The tools we have for birding have increased and improved since the old days when ornithologists had to shoot birds and hold them in their hands to get a positive identification. The invention of binoculars available to the masses was the first major advance. The field guide was next.
Since then, countless internet resources have been invaluable. But nothing in decades has had the impact that the creation of Merlin has. For those not familiar with Merlin, it’s an app for your phone that identifies the birds singing around you.
You can download Merlin for free, walk out in your backyard, hit the “Sound ID” button, and you’re off. I stepped out onto my deck as I was writing this and picked up Canada goose, American crow, black-capped chickadee and American goldfinch in 30 seconds. While listening, a sound spectrogram scrolls across the top of the display showing the time and frequency structure of the sounds being recorded.
I’ll cover more of the history and mechanics of Merlin in a bit. But my main point is that this tool — another masterpiece from the Cornell Lab of Ornithology — has opened up birding to a whole slew of new people. I tell friends of friends about Merlin.
I tell people who write to me with bird questions and observations. And I tell people I run into in Boise parks who stop to chat. By carrying binoculars and a camera, I meet a lot of curious people.
It has amazed me how many have downloaded the app and then reported to me about what birds they’ve found. Not too many years ago, the technology required to identify bird songs was not very good. I remember trying a variety of different apps and basically walking away because they couldn’t identify common songs.
They also tended to “identify” species that weren’t there. As a reasonably good song identifier myself, I could pretty quickly tell whether or not the apps were any good. One of the problems Cornell had to deal with was the issue of overlapping songs.
The human brain quickly and naturally separates aural signals coming in. We don’t even think about how we can separate a robin’s song, a child talking, a car alarm and a squeaky swing from each other. And even when multiple birds are singing at once, within similar frequency ranges — a robin, song sparrow, house finch and chickadee, for example — we readily distinguish them.
But the early technology could not do this very successfully. The algorithms behind the scenes could not match the bird you were hearing to the bird songs in the database. Obvious songs were passed over and misidentifications were frequent.
Now, Merlin can pick out each individual song, typically within a few seconds, if the songs are loud and clear. The Merlin Sound ID project began only in 2020. Merlin uses recordings from the Macaulay Library at Cornell to train computer models to recognize bird vocalizations based on the sound spectrograms you can see yourself when Sound ID is running.
The computer tools are actually “vision tools” similar to those that power photo identification software. So, it’s not the sound itself, but rather the picture of the sound computers work with. The accuracy of the process is continually improved by experts at Cornell.
Of course, having Merlin tell you what birds are singing around you is different from finding those birds and looking at them. A friend recently texted me about how Merlin had identified a Bewick’s wren singing in the shrubs nearby, but she couldn’t find the bird. This is typical feedback I’ve gotten from many users who are new to the birding world.
Yes, finding the bird is another step. Then, people might learn to identify the song without Merlin. Further, birds don’t always sing, so identifying birds by sight might be next.
But all of these advances aren’t required. Just learning what birds are around is a good beginning. My hope isn’t necessarily that people will become better and better birders, but rather that they will take steps to conserve birds by being more in tune with the natural world.
A bit more on the app itself. Once you have recorded a sound segment, it is stored for you. Just click on “My Sound Recordings” to see a list of all the files you’ve recorded.
When you click on one of them, you get a number of options. At the top of the screen is the standard “share” icon. If I get a good recording of something, I will email it to myself so I can add it to my eBird checklist for that particular outing.
Once you’ve sent the sound file to your email, you can download it to your computer and then upload it to the eBird checklist, along with photos and/or videos, if you also got some of those. These media files help Cornell continually refine what is known about each species, and they provide raw material for research projects, some of which have not yet been conceived. Like money in the bank, you don’t always know what it will be used for when you save it.
But why stop with Merlin? There are now a number of autonomous recording units (ARUs) available as well. Like a trail cam, you put this unit out in the woods, let it run for a few days, and then bring it in to see what sounds have been recorded. Some ARUs are designed to record bats, picking up those uber-high frequencies we can’t hear.
As with trail cams, there are many vendors and units have a variety of features and options. Some old-time birders bemoan the reliance on technology by today’s bird enthusiasts. The use of Merlin and ARUs is seen as cheating.
But I totally disagree. Anything that gets people more engaged in the amazing and beautiful world of birds is OK by me. Get Merlin, get trail cams, get ARUs, use them, and discover.
Your life will be better!.
