Subject: song-recognition software
Date: Mar 27 11:06:37 1995
From: Dennis Paulson - dpaulson at ups.edu
James West wrote:
". . . And the second of these criteria would make me uneasy about the
automatic recording of bird species by song recognition software. You'd
have to hope there weren't too many starlings in the vicinity mimicking
everything in earshot. And could the software be made sophisticated enough
to recognize _every_ vocalization of common birds with a very varied
repertoire . . .?"
I think software could be made to recognize just about anything, James.
You'd just have to plug in all the known vocalizations of species A, B, C,
etc., and the software could be programmed to recognize any subtlety that
identified them. "If it quacks like a duck, it *is* a duck." And it would
be easy to program the software so it would specify "flicker" if the call
was repeated several times, but "starling" if it was interspersed with
other vocalizations. Perhaps a single flicker call from a starling and a
flicker might be indistinguishable, but, then again, there may be subtle
differences between them (I wouldn't doubt it) that the computer could pick
up more easily than we do. I don't know if anyone has taken a
mockingbird's song and tried to distinguish all of its imitations from the
real things by using sonagrams, but this would be of some importance for
such projects.
Generic software would have to be produced that could be "loaded" with the
vocalizations from a given area, to take into account the suite of species
present and the local dialects of many of them, but that doesn't seem too
difficult.
They've got surveys going on in many areas that record bat squeaks and
identify them to species, greatly enhancing our knowledge of bat
distribution and abundance. This is done by people, not software, but I
imagine software is just a matter of time. I don't know if all species can
be distinguished or not; I think bird songs would be even easier, but that
may be my ignorance.
Dennis Paulson, Director phone: (206) 756-3798
Slater Museum of Natural History fax: (206) 756-3352
University of Puget Sound e-mail: dpaulson at ups.edu
Tacoma, WA 98416