Record a bird singing and use Automatic Recognition to help identify it. Unlike all other apps, no internet or mobile connection is needed.
• The Times Newspaper said the app “Does what it says on the
tin” and were amazed that “this astonishing piece of software
costs about the same as a cup of coffee”.
• “In direct comparison Bird Song Id showed better results than Chirpomatic”. – naturschutz.ch – Swiss Birding Magazine
Copied by a few, bettered by none, Bird Song Id is three years
ahead of anyone else in this field !
• The Telegraph Newspaper said “Bird Song Id is a good walk improved, and may even put a spring in your step”.
• over 1000 good ratings
Automatic Recognition takes place on the device, no connection to the internet or mobile network is needed.
Automatic Recognition needs good quality recordings to work. See the HELP below, or follow the Developer Website link to find out how to improve your recordings.
Automatic Recognition will help identify a good cross section of bird song in the British Isles, but you will need a good quality recording. Lots of blue and green peaks reaching to the top of the analysis graph is a good indication of a quality recording. Help and examples are included in the app to guide you.
How successful is it ? Our aim is to get the correct bird in the top three of those listed. In our testing of over 1,000 samples we found the success rate to be 85%.
Follow the Developer Website link to see a demo of Automatic Recognition on YouTube
————— HELP —————-
Automatic Recognition needs a good quality recording of the bird.
Only record when a single bird is singing, or when a single bird is dominant
The analysis graph should have plenty of blue and green colors with few or no black peaks.
The peaks on the graph should reach all the way to the top. Low peaks indicate poor quality and recognition is unlikely.
The Audio Quality, shown at the top of the results screen, should be at least Good, and preferably Very Good or Excellent
The highest score should be above 70% to be useful. Below this and the system is telling you it is not sure what the song is.
A difference between two birds of less than 10% is not that significant, so listen to the samples from both to help you further
See the samples for how the analysis graph should look (to see the samples press the sample button at the top right of the help page)
Make sure it is very quiet apart from the bird song which should be nice and loud
Be as close to the bird as you can be without disturbing it
Be patient, dont expect miracles, the system works but needs good clear recordings
The microphone is at the bottom of an iPhone, point the bottom of the phone towards the bird
Remove the case from your iPhone if you use one as this can dampen the microphone
Record a single bird at a time
Capture at least two singing episodes if you can.
Record for the full 30 seconds
Avoid windy conditions. Air moving over the microphone will mask the bird sounds
Make sure there is plenty of life in the battery. Your device will reduce power going to the microphone as the battery runs down
Playing songs to the app from recorded material on the internet, other apps or CDs is unlikely to work as this material is digitally processed and edited (inter phase gaps are often shortened). They sound very different to recording birds for real on an iPhone or iPad in the natural environment. Its a bit like putting rocket fuel in a lawn mower.
Version iOS-App 8.9
Here is what’s new and what we’ve done for you:
– We added more songs and calls
– improved automatic recognition!
– new design, new icon
– more bird images
– see where other users have recorded birds in your region