It’s an almost universal feeling: the thrill of hearing a mysterious new bird song. And that’s usually followed by a question: what was this bird?
The question has become much easier to answer. The free Merlin Bird ID app from the Cornell Lab of Ornithology can now identify bird sounds. Merlin can recognize the sounds of over 400 species from the United States and Canada, and that number is expected to increase rapidly in future updates.
As Merlin listens, he uses artificial intelligence (AI) technology to identify each species, displaying a real-time list and photos of birds singing or calling.
Automatic song identification has been a dream for decades, but sound analysis has always been extremely difficult. The breakthrough came when researchers, including Grant Van Horn, Merlin’s principal investigator, began to treat sounds like images and apply new and powerful image classification algorithms like the ones that already power the identification function. photo of Merlin.
“Every sound recording made by a user is converted from a waveform to a spectrogram – a way to visualize the amplitude [volume], frequency [pitch] and the duration of the sound, â€said Van Horn. “So just like Merlin can identify an image of a bird, he can now use that image of a bird’s sound to make an identification.”
Merlin’s pioneering approach to sound identification is fueled by tens of thousands of citizen scientists who have contributed their bird sightings and sound recordings to eBird, the Cornell Lab’s global database.
“Thousands of sound recordings train Merlin to recognize each species of bird, and over a billion bird sightings in eBird tell Merlin which birds are likely to be present in a particular place and time. Said Drew Weber, coordinator of the Merlin project. “Having this incredibly robust bird dataset – and integrating it with faster and more powerful machine learning tools – allows Merlin to identify birds by sound now, when it seemed like a challenge from size just a few years ago. “
This pioneering sound identification technology is integrated into the existing Merlin Bird ID application, which means that Merlin now offers four ways to identify a bird: by sound, by photo, by answering five questions about a bird that you’ve seen, or by exploring a list of expected birds in the user’s location.
Merlin helps identify the sounds of individual birds even when several birds are singing at the same time. Once a user has recorded the sounds of the bird, they can select a species and return to the location in the recording where their song or call occurred. The recordings are saved automatically so that users can listen to and watch the sounds again, making Merlin a great learning tool.
With one-touch access, users can also learn more about each bird with identification tips, maps and over 80,000 photos and sounds from Cornell Lab’s Macaulay Library.
“The Merlin app really opens up a whole new world of sound,†said Jessie Barry of Cornell Lab, whose team led the project. “It helps everyone solve the mysterious birds they hear around them, and the technology that powers Merlin’s sound identification can also be used for research and conservation, opening up new possibilities for how scientists can monitor, study and protect birds. “
Merlin Sound ID marks a big step forward in people’s ability to connect and understand the sounds of the natural world around them, Barry said.
Macaulay Library web designer Matt Schloss, who describes himself as an advanced newbie and has beta tested the app, agrees.
“Merlin helped me find birds that I might have overlooked before. It actually makes me feel like I have super power, or at least enhanced skills, â€Schloss said. “I really think this will change the way people fly.”
The Merlin Bird ID app with the new Sound ID function is available for free on iOS and Android devices. Click here to download the Merlin Bird ID app and follow the instructions. If Merlin is already installed on your phone, tap “Get sound id”.
Marc Devokaitis is associate editor at the Cornell Lab of Ornithology.