Introduction of the speaking rate in the model of speech recognition (Q1607824)

Summary: We propose an improvement to the centisecond TLHMM model of the second author (1999) applied to the sound duration. Indeed, the distribution of the sound duration depends on the speaking rate. An adaptation in a post-processing step is needed. This adaptation is studied by proposing a model of the speaking rate based on average syllabic duration. The experiments elaborated on a set of BDSONS show the interest of this approach. This work is a continuation of those by the second author (1999) and \textit{N. Suaudeau} [Un modèle probabiliste pour intégrer la dimension temporelle dans un système de reconnaissance automatique de parole, Thèses parole avec résumés, Université Rennes I, March 1994].