[Lexicog] Re: audio component

Bill Poser billposer2 at GMAIL.COM
Mon Apr 29 02:16:23 UTC 2013


I use a program called SndBite, which is a specialized audio editor
intended specifically for this purpose. If you are confident of the
sequence in your file, you can give it a list of file names to use. An
alternative is to let it generate file names itself (fragment001.wav,
fragment002.wav, etc.) and map these to the desired names later. It
generates a log of every write, into which you can insert notes if desired.
The other thing that speeds things up is that it automatically computes
transitions between intervals of silence and sound and makes a list of
silences greater than a certain duration. It can then be asked to move so
as to straddle the next chunk of sound. The sound/silence decision is
tunable. I have used this program to cut out thousands of words and
sentences and find that most of the time it works nearly automatically -
you tell it to advance, look and listen, decide that it has done the write
thing, hit 'write window', and repeat. Of course it is easier to use your
own software, but I taught a young woman, bright but with just basic
computer literacy and no linguistics, to use it in 15 minutes.

The editor is actually scriptable, in Tcl, and can be made to perform this
task without human intervention. If your files are sufficiently clean and
regular, this may work but you should experiment first.

The interface is a little unusual and various things could use improvement
or updating but it seems to work pretty well. Anyhow, have a look at:

http://billposer.org/Software/SndBite.html
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.linguistlist.org/pipermail/lexicography/attachments/20130428/ecac9e78/attachment.html>


More information about the Lexicography mailing list