blob: 0dec93c9270dd4301d4e4dcbb9510748315933d5 [file] [log] [blame] [edit]
This directory contains the BRLTTY speech driver for the Theta text to speech
engine from Cepstral [http://www.cepstral.com/]. It was implemented, and is
being maintained, by Dave Mielke <dave@mielke.cc>.
BRLTTY's configure script automatically includes this driver if Theta has been
installed. The default is to check the directories /usr, /usr/local,
/usr/local/Theta, /usr/local/theta, /opt/Theta, and /opt/theta. The actual
location can be explicitly specified via --with-theta.
This driver and that for the FestivalLite text to speech engine cannot be
linked into BRLTTY's binary at the same time (via --enable-speech-driver)
because their run-time libraries contain conflicting symbols.
If this driver is built as a dynamically loaded shared object (the default),
i.e. it isn't linked into BRLTTY's binary via --with-speech-driver, then the
directory $THETA_HOME/lib must be added to the colon-delimited list of
directories in the LD_LIBRARY_PATH environment variable before BRLTTY is
started because Theta's own shared objects don't contain run-time search paths
for their internal dependencies.
This driver recognizes the following parameters:
Parameter Settings
age 1-99 (years)
gender male, female, neuter
language two-letter language code
name name (human), /path/to/directory
pitch -2.0-2.0 (octaves relative to default)
If the human name of the voice is specified then the voice path is searched.
This path can be explicitly defined by storing a colon-delimited list of
directories in the THETA_VOXPATH environment variable. The default is to look
in $THETA_HOME/voices.