Humans are still the best systems in recognizing acoustic patterns. For example, volunteers perform much better than machines in associating vocalizations to the corresponding species, by listening to the recordings obtained by placing sensors into the environment. But can a human being listen to hundreds of hours of wildlife recordings and annotate all the sounds she/he hear? Of course not – machines have been invented for that. But if you couple both approaches, you could get a powerful tool: a crowdsourcing technique to detect large numbers of animal voices, as the one developed by a team of australian researchers. It is not just a weekend hobby: recognizing and counting species tells a lot about the biodiversity of an ecosystem and how it is evolving. Birdwatchers, are you listening?
x
