BioAcoustica: A wildlife sound repository with an ear to the future

Biodiversity informatics (BI) applies technology to the collection, analysis, and management of basic species data to better understand how human decisions affect the biosphere. BioAcoustica, a free online open repository and analysis platform, provides BI researchers with a wealth of sound recordings of animals ranging from wild to domestic mammals and birds. 

Understanding an ecosystem requires a lot of time, and multiple observations recorded through all seasons and times of day. With BioAcoustica, anyone in the world can submit and access acoustic samples — an interest in the natural world is all that is required — and distributed analysis is automated to speed data collection and analysis. (Metadata is also available to communities outside the field of bioacoustics.)

The future of BI is in automation of these lengthy processes that use microphones to record sounds and environmental conditions for months on end. The resulting multi-gigabyte audio files will also require strong computational resources and an integrated cyberinfrastructure so that researchers worldwide can quickly run sound files through BioAcoustica algorithms for identification.

“BioAcoustica is a part of a global effort to model the biosphere in a way that we can answer the hard questions we face with a good degree of accuracy and precision,” says Edward Baker, project leader for BioAcoustica at the Natural History Museum, London. “As we move to long term acoustic studies, and as the idea of automated acoustic identification becomes a reality, processing the data becomes a greater challenge.”

Hosted by the Natural History Museum, London, BioAcoustica stores field recordings in a waveform audio or mp3 format in the Scratchpads virtual online biodiversity research environment. For sound analysis, BioAcoustica uses the R statistical language package seewave in the Biodiversity Virtual eLaboratory (BioVeL).

Ultimately, BioAcoustica is striving to make recordings available to as large an audience as possible in both human- and machine-readable formats. When researchers explore images of an audio file in a browser window, they can examine the shape of the waveform and add commentary and labels. By crowdsourcing these annotations and integrating common acoustic analyses that eliminate the need for specialist programs, BioAcoustica greatly expands its reach.

“Our work has a global perspective — the problems facing our ecosystems are universal, and the same techniques can be applied to understanding them throughout the world,” says Baker. “Automated identification is going to create a paradigm shift not only in the amount of data we can collect, but the kind of questions we can answer using it.”

--Lance Farrell