[Tweeters] Merlin App fallibility

[Gary Bletsch via Tweeters]

Dear Tweeters,
Kudos to Kevin Lucas for bringing this up again. Unfortunately, Tweeters is probably no longer reaches as big a part of the birding community as formerly, so this discussion may not reach as many birders as it ought to.
Here in Western New York, I was amazed to hear a Rufous-sided Towhee the other day. I searched and searched for it, but could not see it, although the "chewink" call was reaching my ears loud and clear from somewhere across the street. It's rather early for towhees here right now; they desert the area when winter sets in, and it's still quite wintery here.
I pulled out the mobile phone and tried the Merlin App's "Sound ID" function, which immediately ID'd the bird as an Eastern Towhee. "Nice to have some corroboration," I thought. Imagine my surprise when the calling bird popped out of a bush and called one more time, before flying off--it was a Common Grackle! I don't remember hearing mimicry from a grackle before, so it was probably just an individual with an oddball call note.
Just for fun, I tried the Merlin App a bunch of times in Namibia recently. The "Sound ID" function was remarkably unsound, identifying zero point zero percent of the bird vocalizations correctly. After about 20 tries, I gave up. I had had hopes that it might be of use with some of the tricky vocalizatons, such as those of the cisticolas. At one point, the app suggested both Pileated Woodpecker and Chiguanco Thrush for one call. Hmm, a North American Woodpecker and a South American Thrush calling away in southwestern Africa!
Yours truly,
Gary Bletsch

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman11.u.washington.edu/pipermail/tweeters/attachments/20240404/9dc04c00/attachment.html>