Closed saulalbert closed 6 years ago
Latching (i.e. speaker transition without any discernible silence) is marked in CAlite with the '≈' unicode symbol.
The problem is that WITHIN-speaker latching (i.e. when a speaker rushes to start the next utterance) makes the latched-together words unreadable.
So e.g.:
hello≈how are you
Is not parseable by a lexical search for 'hello' or 'how'.
So CAlite2CHAT will add spaces around the '≈' marks, and CHAT2CAlite should remove them.
should be converted to
hello ≈ how are you
Latching (i.e. speaker transition without any discernible silence) is marked in CAlite with the '≈' unicode symbol.
The problem is that WITHIN-speaker latching (i.e. when a speaker rushes to start the next utterance) makes the latched-together words unreadable.
So e.g.:
hello≈how are you
Is not parseable by a lexical search for 'hello' or 'how'.
So CAlite2CHAT will add spaces around the '≈' marks, and CHAT2CAlite should remove them.
So e.g.:
hello≈how are youshould be converted to
hello ≈ how are you