Digital Bird Song Identification a Reality?
by Nate Swick
Even since the introduction of the smartphone app Shazam, which, when held up to nearly any piece of recorded music can quickly and seemingly miraculously name the song and artist, techy birders have been clammoring for something similar for bird vocalizations. Imagine a computer program that could identify a bird in the field simply by analyzing a short recording. It would revolutionize birding in a way not seen since the advent of the binocular.
For years such a device was prohibitively difficult to implement for fairly obvious reasons. Birds, after all, have myriad vocalizations, and while recorded music is predictable and repeatable, a singing bird is practically anything but. But Mark Berres, an ornithologist at the University of Wisconsin in Madison, thinks he's finally cracked the code with WeBIRD, the Wisconsin Electronic Bird Identification Resource Database.
Barres plans to have WeBIRD up as a free app in the spring of 2012, but the questions as to what this means for birding are obviously huge. They pertain both to the limits of this technology (does the bird need to be singing loudly, or nearby?) and the ethics of using it (heard birds are obviously allowed on ABA lists, but what about birds identified by a computer?).
I'm certainly curious to put it through its paces and, at very least, to see if the thing blows up when faced with the spring's first American Redstart.