Prosodic focus control in reply speech generation for a spoken dialogue system of information retrieval

Kiriyama, Shinya; Hirose, Keikichi; Minematsu, Nobuaki

doi:10.1109/wss.2002.1224393

Cited by 6 publications

(4 citation statements)

References 3 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The types of prominence relationships that are modeled in Pierrehumbert and Beckman's synthesis program are also modeled in a comparably direct way in Hiroya Fujisaki's model that was first described in Fujisaki and Sudo 1971 and which has been adopted in an impressive body of later studies that became the basis for the intonation synthesis modules in several leading Japanese text-to-speech and concept-to-speech systems (e.g., Fujisaki and Hirose 1993, Fujisaki et al 1994, Hirai et al 1996, Kiriyama et al 2002. In the Fujisaki framework, the prominence relationship among different intonation phrases is modeled by specifying different amplitudes for their phrase commands.…”

Section: The Relationship Between Tone Features and Pitch Range Featuresmentioning

confidence: 99%

Prominence Marking in the Japanese Intonation System

Venditti

Maekawa

Beckman³

2008

Oxford Handbooks Online

View full text Add to dashboard Cite

Japanese uses a variety of prosodic mechanisms to mark focal prominence, including local pitch range expansion, prosodic restructuring to set off the focal constituent, post-focal subordination, and prominence-lending boundary pitch movements, but (notably) not manipulation of accent. In this paper, we describe the Japanese intonation system within the Autosegmental-Metrical model of intonational phonology, and review these prosodic mechanisms that have been shown to mark focal prominence. We point out potentials for ambiguity in the prosodic parse, and discuss their larger implications for the development of a tenable general theory of prosody and its role in the marking of discourse structure. Venditti, Maekawa, & Beckman: Prominence marking in the Japanese intonation system.-p. 2 In this chapter, we will review the currently standard AM framework account of how the prosodic marking of prominence works in the Japanese system (in Section 3). We will show that, just as in the English prosodic system, there are several different prosodic mechanisms in Japanese for marking some constituents as having focal prominence and others as being relatively reduced and out of focus. We will then describe four phenomena that are the locus of lively discussion and controversy in the further development of this AM framework account (in Section 4), before discussing the larger implications that these phenomena have for the development of a tenable general theory of the role of prosody in the marking of discourse prominence (in Section 5). Before we begin this description, however, we need to make explicit some assumptions about what prosody is and about the nature of phonetic representations that we use to study it. An understanding of these assumptions is essential background for understanding the AM framework description of Japanese. Section 2 lays out this background. 2 Elements of an AM description of Japanese intonation In describing the prominence-marking mechanisms of Japanese, we will use the account of the Japanese intonation system that is encoded in the X-JToBI labeling conventions (Maekawa et al., 2002). These conventions build on the earlier J_ToBI conventions described by Venditti (1997, 2005), adding tags for elements of intonation contours that have not been noted in the previous literature in English, although they are observed even in relatively formal spontaneous speech such as the conference presentations in the Corpus of Spontaneous Japanese (henceforth "CSJ"; Maekawa 2003, Kokuritsu Kokugo Kenkyuujo [National Institute for Japanese Language] 2006). Both the original and the expanded tag set are intended to describe intonation contours for standard (Tokyo) Japanese. The intonation systems of other dialects are known to differ, some rather dramatically. However, they have either not been studied at all (most regional dialects) or have been studied less extensively (e.g., the Osaka system, as described by Kori (1987) and others). Both tag sets also assume a description couched in the AM framework. 2.1 The AM f...

show abstract

Section: The Relationship Between Tone Features and Pitch Range Featuresmentioning

confidence: 99%

Prominence Marking in the Japanese Intonation System

Venditti

Maekawa

Beckman³

2008

Oxford Handbooks Online

View full text Add to dashboard Cite

show abstract

“…The prosodic variations are statistically modeled by using the average difference in prosodic features between the original and synthetic speech of the training data. The Fujisaki model (Fujisaki, 1983) is another wellknown prosodic model in ESS (Chen et al, 2004;Kiriyama et al, 2002). Ochi et al (2009) used this model to control focus by modifying the Fujisaki model parameters.…”

Section: Introductionmentioning

confidence: 99%

Hierarchical stress modeling and generation in mandarin for expressive Text-to-Speech

Tao

Hirose

et al. 2015

Speech Communication

View full text Add to dashboard Cite

“…Nevertheless, in real communication, people do not always answer, they can also ask for further information to clarify the question, etc. To provide a natural SDS and expand its application, SDS system should have the capability to generate natural and expressive interrogative sentence [2].…”

Section: Introductionmentioning

confidence: 99%

Assign stress for interrogative sentences via syntax structure mapping

Li,

Liu,

et al. 2012

Speech Prosody 2012

View full text Add to dashboard Cite

Compared to the prosody study in statement-like sentence, little research has been conducted in interrogative sentence, which will definitely improve the naturalness of spoken dialog system. This paper proposes a computational model to assign stress in interrogative sentences using linguistic knowledge. We firstly investigate the interrogative sentences perceptually to label their stress locations and then summarize a series of syntax-stress mapping rules. The dependency syntax is adopted in this work and automatically obtained by Stanford Parser. All the dependency relations in Stanford Parser are studied and stable stress patterns are assigned. Syntax depth is considered as a factor to weight each dependency relation. Finally, the linear superposition of stress score calculated from each dependency relation is assigned to prosodic words. Experiments show that the weighted accuracy of syntax-stress mapping is 74%. A detailed analysis of results is also provided.

show abstract

Prosodic focus control in reply speech generation for a spoken dialogue system of information retrieval

Cited by 6 publications

References 3 publications

Prominence Marking in the Japanese Intonation System

Prominence Marking in the Japanese Intonation System

Hierarchical stress modeling and generation in mandarin for expressive Text-to-Speech

Assign stress for interrogative sentences via syntax structure mapping

Contact Info

Product

Resources

About