Automatic Modelling of Rhythm and Intonation for Language Identification

This paper deals with an approach to Automatic Language Identification using only prosodic modeling. The traditional approach for language identification focuses mainly on phonotactics because it gives the best results. Recent studies reveal that humans use different levels of perception to identify a language, in particular prosodic cues. Among prosodic features, rhythm is known to carry a substantial information about language identity. Rhythm is produced by the periodicity of a pattern that can be syllable, which is a language specific unit. That is why we introduced the notion of ”Pseudo-Syllable”, derived from the most frequent syllable structure in the world, the Consonant-Vowel structure. In this paper, an automatic and language independent rhythmic units extraction algorithm is described: using a vowel detection algorithm, rhythmic units matching the Consonant-Vowel structure are segmented. Two models describing rhythm and intonation of each language are then learned using Gaussian Mixtures.