C-PROM-Task: A New Annotated Dataset for the Study of French Speech Prosody
Date issued
August 30, 2013
From page
27
To page
30
Subjects
corpus spoken French prosodic annotation prominence phrasing.
Abstract
The aim of this paper is to describe C-PROM-Task, a dataset created and annotated in the same spirit as C-PROM [1]. CPROM-Task was annotated following a perceptually-based and computer-assisted procedure for the study of syllabic prominences, syllabic disfluences and two ranges of prosodic units. All told, C-PROM-Task comprises recordings and annotated TextGrids of story-telling by 20 native French
speakers from Switzerland. The entire dataset is 2 hours 20 minutes long. Some observations are also made regarding accentuation (prominence rate), disfluence rate and phrasing (length of prosodic units) in the corpus.
speakers from Switzerland. The entire dataset is 2 hours 20 minutes long. Some observations are also made regarding accentuation (prominence rate), disfluence rate and phrasing (length of prosodic units) in the corpus.
Notes
, 2013
Event name
TRASP'2013
Location
Aix-en-Provence, France
Publication type
conference paper