The intellegibility and naturalness of synthetic speech strongly depends on its prosodic quality. Departing from works by Mixdorff on a linguistically motivated model of German intonation based on the Fujisaki model, the current paper presents statistical results concerning the relationship between linguistic and phonetic information underlying an utterance and its prosodic features. Statistica...

The authors have already presented a method for automatic extraction of accent and phrase commands of a model from a given F0 contour of speech. This paper describes improvements introduced to cope with difficulties encountered by the previous method, especially in connection with the extraction of accent commands, and reports the results of experiments conducted for the evaluation of the curre...

The main body of our study derives from the processing of 5 subjects’ reading of a corpus from a Swedish novel. Two of the subjects were females. Intonation contours on a log frequency scale have been sampled and normalised to eliminate differences in mean tonal level and duration. As a result, intonation patterns across speakers are brought out revealing individual performances as well as grou...

Automatic prosodic events detection and annotation are important for both speech understanding and natural speech synthesis. In this paper, the complementary model method is proposed to detect prosodic events. This method discards the independent assumption between the acoustic features and the lexical and syntactic features, models not only the features of the current syllable but also the con...

This paper examines word-level prosody in Fataluku, an underdocumented Papuan language of East Timor. Although information about Fataluku remains very limited, at least three alternative analyses of word-level prosody have been suggested: predictable accent, phonemic accent and phonemic tone. This study investigates the phonetic facts of the f0 contour in Fataluku, comparing the behavior of f0 ...

This paper describes a novel pitch pattern generation method for speech synthesis using Hidden Markov Models (HMMs). In the proposed method, the F 0 contours of minor phrase are modeled by HMMs (pitch-HMMs). The pitch-HMMs are trained using F 0 and F 0 considering phonetic environments (e.g. accent type, mora count, mora position, phonemic category, etc.). To evaluate the pitch-HMMs, accent ide...

Recently it has been proposed to base intonational typology on the way languages convey focus (pragmatic prominence). Generally, languages can enhance, add or delete phrase boundary tones (phrase languages) or add an extra pitch accent (intonation languages) to mark focus. Tone languages are somewhat problematic for this typology as it is difficult to make direct observations on their sentence ...

We argue that phonetically invariant realizations of phonological categories imply the synchronic and diachronic imperviousness of such categories to phonological rules and sound laws. We claim that phonetic invariance is the foundation of phonological stability. The category we discuss in this contribution is the pitchaccent. We provide a parametric phonetic description of this phonological ca...

The present study examines the in uence of intonational contrasts of German and Japanese on the F0 patterns of Japanese learners of German. It was found that Japanese learners generally place the word accent in compound words on the right-most (the generic) term which suggests that they adopt rules of accent concatenation of Japanese. In contrast to native speakers of German, they tend to place...

Languages with primarily delimitative function of word stress commonly make use of accentual phrases (APs) in their intonational phonology (e.g. Tamil or French). Slovak and Hungarian are genetically unrelated but geographically close languages with word-initial lexical stress. In this paper we compared the stylised f0 of single accent groups (AGs) with the f0 level pattern of the entire intona...