Once upon a time, O younger readers, if you wanted to know the title and artist of a new song you heard on the radio, you had to hope and pray that, when the music came to an end, the DJ would remember to tell you. They usually didn’t, which meant your only option was to visit the nearest record shop and hum the tune to a bemused assistant.
Now, of course, you simply take out your smartphone, open the Shazam app, and within seconds you can find out if you’re listening to the Shires or Cher, Bach or the Beatles, Taylor Swift or Meatloaf.
Likewise, until recently, if you wanted to tell a singing robin from a trilling wren, you needed a lifetime’s ornithological knowledge. But now, like buses, two new birdsong apps have come along at once. And both claim they can help us work out which species is serenading us from deep inside the shrubbery.
Chirpomatic and Warblr use the same sort of technology as Shazam. But instead of holding your smartphone up to the radio, you just have to get close enough to a bird to record a 10-second snatch of its song.
This is not as easy as it sounds, as I discovered when I tested them out in my Somerset garden. Just as I pressed the record button, the bird would fall silent; or a gust of wind would blow; or one of my children would ask why I was holding my phone up against a hedge.
But eventually I did manage to record enough material for the analysis to begin. And that’s where the trouble started.
Chirpomatic helpfully suggests three likely candidates, and then two more possible alternatives. I say helpfully, but when the choices bring up combinations such as wood pigeon, greenfinch and great tit (when the actual songster was a chaffinch), it isn’t terribly useful.
At present, despite being marketed as including the birds of north-west Europe, Chirpomatic mainly covers birds of parks, gardens and woodland, which excludes several species on which I wanted to test it, such as reed warbler. However I’m sure that this will rapidly improve as the software gets more sophisticated.
One problem for the app designers is that although individual recordings of pop songs are exactly the same, bird songs are not – even the same individual will vary its song each time it opens its beak; while chaffinches, for instance, have different accents in different parts of the country.
Warblr – developed by a team from Queen Mary University of London – claims to be able to identify 95% of bird songs accurately, under optimum conditions. While the same caveats apply as to Chirpomatic – you do need to be fairly close to the bird, to minimise interference – the results were better.
As a bonus, the teams behind both apps collect data and recordings from successful identifications, which they then pass on to researchers and conservationists.
This kind of app is still in its infancy, so there’s certainly room for improvement. And they could be a useful tool for beginners. But this raises an important question. Do we really want to know exactly which bird we’re listening to, without having to make any real effort?
There is a lot to be said for the long – and admittedly challenging – process of learning each species’ song and calls, by a process of trial and error. When you do finally manage to identify the songster, you feel a real sense of achievement. You also get to know the bird itself, which leads to a better understanding of its lifecycle.
So what’s next? Suppose you could look through your binoculars, press a button, and discover exactly which species you are looking at? Reducing birding to a kind of organic trainspotting would, in my view, take all the fun out of it – learning from your mistakes is what spurs you on to gain a genuine understanding of the natural world.
And sometimes it’s better not to try to identify every bird you hear. The wildlife sound recordist Gary Moore has described listening to the dawn chorus as ‘audio yoga’ – and he’s right. Trying to pick out each individual songster is not only difficult, but can be pointless – just sit back and enjoy the greatest free show on earth.
• This article was amended on 26 August 2015. An earlier version said Chirpomatic suggested the likeliest candidate together with three possible alternatives; in fact it suggests three likely candidates and then two more alternatives. References to Chirpomatic misidentifying some birds was removed, as it does not suggest one likeliest candidate, and replaced by a section about the birds the app covers. A reference to the results from a rival app being “far better” than those from Chirpomatic was changed to say they were better.
