mjsull:I'd like to see a "Shazam" type function for bird sounds. It would be awesome to add this capability to iBird. I would think it would be easier to develop than the music oriented application due to a significantly smaller variety of sounds.
Its a good idea which I address in our FAQ. While it is a cool idea, there are some issues that make it very difficult.
The Shazam program takes advantage of a large digital library of music
that can be compared to the sounds received by the iPhone microphone.
And since digital music is like a finger print the comparison process
can be very accurate.
Bird calls and songs on the other hand are more
complex than human speech. Not only does every species have a repertory
of sounds - no two birds sound alike - indeed the same species of bird
in the same area can have different dialects. The sounds of most song birds are also delivered by two voice boxes, which means the frequency spectrum is extremely complex - you can see this in some of the sonograms on the Cornell All About Birds web site. And even if similar birds did sounds alike the frequencies of a bird call are difficult to categorize.
Then there is the problem
of the iPhone microphone, its just not up to the job of isolating a
bird song from background noise. For that you need a special parabolic
mic and a high gain amplifier with filters to remove noise. There are
other reasons that make this difficult that are fairly technical.
Here is a metaphor: Imagine a computer program that can recognize human speech in English
and convert it to text. Now consider it has to also recognize and
convert speech in Spanish, Russian, Chinese, etc. Sound like fun?
So
while we expect someday this will be feasible for now its remains on
the wish list for great ideas. If it ever does happen it will take a super computer. Of course in technology never say never so I leave the door open that someone may invent a clever way around these issues. I'm just not holding my breath.
In fact what is more interesting is how regularly people who have purchased iBird and downloaded and used Shazam suggest this idea. What do you call that phenomena?
Mitch Waite
