Quantifying group specificity of animal vocalizations without specific sender information

Area of Science:

• Bioacoustics
• Animal Communication
• Marine Mammal Research

Background:

• Automated bioacoustics monitoring and marine mammal studies often lack sender and contextual information in vocalization recordings.
• Distinguishing vocalizations between social groups is crucial for understanding animal behavior and social structures.

Purpose of the Study:

• To develop a framework for estimating group specificity in animal vocalizations without requiring sender identification.
• To introduce and apply the bag-of-calls-and-coefficients approach (BOCCA) for analyzing vocalization ensembles.

Main Methods:

• The bag-of-calls-and-coefficients approach (BOCCA) was developed to analyze ensembles of cepstral coefficients from vocalization signals.
• Relative entropies were computed to compare distributions of coefficient ensembles, revealing group-specific differences.
• BOCCA was applied to vocalization recordings from long-finned pilot whales (Globicephala melas) in northern Norway.

Main Results:

• The study successfully applied BOCCA to identify group-specific vocalization differences in long-finned pilot whales.
• Vocalizations within social groups of pilot whales exhibited significantly lower differences compared to intergroup variations.
• The findings demonstrate the effectiveness of BOCCA in distinguishing between social groups based on their vocalizations.

Conclusions:

• The BOCCA framework provides a robust method for analyzing group specificity in animal vocalizations, even with limited sender information.
• This approach enhances the study of marine mammal bioacoustics and automated monitoring systems.
• The research highlights significant vocal distinctiveness among different social groups of long-finned pilot whales.