BirdNET: Tips, Tricks, and Insights for Ecoacoustic Analysis

About the Event

BirdNET is an open-source AI-powered system designed to automatically identify bird species from audio recordings. By automating the processing of large-scale environmental sound data, BirdNET enables researchers to efficiently assess bird community composition and monitor biodiversity. However, the automatic identifications that BirdNET produces can vary in accuracy, and that accuracy may not directly correlate to the associated confidence score that BirdNET generates. How much manual call verification should a user rely on? Can thresholds levels in confidence scores be set that remove the need for in-depth manual call verifications? What are some of the tips related to using BirdNET on large datasets?

It is also important to consider technical aspects such as consistent file naming (including site information), standardising terms and spellings and ensuring robust metadata practices are used. Join us to learn best practices and avoid common pitfalls in managing your BirdNET data and bring your question along as there will be plenty of Q&A time.

Whether you’re new to ecoacoustics or an experienced researcher, this session will offer valuable insights into the future of sound-based environmental monitoring.

Recording

A recording of this session is now available:

Speakers

• Philip Eichinski, QUT
• Dr Nina Scarpelli, CSIRO and Museums Victoria
• Callan Alexander, QUT

About the 2025 Open Ecoacoustics Webinar Series

The Open Ecoacoustics Webinar Series will feature one-hour sessions designed to engage and connect the Australian ecoacoustic community. This series offers a platform for exploring compelling and thought-provoking topics related to ecoacoustics while fostering ongoing collaboration among community members. Each webinar will include structured segments such as expert panels and lightning talks, followed by an open discussion to encourage audience interaction and questions.

Who should attend?

People using or interested in using acoustic monitoring, and curious to hear more about current projects including:

• Researchers with large datasets needing storage and analysis tools for ecoacoustic research
• Ecologists working with acoustics
• Government agencies with conservation goals
• Research Software Engineers
• Researchers who coach and mentor peers and colleagues
• Higher Degree Research Supervisors
• HDR Students

What will participants gain from the session?

• Learn from experts about data analysis techniques for ecoacoustic research
• Connect with other people working with ecoacoustic data
• Gain insights into the value proposition for ecologists, including practical applications and case studies

To keep up-to-date with Open Ecoacoustics, subscribe to the newsletter

Will the session be recorded?

Please note that this event will be recorded and published by the ARDC. This may include your contributions during the session. ARDC respects the privacy of individuals. Information collected is in accordance with the ARDC Privacy Policy.

Have questions?

Email [email protected]

Please note that parts of this event will be recorded and shared by the ARDC. ARDC respects the privacy of individuals. Information collected is in accordance with the ARDC Privacy Policy.