Orcasound app: An open-source solution for streaming live ocean sound to citizen scientists and cloud-based algorithms

Abstract

A growing number of hydrophones stream ocean sound data in near-real-time to automated detectors. Many of these streams are also provided to the public, but none yet appeal for citizen scientists to provide real-time detection or classification services. Within the Orcasound hydrophone network—orcasound.net (WA, USA)—live human listeners have always been in friendly competition with real-time automated detectors. To facilitate the participation of citizen scientists in studies of signals from southern resident killer whales and other soniferous organisms of the Salish Sea, in 2017–2018, we developed a low-cost combination of hardware and open-source software to stream audio data from our hydrophones to the headphones of our listeners. The data are archived at and stream from AWS/S3 in real-time (<60 seconds latency) to a browser-based player that is platform- and device-independent. This sets the stage for human detection of signals—mainly the calls, whistles, or clicks of fish-eating orcas—but also novel signals for which we do not yet have classifiers. We will demonstrate the Orcasound app and explain how it interacts synergistically with automated detectors and classifiers, as well as crowdsourced citizen science projects.