This paper presents a novel technique for empirically identifying regions of analysis for time/value information. The technique relies on analysis of mutual information between the contour, and some variable of interest. We present the use of this technique in the analysis of prosody in American English speech, where we identify valuable regions of analysis for the classification of phrase endi...

Our approach for unsupervised extraction of prosodic structure in spontaneous speech consists of the four steps: chunking into interpausal units, syllable nucleus extraction, prosodic boundary detection, and pitch accent detection. The extraction is based on acoustic features derived from F0 parameterization, and on energy and segment duration features. Phrase boundaries and accents are detecte...

Many aspects of prosody prediction in speech synthesis could be improved, from placement of symbolic accent and phrase boundary markers to control of continuously varying parameters (e.g., duration, fundamental frequency). The goal of this work is to develop algorithms for predicting aspects of fundamental frequency typically said to have gradient variation: pitch range and prominence. In addit...

This paper examines predictions made by two theories of the relationship between pitch accent and focus. The empirical evidence presented suggests that listeners are sensitive to a variety of factors that may affect the focus projection ability of pitch accents, that is the ability of a pitch accent on one word to mark focus on a larger constituent. The findings suggest that listeners' interpre...

A read speech corpus was collected in order to study the tonal targets alignment of two Pisa Italian peak accents. The production of the two accents was forced by inducing a broad and a contrastive interpretation of the target words inserted in small paragraphs. Results of latency and F0 measurements reveal that the contrastive accent has a low target at a 'fixed' distance after the peak wherea...

The paper proposes a framework for automatic prosody labeling. The labeling involves detection of the location of accented syllables and phrase boundaries, and recognition of pitch accent and boundary tone types. A number of classification models are designed to perform these tasks on the basis of small vectors of acoustic features. The models achieve high accuracy and their performance is comp...

In the present paper, we investigate whether decreasing phrase-initial accent peaks over the course of several prosodic phrases is used to signal coherence among phrases in spontaneous Swedish. Since little is known about prosodic phrasing and discourse prosody in spontaneous Swedish, the investigation has an exploratory nature and a number of different methods have been employed in order to an...

Much recent work on German and English intonation has addressed the impact of information structure on prosodic patterns in terms of the focus/background partition. In contrast with stress-accent languages such as Italian, Spanish or English, French does not appear to signal focus through pitch accent assignment, rather it appears to mainly exploit prosodic edge marking for the same purposes. T...

We performed a cognitive experiment in order to obtain evidence that prosody is used for the segmentation of continuous speech for understanding in real time. We assume that the cognitive unit in Japanese speech is the accentual phrase (AP) which is represented by low and high pitch patterns according to the Japanese accent rule. Meaningless speech which consisted of two APs was used as the exp...

This paper reports on a comparative analysis of accentuation strategies within Italian and Dutch NPs. Accent-patterns were obtained in a (semi-)spontaneous way via a simple dialogue-game with 8 Dutch speakers and 4 Italian ones. In this way, target descriptions of all speakers were obtained in the following four contexts: all new, single contrast in the adjective, single contrast in the noun, a...