Volume 50, Issue 3

This article presents a preliminary study of the phonetic interpretation of tone in Igbo, a Kwa language of southeastern Nigeria. The experimental method varies the speaker’s pitch range orthogonally with variation in tonal material, and fits a model to measurements of maxima and minima in the resulting Fo contour. A new interpretation of downstep is proposed as a result.

The goal of this experiment is to find the most important phonetic features of Dutch accent-lending pitch movements, in terms of shape, pitch level and alignment with the segmental structure. Time pressure is used as a heuristic method to isolate important phonetic aspects of pitch movements, assuming that under time pressure the speaker will preserve those aspects. In a production experiment, accent-lending rises (‘l’) and falls (‘A’) were realized under various types of time pressure. The pitch rise is time-compressed under all pressure types, which would mean that the shape of the rise is relatively unimportant. The segmental alignment of the rise proved to be more important: the onset of the rise is synchronized with the syllable onset. For the fall no fixed synchronization point was found, but its shape was relatively invariant, indicating that shape rather than exact timing is the more important feature of the fall.

Clause-internal filled pauses and preceding peak fundamental frequency (F₀) values were analyzed to determine whether the intonation of filled pauses is relative to, or independent of, prior prosodic context. Higher peaks were found to be systematically associated with higher filled-pause values, supporting the ‘relative’ hypothesis. A linear model, in which filled-pause F₀ was expressed as an invariant (over speakers) proportion of the distance between preceding peak F₀ and a speaker-dependent baseline F₀, produced results nearly identical to those of a two-parameter model in which the coefficients of peak and baseline were allowed to vary freely. The model was less appropriate for filled pauses after sentence-initial peaks, but unaffected by temporal variables.

The present research aims at a partial model for synthesized text intonation, restricted to utterances consisting of a single intonation phrase. Intonational characteristics of short utterances in different positions within a paragraph are investigated. The effects of position in the paragraph on sentence prosody are mainly located in the beginning of the utterance. A model is formulated to capture the main findings. The model is then perceptually evaluated. The results indicate that the model can be used to improve the naturalness of rule-based intonation in a system for automatic text-to-speech conversion.

This article describes a perceptual evaluation of the prosodic structure of a spontaneously produced monologue. It was found that the speaker studied demarcates larger-scale topical units in spoken discourse by means of intonation (use of melodic boundary markers, scaling of maxima in pitchmovements, general decline in average pitch) and by the use of pauses with variable durations. In a perception test, it was examined to what extent these prosodic devices may be important to listeners. Subjects were confronted with three unintelligible (band-pass-filtered) versions of a fragment of the elicited monologue: (1) with the original prosody unchanged; (2) with constant pause duration and the original speech melody; (3) with monotonous pitch and the original pause structure. They were instructed to indicate the boundaries of the larger-scale topical units in the three versions. Subjects were able to detect correctly the major discourse boundaries in all three filtered versions in a significant number of cases. They performed best when confronted with version 1. Versions 1 and 2, in their turn, did better than version 3, which suggests that, in the performance of this speaker, intonation is a perceptually more important factor than pause for the clarification of the topical make-up of a text, though the latter dimension is certainly not negligible.

This article describes how well prosodic information correlates with the topic structure of discourse. To investigate this correlation systematically, we first introduce the notion of utterance unit , which can be viewed as the basic unit in conversations. We then define four topic boundary classes: topic shift , topic continuation , elaboration , and speech - act continuation . The prosodic parameters – onset fundamental frequency (F0), final F0, and peak F0 ratio – are measured at these topic boundaries to show how these prosodic parameters change with the topic structure. Finally, we discuss discourse structure identification via the prosodic parameters.