Automatic prosodic clustering of humpback whales song

Gautier Picot (1) , Olivier Adam (2) , Maïtine Bergounioux (1) , Hervé Glotin (3) , François-Xavier Mayer (4)

1 MAPMO - Mathématiques - Analyse, Probabilités, Modélisation - Orléans
2 LISSI - Laboratoire Images, Signaux et Systèmes Intelligents
3 LSIS - Laboratoire des Sciences de l'Information et des Systèmes
4 Association MEGAPTERA

We automatically segmented a humpback whale song using the Roger Payne's principle of sound unit and we patterned and automatically classified intonations of sound units thanks to an unsupervised algorithm to describe recurrent patterns. We determined 6 different patterns of intonation. Some different sound units present the same pattern of intonation and one sound unit can be submitted to different patterns of intonation. Intonation could be related to information's transmission between humpback whales while they are singing.

Mots clés: prosody; whale song

Domaines: Traitement du signal et de l'image [eess.SP]; Biodiversité et Ecologie; Traitement du signal et de l'image [eess.SP]

Gautier Picot, Olivier Adam, Maïtine Bergounioux, Hervé Glotin, François-Xavier Mayer. Automatic prosodic clustering of humpback whales song. 2008. ⟨hal-00315949⟩