Caracterización y modelado de la producción de sonidos de las ballenas beluga (Delphinapterus Leucas) basado en modelos de análisis / síntesis de voz
[EN] This thesis deals with the study of the sounds produced by beluga whales (Delphinapterus leucas) with a fundamental objective: its characterization and modeling. To this end, analysis / synthesis algorithms of the sounds produced by these animals are proposed. These algorithms are inspired by r...
| Main Author: | |
|---|---|
| Other Authors: | , |
| Format: | Doctoral or Postdoctoral Thesis |
| Language: | Spanish |
| Published: |
Universitat Politècnica de València
2016
|
| Subjects: | |
| Online Access: | http://hdl.handle.net/10251/74645 https://doi.org/10.4995/Thesis/10251/74645 |
| Summary: | [EN] This thesis deals with the study of the sounds produced by beluga whales (Delphinapterus leucas) with a fundamental objective: its characterization and modeling. To this end, analysis / synthesis algorithms of the sounds produced by these animals are proposed. These algorithms are inspired by recent researches on how beluga whales produce the sound and in the physiognomy of the sound production organs. This is a multidisciplinary work, and in order to achieve this goal many fields and topics have to be studied: the generation of sounds in musical instruments, time-frequency analysis techniques along with pattern recognition methods, feature selection, the potential to include algorithms that work in the cepstral domain and a quantitative analysis of the Recurrence Plots. All this allows to propose a sound production model capable to adapt to the peculiarities of this specie and reproduce with high fidelity its wide repertoire of sounds. In addition, in order to validate the proposed model, different sounds from a database of beluga whale vocalisations were analysed. These sounds were compared with those generated with a generic analysis / synthesis model. Furthermore, it is proposed to use the synthesis model parameters for a new sound classification based on how the sounds are produced, achieving better results than those obtained with classifiers based on characteristics of the time-frequency diagram. All the proposed hypotheses have been validated by doing acoustic measurements of beluga whales from the Oceanografic of Valencia (supervised by park biologists), as well as a large number of audio laboratory recordings under controlled conditions. Finally, a passive acoustic monitoring device called SAMARUC design in the framework of the projects related to this thesis is described. The SAMARUC system has the possibility to include different signal processing algorithms for sound analysis in open water environment and is capable of recording high quality sounds. It can also provide the classification of the ... |
|---|