Data Sets



Transcription Conventions

The number in the title of each entry (e.g Ng : 43) indicates the time location (in seconds) from the start of the recording.

The following are the conventions used in transcribing the data. They are based on those used in the VOICE project.

S1, S2 Speaker 1, Speaker 2
(.) short pause
(2) pause of about 2 seconds
hh audible intake of air
@@@ laughter
<@> words </@> words spoken accompanied by laughter
<coughs> non-linguistic sound
word: word that is extended
word:: word that is extended considerably
(word) uncertain transcription
WORD word (or syllable) said with extra emphasis
. falling intonation
? rising intonation
<1> words </1>   overlapping speech
word word that was misunderstood
<spel> u b d </spel>   word that is spelt out
<un> xxx </un>   unclear speech; each 'x' indicates one syllable
... indicates it is not the start of the speaker's turn