Next: Prompting style
Up: Classification of general strategies
Previous: Classification of general strategies
The RECORDING MODE defines the exact method of controlling transfer of speech
samples to the storage medium. It is defined in terms of the TAKE, SEGMENTS and
SPEECH FILES. It is concerned with the stopping and starting of the sampling process and
the way the recording system treats any possible speaking errors.
- MODE 1
- A TAKE is recorded in one complete SEGMENT, errors are
not excised.
(Explanation - this means that the sampling and transfer process
is started at the beginning of the TAKE, all the acoustic signal,
including breath noise, speaking errors are recorded and the
sampling and transfer process is only stopped at the end of the
TAKE.)
- MODE 2
- A TAKE is recorded in one complete SEGMENT, except when
a speaking error is detected during the TAKE, and then that
error is excised from the storage medium. If there have been
no speaking errors there will be no DISCONTINUITIES in the
SPEECH FILE. For each section of speech removed due to a
speaking error there will be one DISCONTINUITY.
- MODE 3
- A TAKE is recorded in a number of SEGMENTS. There will
be one SEGMENT per utterance, and speaking errors will be
included in the SPEECH FILE.
(Explanation - in general this means that an endpoint detector is
used to start and stop the sampling and transfer process. Apart
from the ``back-off'' time no inter-utterance acoustic signal will
be stored in the SPEECH FILE.)
- MODE 4
- A TAKE is recorded in a number of SEGMENTS. Speaking
errors will be excised from the storage medium, and so there will
be one SEGMENT per required utterance. This MODE is the
same as MODE 3 except that the resulting file will contain no
errors.
EAGLES SWLG SoftEdition, May 1997. Get the book...