v:* { behavior: url(#default#VML) }
o:* { behavior: url(#default#VML) }
.shape { behavior: url(#default#VML) }
Speech Data for DTW
Six continuous speech utterances produced by
six different speakers, 3 males and 3 females, are selected from the TIMIT
database. The contents of these six utterances have the same transcription.
However, because of the different acoustic characteristics of the six speakers,
it turns out to be six wave files with very different phonetic properties and
time periods.
The Speech Data for Waveform
and Transcription Files
The speech data is saved as a zip file
DTW.zip
. Download and extract the zip file to get the waveform and
the transcription files. There will be only one folder named “DTW”
after extracting. Six speech waveforms and one transcription file will
be obtained. The transcription file,
transcription_word.txt, is the word transcription for the speech
sentences. The
sizes and the time periods of the six
waveform
files are listed below:
female1.wav = size: 103 KB
length:
3.298 sec
female2.wav = size: 103 KB
length: 3.299 sec
female3.wav = size: 109 KB
length: 3.512 sec
male1.wav = size: 97.3 KB
length: 3.112 sec
male2.wav = size: 97.4 KB
length: 3.116 sec
male3.wav = size: 104 KB
length: 3.349 sec