warning on formats and media

[Brian MacWhinney]

Dear Info-CHILDES,

     The rise of MP3 and DVD as popular consumer formats with readily  
available playback devices may lead researchers to think that these  
formats are also suitable for research on spoken language in natural  
contexts.  In fact, both of these formats have problems when used   
for transcript-based analyses.  Both are compressed in ways that make  
it difficult for software to find exact begin and end points of  
utterances, words, and gestures.  The CHILDES CLAN program can play  
back MP3 because MP3 is a format supported directly by Quicktime.   
However, because of the nature of MP3 compression, QuickTime and CLAN  
cannot locate precise time points within an MP3 file.  For this  
reason, I strongly advise researchers to use uncompressed audio  
formats like WAV and AIFF or OGG.  Currently, all audio in the  
database is in WAV format.  If you wish to generate MP3 files for  
special purposes such as student work,  please always be sure to  
maintain uncompressed WAV files for archival purposes.  It is also  
true that careful phonological analysis in programs like Praat or  
Phon will work much better with WAV files than with MP3 files.
     On the video side, there is an even more serious issue regarding  
the commercial DVD format.  Currently, there are no methods at all  
for linking transcripts to material in DVD format.  Instead, CLAN  
requires linkage to any of the compressed file formats supported by  
QuickTime, which include the various MPEG formats.  Please note that  
using DVD media to store QuickTime or WAV files is not a problem at  
all.  It is only the full DVD movie format that is a problem.
    For both video and audio, it is important to preserve your  
original uncompressed media as your archival source.  For audio, this  
means the WAV files.  For video, it typically means the mini-DV  
cassettes used by a digital recorder or the original VHS.
    If you are participating in a project that is having problem with  
the monetary resources needed for the archival storage of  
uncompressed media, please contact me and we will try to figure out  
some way of providing assistance or at least minimizing the negative  
consequences of destroying your original media.
    Many thanks,

Brian MacWhinney