Skip to Main content Skip to Navigation
New interface
Conference papers

Disfluency Insertion for Spontaneous TTS: Formalization and Proof of Concept

Raheel Qader 1 Gwénolé Lecorvé 1 Damien Lolive 1 Pascale Sébillot 2 
2 LinkMedia - Creating and exploiting explicit links between multimedia fragments
Inria Rennes – Bretagne Atlantique , IRISA-D6 - MEDIA ET INTERACTIONS
Abstract : This paper presents an exploratory work to automatically insert disfluencies in text-to-speech (TTS) systems. The objective is to make TTS more spontaneous and expressive. To achieve this, we propose to focus on the linguistic level of speech through the insertion of pauses, repetitions and revisions. We formalize the problem as a theoretical process, where transformations are iteratively composed. This is a novel contribution since most of the previous work either focus on the detection or cleaning of linguistic disfluencies in speech transcripts, or solely concentrate on acoustic phenomena in TTS, especially pauses. We present a first implementation of the proposed process using conditional random fields and language models. The objective and perceptual evalation conducted on an English corpus of spontaneous speech show that our proposition is effective to generate disfluencies, and highlights perspectives for future improvements.
Complete list of metadata

Cited literature [25 references]  Display  Hide  Download
Contributor : Gwénolé Lecorvé Connect in order to contact the contributor
Submitted on : Monday, July 16, 2018 - 4:44:18 PM
Last modification on : Saturday, August 6, 2022 - 3:33:01 AM
Long-term archiving on: : Wednesday, October 17, 2018 - 3:56:26 PM


Files produced by the author(s)



Raheel Qader, Gwénolé Lecorvé, Damien Lolive, Pascale Sébillot. Disfluency Insertion for Spontaneous TTS: Formalization and Proof of Concept. SLSP 2018 - 6th International Conference on Statistical Language and Speech Processing, Oct 2018, Mons, Belgium. pp.1-12, ⟨10.1007/978-3-030-00810-9_4⟩. ⟨hal-01840798⟩



Record views


Files downloads