Skip to Main content Skip to Navigation
New interface
Conference papers

Optimal feature set and minimal training size for pronunciation adaptation in TTS

Marie Tahon 1 Raheel Qader 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 : Text-to-Speech (TTS) systems rely on a grapheme-to-phoneme converter which is built to produce canonical, or statically stylized, pronunciations. Hence, the TTS quality drops when phoneme sequences generated by this converter are inconsistent with those labeled in the speech corpus on which the TTS system is built, or when a given expressivity is desired. To solve this problem, the present work aims at automatically adapting generated pronunciations to a given style by training a phoneme-to-phoneme conditional random field (CRF). Precisely, our work investigates (i) the choice of optimal features among acoustic, articulatory, phonological and linguistic ones, and (ii) the selection of a minimal data size to train the CRF. As a case study, adaptation to a TTS-dedicated speech corpus is performed. Cross-validation experiments show that small training corpora can be used without much degrading performance. Apart from improving TTS quality, these results bring interesting perspectives for more complex adaptation scenarios towards expressive speech synthesis.
Complete list of metadata
Contributor : Damien Lolive Connect in order to contact the contributor
Submitted on : Wednesday, June 29, 2016 - 11:52:35 AM
Last modification on : Wednesday, November 3, 2021 - 6:04:13 AM


  • HAL Id : hal-01338853, version 1


Marie Tahon, Raheel Qader, Gwénolé Lecorvé, Damien Lolive. Optimal feature set and minimal training size for pronunciation adaptation in TTS. International Conference on Statistical Language and Speech Processing (SLSP), Oct 2016, Pilsen, Czech Republic. ⟨hal-01338853⟩



Record views