chopping up sound files

[Nicholas Thieberger]

I strongly believe that chopping up sound files into smalll parts is 
a problem. First, because it creates lots of files that can easily be 
separated or misplaced. Second, because it is better to reference a 
complete sound file to give context to any piece that you want to 
quote. So using timecodes to annotate a whole file is the basic first 
step and creates an index of the archival form of the media.

However, there are times when you do need to chop up a sound file. 
For example, you may need to insert audio clips into an online 
dictionary. If a speaker of the language has had the patience to read 
all the headwords, then you can transcribe and time-align this 
material using Transcriber, and then export the text using the 'Limsi 
label' format.

The freely available program Audacity can import this file as 
Audacity labels. Having opened the audio file and then imported the 
labels you can select the 'Export multiple' option and Audacity will 
chop up the file along the labelled timecodes, and will name each 
file with the label. This means you will then have many named files, 
ready to be referenced by your dictionary, and, if the labels are the 
same as the headword then it should be easy enough to automate the 
linking of the audio and text of the dictionary.



-- 
Project Manager
PARADISEC
Department of Linguistics and Applied Linguistics
University of Melbourne, Vic 3010
Australia

nicholas.thieberger at paradisec.org.au
Ph 61 (0)3 8344 5185

PARADISEC
Pacific And Regional Archive for Digital Sources in Endangered Cultures
http://paradisec.org.au