Computer scientists develop birdsong identifier that works even in a chorus of tweets
Researchers at the Queen Mary University of London have developed a technology that can identify individual birds present in an area based on their songs, even when many different types of birds are singing at once. This type of information could be a huge benefit to conservation efforts as well as for studying the behavior and language of different bird species.
“Automatic classification of bird sounds is useful when trying to understand how many and what type of birds you might have in one location,” said Dr. Dan Stowell from QMUL’s School of Electronic Engineering and Computer Science and Centre for Digital Music.
The computer scientists designed an automatic analysis technique and classification algorithm to identify many bird species, capable of singling out individual bird species among other birdsongs and other background noise. The researchers used birdsong recordings in the British Sound Library Archive as well as online sources for testing the technology.
The researchers hope that their technology will be used by both professional scientists and everyday bird watchers too and they don't intend to stop at just identifying birds by their calls, they want to learn what those calls mean too.
Dr. Stowell says: “I'm working on techniques that can transcribe all the bird sounds in an audio scene: not just who is talking, but when, in response to whom, and what relationships are reflected in the sound, for example who is dominating the conversation.”
Beyond just understanding birds, the researchers think this technology could also help us to learn more about human language since birdsong and human language have a lot in common. Birds go through the same stages of vocal learning that humans do, so studying those sounds could tell us more about how our own language evolved.