canberrabirds

Sort of related - bird call ID apps

To: 'CanberraBirds' <>
Subject: Sort of related - bird call ID apps
From: Philip Veerman <>
Date: Mon, 27 Nov 2017 23:53:49 +0000

I reckon anyone who can actually produce something along these lines, that actually will work to any useful extent, as in be mostly correct, will do something amazing. If they can do so for any kind of financial reward, in our small market they would also be incredibly lucky. Maybe Martin and a few others should think about the challenges of doing this, as part of the wondering why they haven’t.

 

It can only work by comparison to the small subset of almost the same sounds already contained in the software, so anything little known won’t be on it. Mostly lyrebirds are easy to pick. Very different context to model sounds. Quick changes of sounds, many different sounds included and pretty much all sounds at the same high volume. Other bird vocal mimics could be harder to pick. For example a few times I have had a calling Wedge-tailed Eagle in my garden that turned out to be an O-b Oriole. A machine wouldn’t distinguish the Regent Honeyeater mimicry from the Red Wattlebird model.

 

No doubt David is right. It is not just how hard it is to record the sounds it is that often the sound we are interested in is just one of many. Our ears are good at discriminating.

 

Philip

 

From: David Rees [
Sent: Tuesday, 28 November, 2017 9:57 AM
To: David McDonald (personal)
Cc: CanberraBirds
Subject: Re: [canberrabirds] Sort of related - bird call ID apps

 

Interesting, as someone who does a lot of bird sound recording when filming and in addition to filming I find that without a fully wind muffed microphone, most of the time what you will record will be crap.   Getting good audio is sometimes harder than the filming.  The microphones on all mobile phones, regardless of price, are not wind protected, similar to on-camera mics outdoors, mostly useless.  Software can help with a low cut filter but there are limits. You can however plug a properly protected mic into a mobile phone (though Apple are making it harder with their 'playing with ports' game), this would allow the app to get a 'clean feed' in anything other than dead still conditions.  Wonder how it will cope with mimics, such as the lyrebirds!.

 

David

 

On Sat, Nov 18, 2017 at 9:00 PM, David McDonald (personal) <> wrote:

There are a number of Shazam-type apps for identifying bird calls using a smart phone for the USA and other regons, and they are now appearing for Oz birds as well.

See Bird Song Id Australia - Automatic Recognition https://itunes.apple.com/au/app/bird-song-id-australia-automatic-recognition/id1079325343?mt=8
quote, my emphasis:
Identify birds of Australia by their Songs and Calls with this stunning guide!
Record a bird singing and use Automatic Recognition to help identify it. Includes high quality images and audio. No internet connection is needed!

With Bird Song Id, identification can be enjoyed by the whole family.

Test your knowledge with our Bird Song Quiz, see if you can become the highest scorer!

Bird Song Id has been created by professional published ecologists, and uses propriety image and audio libraries. Bird Song Id can also be used to keep a log of your own bird recordings.

This app includes 110 bird species and over 200 songs and calls. All the most common garden birds are included.
unquote

The mention of 'No internet connection is needed' tells us that the app is large - it has to store the audio files on your phone. Yes, 261 MB, but that is only half the size of the Frog ID app and a third the size of PK Birds.

David
-

On 10/11/2017 9:04 AM, Martin Butterfield wrote:

I and a few others have occasionally wondered why there isn't a bird call identification app along the lines of Shazam.  The reasons given all seem to be along the lines of "Its too difficult to do every bird call so we'll do none."' Which is not particularly helpful!

 

As a small step along the way the Australian Museum has launched what could be called Shazam for Frogs.

 

 

 

 

<Prev in Thread] Current Thread [Next in Thread>
Admin

The University of NSW School of Computer and Engineering takes no responsibility for the contents of this archive. It is purely a compilation of material sent by many people to the Canberra Ornithologists Group mailing list. It has not been checked for accuracy nor its content verified in any way. If you wish to get material removed from the archive or have other queries about the list contact David McDonald, list manager, phone (02) 6231 8904 or email . If you can not contact David McDonald e-mail Andrew Taylor at this address: andrewt@cse.unsw.EDU.AU