Merlin app

[Wowbagger]

This is an app designed to recognise birdsong. I downloaded it few days ago on the recommendation of My Mate Terry Who Art in Sibton and have been giving it a trial run.

It seems to be very good at identifying every high-pitched squeak and attributing it to a particular species. Yesterday it immediately spotted a long-tailed tit, which I hadn't heard, but of which there are often loads in the park, plus the usual stuff I could hear: robins, blackbirds, blackcaps, great tits, blue tits, wrens, house sparrows, chiffchaffs, goldfinches, chaffinches. It seemed to have a bit more difficulty with the caw of the carrion crow and the quack of the mallard, but got them eventually. It completely missed what, to me, is the most obvious cry, the yaffle of the green woodpecker. To my surprise, it picked up a song thrush yesterday. Not impossible, but it's at least a year since I saw or heard one in the park.

Today, it claimed to have heard a grey wagtail (spelt the USAnian way because this app has been produced by Cornell University), and a greenfinch. I'm not that surprised by the grey wagtail because I was walking pretty close to the brook, where I have occasionally seen them in the past. Greenfinches used to be very numerous but I've not seen one for a long time. There's some horrible disease doing the rounds which has all but wiped them out in some places. 20 years ago there used to be massive winter roosts of them - several hundred birds - in our local park.

But it really let itself down when it solemnly informed me it had heard a great crested grebe. In 30+ years of visiting that park, I have never seen a great crested grebe there, and it's a very obvious bird when you see them. They swim in open water and frequently dive, and you can see them from a long way away. I'm 100% certain that there was no great crested grebe anywhere in the vicinity. There was, however, a squabble amongst some mallards, including a mother with two ducklings, so I suspect one of them made a grebish noise that the device picked up.

But I'd give it 8/10 for its usefulness.

[Sergeant Pluck]

I use one called BirdNet. It has similar performance to Merlin based on what you’ve said.

I see it is also developed by Cornell. I wonder what the difference is. Does Merlin need a data connection? BirdNet does - you record, snip a sample, and send it off for analysis.

[Mrs Pingu]

I use Birdnet which is linked to Cornell somehow too. It also saves what you send for analysis so if you hear something pretty exciting you still have a record of it.

[hbunnet]

I use Birdnet which is linked to Cornell somehow too. It also saves what you send for analysis so if you hear something pretty exciting you still have a record of it.

Jings, I've just found 80+ records on Birdnet, I hadn't realised they were saved. It shows a map of the location too.
Thanks for that. I find the application very useful when the leaves are on the trees.

[lissotriton]

Maybe a better microphone would help? A little shotgun mic would make it more directional. Or a parabolic reflector.

[Pingu]

This is an app designed to recognise birdsong. I downloaded it few days ago on the recommendation of My Mate Terry Who Art in Sibton and have been giving it a trial run.

It seems to be very good at identifying every high-pitched squeak and attributing it to a particular species. Yesterday it immediately spotted a long-tailed tit, which I hadn't heard, but of which there are often loads in the park, plus the usual stuff I could hear: robins, blackbirds, blackcaps, great tits, blue tits, wrens, house sparrows, chiffchaffs, goldfinches, chaffinches. It seemed to have a bit more difficulty with the caw of the carrion crow and the quack of the mallard, but got them eventually. It completely missed what, to me, is the most obvious cry, the yaffle of the green woodpecker. To my surprise, it picked up a song thrush yesterday. Not impossible, but it's at least a year since I saw or heard one in the park.

Today, it claimed to have heard a grey wagtail (spelt the USAnian way because this app has been produced by Cornell University), and a greenfinch. I'm not that surprised by the grey wagtail because I was walking pretty close to the brook, where I have occasionally seen them in the past. Greenfinches used to be very numerous but I've not seen one for a long time. There's some horrible disease doing the rounds which has all but wiped them out in some places. 20 years ago there used to be massive winter roosts of them - several hundred birds - in our local park.

But it really let itself down when it solemnly informed me it had heard a great crested grebe. In 30+ years of visiting that park, I have never seen a great crested grebe there, and it's a very obvious bird when you see them. They swim in open water and frequently dive, and you can see them from a long way away. I'm 100% certain that there was no great crested grebe anywhere in the vicinity. There was, however, a squabble amongst some mallards, including a mother with two ducklings, so I suspect one of them made a grebish noise that the device picked up.

But I'd give it 8/10 for its usefulness.

The Birdnet app gives a confidence rating of its findings. Does Merlin?

[Wowbagger]

It doesn't seem to, although it did say something about returning results to Cornell to improve the product, but I couldn't see how to do this either.