The Merlin app by Cornell Lab of Ornithology is fantastic for identifying bird sounds in real-time. Besides the machine-learning tech the overall app is a great experience, clearly made by talented developers who probably also love birds and dogfood the product. Every thing about it is tailored to quickly giving you all the information you need to look up what kinds of birds to expect in different areas at different times of year and quickly narrow down possible candidates.
I've been a user of BirdNet (also from Cornell, but apparently a different group?), looking forward to checking out Merlin. With Birdnet you have to record a sound, pick out the specific portion of the audio and click identify.
One could say BirdNet is the "pro" tool. Merlin uses much of the same sound database and AI but as you've seen - is automated - set it recording and it will mark every bird sound it can hear and identify.
Another good one is SongSleuth[1] - recently recommended to me by Nathan Pieplow because it has a "pre-record" buffer so you can hit record as or immediately after you hear a bird and it captures a few seconds before that instead of having to run Merlin continuously waiting for the bird to call next.
I've tried setting it up with the birdnetlib package but had trouble getting it running with a model and pack that was for my region. The default model for birdnetlib is in Europe.
Merlin is automatic and realtime. There does currently seem to be a bug with recordings. It will highlight the wrong bird in the set during playback even though it was correct when the live recording was made.
Also love Merlin app and my life since finding it. A few years ago I kept hearing a beautiful and haunting bird song on hikes but after asking other hikers nobody could tell me what bird it was!
Enter Merlin app and it picked it up easily as a Swainson's Thrush so I am forever grateful to the technology.
Caution for folks learning about the Merlin app from this thread: it's good at identifying bird calls, especially in North America, where there's a lot of training data, but it's not conclusive and can sometimes lead you astray by misidentifying a common bird as something exotic, or getting confused by birds that mimic other species (like mockingbirds). Treat Merlin Sound ID's suggestions as just suggestions, and use multiple inputs to identify a bird - appearance, structure, plumage, time of year, time of day, behavior, habitat for example.
Yes, solid advice. Generally I find if I'm fairly close to the sound source, and it's repeating, it's almost always accurate, to a surprising degree (when I can confirm by sighting it); if it's a one-off call from farther away, especially of a rarer bird, it's often not. Merlin helps by printing a red circle next to rare candidates for your time/location, and I've learned to treat those with suspicion. I've definitely seen it get confused - once it kept pegging a Mallard when I was nowhere near water, and I finally figured out it was interpreting the jostling sound of lens caps in my bag as I walked that apparently resembled a quack!
Would you say more about mockingbirds? Does Merlin tend to output mockingbird when it’s really a bird that a mockingbird imitates, or output a non-mockingbird when it’s actually a mockingbird imitating that non-mockingbird? (I found an amazing-sounding bird and downloaded Merlin on the spot to id it, and it output mockingbird, that’s why I am especially curious. It was a dark night so I could not visually id at all)
Merlin's pretty at good at identifying mockingbirds by sound alone but once in a while it outputs a bird that's not even expected in your area. It rarely makes the reverse error (identifying the call of a non-mockingbird as the call of a mockingbird) in my experience!
BirdNET seems to handle mockingbirds pretty well in my experience. I've tried it a few times when I know what they're mimicking, and it popped up as a mockingbird.
What I don't see mentioned in the article is that most birds seem to live on a much faster timescale than humans. If you record and slow the song down, you can hear far more nuance in it.
BTW the Merlin app mentioned in the article is fantastic. Having it for birdwatching is like having a GPS map in the city, especially when paired with a good directional microphone. It only tries to identify a bird itself, though, without attempting to decode the type of the song.
I know enough when some birds are upset. Crows are usually the most vocal. I'll try the test later but it's funny how well we humans with practice know. Except pigeons they are stoic.
My cat was sitting quietly one day and crows made noise and my cat uncharacteristically growled a deep growl. She know something was up.
I have a paregrene falcon that sometimes comes to hang out in a tree in my backyard, grackles and bluejays will follow it around making noise and harassing it. The sound is unmistakable once you've heard it a few times.
Bluejays themselves are known for their mobbing behaviour towards predators.
Some related anecdotes about listening to birds and sensing the rhythms nature:
Out my window while I work I watch blackbirds harass crows over a field where the blackbirds nest. They don't show up until later in the morning, and I can tell I've slept in way too late if I hear crows. I can tell the sun isn't up yet if I can still hear bullfrogs, and at night I know to grab a blanket when the crickets chirps slow down, or to close the windows when the cricket chirps are dense, because it's probably cooler inside than out and I can preserve the cool in the house if it's a hot night. The robins seem to come out looking for worms before a storm, because the worms likely sense the pressure change and surface before hand.
I had a squirrel problem over the winter, and I progressively started spreading a few peanuts out over the snow in the field further and further from the house and where they had less access to retreat to cover, as the hawks would show up later in the day where their shadows would be cast further out from their prey on the ground. I got the squirrels venturning pretty far out into the field for a while. I never saw the hawks get one, but one day I stopped hearing the squirrel and seeing it around. Not sure it worked, but it seemed like the most fair solution.
I'm consiering a bat house to keep the mosquitos down as well. If you have the peace and quiet and time to listen, you can really integrate with nature pretty well.
The angry/warning caw of a crow is sometimes called scolding, I think. It has a wide range of intensity. There were many American crows around my last house. The degree to which they cawed, and in particular started up a chorus, was proportional to the threat. On my normal routine, I would get a few caws, and then they'd settle. If I diverged from my routine and for example, walked off the path, or if I had others with me, it would get a bigger response. And when I let the dog out it was just a flood of angry cawing.
Crows are very stubborn in their suspicion. More than most birds. Jays and Oak Titmouses (that looks wrong) have definitely warmed to us at home, since we feed them and the Crows. The Crows won't trust us at all, even after years.
Later in the day a crow was outside. It "cawed" and was quite I could hear another crow far away doing the same. So the crow near me "cawed" again and then stopped and listened again. It was obvious it was communicating.
This crow Broken Wing as I call it due to a drooping wing is a local crow. A family of five lives in a nearby tree it's a parent.
I love the merlin app by cornell, but some reason, while it can recognise some pretty exotic birds, It didn't recognise the common crow cawing right over my head.
I know it has the recognition, because the other app (bird.net) recognised the sound as a crow, but it just didn't have it the region pack for my country.
Nathan Pieplow[1] wrote the recent The Peterson Field Guide to Bird Sounds series[2]. He delves deeply into the spectrograms of bird songs and calls and describes and names the different features thereof. He offers the basics the basics on his site3[3]. I got to hear him speak recently - pretty interesting stuff.
I can not only understand some species of birds, but they can understand me.
Developed this habit during the COVID-19 lockdows, and I happen to have blackbirds and ducks as neighbours. Was listening to them for some months, and finally decided to chirp in.
I don't give them food but they are truly curious of me. The blackbirds stare at me in silence before breaking into the most difficult sequence in their pocket - and laugh at me when I try to versionate.
The mature ducks come to me when I say qvuaak, and the chicks when I do the high "blivblivlinli" -sound.
I don't feed them or want them to get too attached to humans, but it has been funny experiment.
I have noticed that, too! I am still up and coming in my knowledge of identifying birds by sight and sound, for which the latter I use the Merlin app. There's been several times where I've been like "ooh, what's that new sound/bird?", only to find out it's one of the local (perhaps to my yard) robins.
Being from rural India, I have heard birds make similar noises and have successfully avoided walking into a snake or two. So that was the easiest for me honestly.
http://web.archive.org/web/20230723151843/https://www.nytime...