United States. ⟨hal-01375897⟩

Share

Metrics

Record views

651

Files downloads

147