Automatic Stylisation and Modelling of French and Italian Intonation

Campione, Estelle; Hirst, Daniel; Véronis, Jean

doi:10.1007/978-94-011-4317-2_9

Cited by 26 publications

(6 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A particularly salient result from this study is that, regardless of pause class and the context in which a pause is deployed, pauses rarely exceeded 1.5 s (<2%). This finding is reiterated throughout the pause duration literature (see e.g., Campione & Véronis, 2002), evidence in itself that 1.5 s may be a landmark in the ethology of human timing. The value of 1.5 s is not to be taken as a fixed constant such as might be encountered in physics, but as a narrow region, perhaps better notated as 2 ± 1 s. The notion that 2 ± 1 s is a watershed in the experience of temporal duration derives from other sources as well.…”

Section: Resultssupporting

confidence: 57%

More laws for pauses: Replication and generalization.

Mezaraups

Gilden

2024

Journal of Experimental Psychology: Learning, Memory, and Cogni

View full text Add to dashboard Cite

The basic timescales governing animal life are generally determined by body size. Pauses in naturally occurring human speech were investigated to determine if pause timescales are also sensitive to body size. Reported is an analysis of pause duration allometry in recorded interviews of 61 athletes. Pauses were divided into three classes based on whether they occurred during fluid speech or whether they preceded or followed a filled pause (i.e., “um”). Allometric laws relating body size to pause size were found for all three classes—larger people take longer pauses. The derived allometric exponents were used to evaluate a theory of how people experience the passage of time. The theory associates the experience of time passage with the distal flow of time through the mathematics of bounded exponential growth. Nonlinearities inherent in the theory are shown to predict, in detail, the way body size interacts with linguistic context in the deployment of pauses. The theory provides a meaningful framework for understanding how time is experienced as a felt quantity and how pauses are negotiated in everyday speech.

show abstract

Section: Resultssupporting

confidence: 57%

More laws for pauses: Replication and generalization.

Mezaraups

Gilden

2024

Journal of Experimental Psychology: Learning, Memory, and Cogni

View full text Add to dashboard Cite

show abstract

“…INTSINT as in Campione et al, 2000). Ananthakrishnan and Narayanan (2008) and Jeon and Liu (2012) retain only two possibilities for pitch accent (presence/absence) and only two for boundaries (presence/absence).…”

Section: Automatic Labeling Of Prosodymentioning

confidence: 99%

“…Third, the algorithms used range from rule-based systems (Mertens, 1987a(Mertens, , 1987b(Mertens, , 1989Campione et al, 2000;Campione and Véronis, 2001; Bartkova et al 2012) to supervised learning techniques, including HMM (Geoffrois, 1995), decision trees (Wightman and Ostendorf, 1994), neural networks, and so on (Braunschweiler, 2005;Ananthakrishnan and Narayanan, 2008;Wagner, 2009;Rosenberg, 2010). Jeon and Liu (2012) give a comparative overview indicating techniques (algorithms) and detection accuracy.…”

Section: Automatic Labeling Of Prosodymentioning

confidence: 99%

Polytonia

Mertens

2021

J. of Speech Sci.

View full text Add to dashboard Cite

This paper first proposes a labeling scheme for tonal aspects of speech and then describes an automatic annotation system using this transcription. This fine-grained transcription provides labels indicating pitch level and pitch movement of individual syllables. Of the five pitch levels, three (low, mid, high) are defined on the basis of pitch changes in the local context and two (bottom, top) are defined relative to the boundaries of the speaker’s global pitch range. For pitch movements, both simple and compound, the transcription indicates direction (rise, fall, level) and size, using size categories (pitch intervals) adjusted relative to the speaker’s pitch range. The automatic tonal annotation system combines several processing steps: segmentation into syllable peaks, pause detection, pitch stylization, pitch range estimation, classification of the intra-syllabic pitch contour, and pitch level assignment. It uses a dedicated and rule-based procedure, which unlike commonly used supervised learning techniques does not require a labeled corpus for training the model. The paper also includes a preliminary evaluation of the annotation system, for a reference corpus of nearly 14 minutes of spontaneous speech in French and Dutch, in order to quantify the annotation errors. The results, expressed in terms of standard measures of precision, recall, accuracy and Fmeasure are encouraging. For pitch levels low, mid and high an F-measure between 0.946 and 0.815 is obtained and for pitch movements a value between 0.708 and 1. Provided additional modules for the detection of prominence and prosodic boundaries, the resulting annotation may serve as an input for a phonological annotation.

show abstract

“…Stylizations come in many flavors; we are interested here in (semi-)automatic procedures. The Momel system [8,9] models the pitch curve by a quadratic spline function, as a sequence of parabolic segments. Other systems use linear regression to determine inflection points.…”

Section: Pitch Stylization Based On Tonal Perceptionmentioning

confidence: 99%