211–220 of 221 posts
I have wondered if perhaps with 3 microphones positioned a specific distance apart from one another, an app could compare their relative volumes and show the user the exact location of a bird. Any smart HNers know if something like this is possible?
I tried using it in New Zealand last year, and it wasn't as effective as in the US: I think it hasn't been trained as well on the native New Zealand birds, many of which aren't found anywhere else.
Amusingly, it identified turkeys when we were in New Zealand, which I was irritated with because there clearly weren't any turkeys in New Zealand. It turned out I was wrong when we came across a flock of them in the Waikato area running around in a sheep field. A local told me that they were brought there around a hundred years ago and are mostly left alone because nobody eats turkey in New Zealand.
I'm friends with some of the researchers on the Sound ID portion of this app! The team's gone to great lengths to make sure the machine learning models and evals are solid. Under the hood, Sound ID is a great example of how "domain-expert-driven" careful research can give more reliable results than just feeding in data and hoping for the best.
The iOS app did seem to lock up when trying to download the region packs though. I should find out where to file a bug!
As an amateur bird photographer I love this app, but I wish there was some way to pinpoint the direction the sound is coming from. I have wondered if perhaps with 3 microphones positioned a specific distance apart from one another, an app could compare their relative volumes and show the user the exact location of a bird. Any smart HNers know if something like this is possible?
I am no expert, but I do know that this is possible. It's called acoustic location [1].
> Civilian uses include locating wildlife and locating the shooting position of a firearm.
As to how spaced out the microphones would have to be and to what accuracy the location can be determined, I cannot say (but my guess would be "pretty accurate").
Another interesting challenge would be to determine the location of multiple birds in different locations. Perhaps isolating specific bird calls with some ML model and then acoustic-locating that specific call. I can imagine the application showing which bird is calling from where in 3D.
[dead]
This app changed my life. I got into hiking and outdoor spaces when I married my wife and she wanted me to join her in camping, something I never did as a kid. This iteratively got me more into natural spaces and wildlife, but only as a passing curiosity. But then one morning while camping, I woke up to a symphony of birdsong. Many calls, but there were at least dozens of a call echoing throughout the forest of a flu…
I love this app. I use it all the time. But I do have a complaint, that you can't upload an image via the web from your PC. For those of us who use a DSLR for our birding photos, the UX to have to transfer images to our phone to send them through the app is really painful. It isn't the end of the world. It just makes me ID my birds in other ways. But it would be nice to provide non-mobile UX options.
I'm the author of the Birder framework (a computer vision toolkit for bird classification) - https://gitlab.com/birder/birder It's still in early alpha, but you might find it useful for your DSLR workflow. If you're interested in trying a web-based alternative, I have a demo space at https://huggingface.co/spaces/birder-project/birder-image-cl... where you can upload images directly from your browser. Fair warning th…
my use case is adding tags to my Lightroom library with identified species.
I really like this app, but last night I had an encounter in the dark with the most weird alien sounding bird, that uses so low frequencies that the app couldnt make anything out of it. I realize this is a fault of my phones mic and not the app, btw! I think the bird was one of the snipes (dvärgbeckasin?) because I met them before.
Did you actually see the bird? Some animals make sounds that are deceptively similar to birds (ex chipmunk chirps). Merlin won’t recognize these as “not a bird”; it just doesn’t recognize the sound at all. Usual disclaimer IANABL (I am not a bird lawyer)
We don't have a huge load of animals making weird sounds around here, so one learns to recognize most of the strange ones after some years in the woods.
Of interest: > Sound ID is trained on audio recordings that are first converted to visual representations (spectrograms), then analyzed using computer vision tools similar to those that power Photo ID. Yeah, the spectrogram scrolling by at the top isn't just a cute gimmick, that's actually how the recognition works...
Neat. When working in a linguistics lab as an undergraduate long ago, we looked at spectrograms to identify sounds (specifically places of articulation) as much as listened to recordings. So it makes some sense to build a model on them rather than some other representation of the sound.