How does Merlin identify a Northern Mocking bird within the first call or two of numerous “mocked” calls. To my ear these calls sound like the original but I’ve never listened to a side-by-side comparison.
Probably, like most AI apps, with unearned confidence and frequent mistakes.
“Or two” is probably the salient factor. I’m pretty sure it can’t reliably do it with one. But if it hears “American Robin…Hermit Thrush” with a predictable pause pattern - Northern Mockingbird it is. Mockingbirds have a stereotypical cadence which I imagine the app can pick out. Or try to pick out. As Darren_Garrison correctly points out, Merlin is far from bulletproof.
It’s going to involve Fourier transforms
So Merlin converts the sound into a picture (a spectrogram) and compares slices of that picture against other slices in its database: Behind the Scenes of Sound ID in Merlin – Macaulay Library
It doesn’t always succeed at this:
(Thread showing how Merlin thinks it’s a mockingbird and several other birds)
This is even discussed in the scholarly literature: Mocking Bird ( Mimus polyglottos ) calls potentially confound acoustic indices of bird diversity and provide a potential heuristic to distinguish them - PubMed
Utilizing the Merlin Bird ID application, we found an average accuracy rate of ~81.3%, with mockingbirds contributing ~31% of false positive identifications. Finding potential solutions for distinguishing mimics in bioacoustic survey data is crucial for enhancing accuracy as researchers increasingly adopt this methodology in the future.
And in an interview with the app developers, they specifically point out mimicry as an issue: https://goldengatebirdalliance.org/blog-posts/the-power-of-suggestion-part-ii-the-wizards-behind-merlin-sound-id/
Q: So we spoke a bit about pattern recognition, and I was curious, what patterns have you noticed or what predictions do you have in regards to occasional Merlin Sound ID challenges to precision or recall?
A: […] Mimicry. Similarly sounding songs, especially if the beginning of a song is similar, the Sound ID will jump the gun and make a prediction based on the first couple seconds […]
So given this…
That’s what the Audubon says too: Are You Listening to a Bird Mimic or the Real Deal? | Audubon
And if you look at the recording samplings for the mockingbirds, the longer ones look pretty different from other bird songs: https://search.macaulaylibrary.org/catalog?taxonCode=normoc&view=grid&mediaType=audio
If you only had the first few seconds, maybe you just got lucky that time? But if you were able to record longer, as it switches between several birds, maybe the accuracy goes up eventually…?