Revealing structure in vocalisations of parrots and social whales

Revealing structure in vocalisations of parrots and social whales

Maria Florencia Noriega Romero Vargas

Doctoral thesis

Date of Examination: 2017-08-07

Date of issue: 2018-08-06

Advisor: Wörgötter, Florentin Prof. Dr.

Referee: Timme, Marc Prof. Dr.

Referee: Hammerschmidt, Kurt Dr.

Persistent Address: http://hdl.handle.net/11858/00-1735-0000-002E-E469-8

Files in this item

Name: thesis.pdf

Size: 16,6 MB

Format: PDF

View/Open

Abstract

English

This thesis proposes methods to investigate structure in bioacoustic signals. For this two frameworks are proposed. The first concerns the automatic annotation of audio recordings by using supervised machine learning methods. The second concerns a quantitative analysis of temporal and combinatorial patterns in vocal sequences of animals by using non-parametric statistics. These methods are used to investigate vocalisations of two wild living animals — known very little — in their natural ecosystems: lilac crowned parrots and pilot whales.

Keywords: bioacoustics; pilot whales; parrots; audio signal processing; machine learning; animal vocalisations