NEAL: an open-source tool for audio annotation.

Area of Science:

• Ecology
• Bioacoustics
• Conservation Technology

Background:

• Passive acoustic monitoring (PAM) is crucial for ecological research, generating vast audio datasets.
• AI and machine learning models require high-quality expert annotations for training, which are often resource-intensive.
• Efficient and cost-effective audio annotation methods are needed to maximize the utility of PAM data.

Purpose of the Study:

• To introduce NEAL (Nature+Energy Audio Labeller), an open-source interactive tool for annotating audio data.
• To provide a user-friendly platform for both expert and citizen scientists to contribute to acoustic annotation projects.
• To demonstrate the adaptability of the tool for various bioacoustics and general audio labeling tasks.

Main Methods:

• Development of an interactive audio data annotation tool using R and the Shiny framework.
• Implementation of a reactive user interface that adjusts automatically based on user settings.
• Application of the tool for labeling acoustic data collected from wind farm sites.

Main Results:

• NEAL offers a cost-effective and efficient solution for annotating large audio datasets.
• The tool facilitates collaboration between domain experts and citizen scientists in acoustic data analysis.
• The R and Shiny framework allows for flexible modification of NEAL for diverse bioacoustic applications.

Conclusions:

• NEAL enhances the process of audio data annotation for ecological studies, particularly those using AI.
• The open-source nature and flexibility of NEAL promote wider adoption and adaptation in bioacoustics research.
• This tool supports the efficient processing of acoustic data for biodiversity and conservation efforts.