Two methods were proposed for the use of prosodic features in speech recognition: one to detect major syntactic (phrase) boundaries as the initial phase of speech recognition, and the other to check the feasibility of the results of ordinary recognition process from the viewpoint of prosodic features. In the rst method, fundamental frequency contours were assumed as waveforms as functions of ti...

Intonational realisation of topic and focus in child Dutch Aoju Chen, Max Planck Institute for Psycholinguistics, Nijmegen Languages express the topic-focus distinction in different ways (e.g. word order, particles, intonation). In this study we focus on intonation and in particular look at type of pitch accent and phrasing. There has been relatively little discussion on the use of intonation t...

A corpus-based method was developed for generating fundamental frequency contours in emotional speech synthesis. The method assumes the generation process model and predicts its command parameters (positions and amplitudes) using binary regression trees with the input of linguistic information of the sentence to be synthesized. Because of the model constraint, a certain quality is still kept in...

In two speeded acceptability experiments we tested which combination of prenuclear accent, nuclear accent and f0interpolation between them is best suited to signal a double contrast in German (i.e., a contrastive topic followed by a contrastive focus). The experimental utterances differed in the prenuclear accent (medialvs. late-peak, i.e., L+H* vs. L*+H), the nuclear accent (earlyvs. medial-pe...

This paper explores the interplay of source correlates of accentuation, examining a hypothesis (the Voice Prominence Hypothesis) that different source parameters are involved and may serve as equivalent. It predicts that where accentuation is not marked by pitch salience there will be more extensive changes in other source parameters. This follows our assumption that prosodic entities such as a...

The automatic prosodic annotation of large speech corpora gains increasing consideration since appropriate databases for the training of prosodic models in speech synthesis and recognition are needed. On linguistic level, correct phrase and accent marking are essential processing steps. The authors developed a neural network based method for signal-based phrase break prediction and tested this ...

This paper presents a model to predict the accent commands (henceforth ACs) of the Fujisaki Model for the F0 contour, being known the phrase commands (henceforth FCs). Accent commands are associated with syllables. For each syllable, an artificial neural network (ANN) decides, with an accuracy of 89.4% whether there will be an associated AC or not. For syllables with associated AC, the amplitud...

A text-to-speech system can effectively imitate distinctive speaking styles when a few critical prosodic features are modeled and controlled. This paper demonstrates the methodology with a number of examples, including the ornamental notes and the amplitude profile that define the singing style of Dinah Shore, the phrase curve that sets off the dramatic speaking style of Martin Luther King Jr, ...

Based on a represention of F0 in mora units (F0mora), several parameters and methods were introduced in order to identify accent types of accentual phrases: new candidates for F0mora estimation method; the addition of relative F0mora parameters and relative power parameters; and a new accent type identification method. The candidates for F0mora were investigated in order to find the best matchi...

Previous studies have proposed that the French phonological system includes an Initial Accent on lexical words, which has rhythmic functions. Descriptions of English prosody propose that phonological phrases may have an early accent as a boundary marker. We report results from a preliminary production experiment, which tested the structural and rhythmic functions of the F0 of initial syllables....