speechd-discuss
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

SAPI server with speech dispatcher


From: Thomas Lloyd
Subject: SAPI server with speech dispatcher
Date: Sun, 18 Jan 2009 11:51:31 +0000

Dear All

The Open SAPI server is now working with speech dispatcher after a
rework today implementing a pipe mode.

I am not familiar with ORCA and it's capabilities but I am using the web
and writing this e-mail with Microsofts speech alone. It seems a little
bit more responsive than espeak but only time and testing from weathered
screen reader and text to speech users will tell , I am a little biased
at the moment :). At this stage the system is not designed for speed but
it is not bad for a first release. 

I have written an open-sapi-generic.conf that calls on the sd-generic
binary and would appreciate a little help getting the rate volume and
voice elements working. It has been designed so that in a future release
I can write a open-sapi binary to communicate directly with the speech
server itself. For now  I am contented with what I have.

I would really like to get the open-sapi configuration files included in
the SD source who should i contact, send them to or ask for permission
from?

I have included the values for the settings in the conf for volume and
rate; as pitch does not currently work. I would appreciate some help
getting the settings correct.

#           SSIP Range  SSIP Default    o-sapi Range    o-sapi Default
#           ----------- ------------    ------------    --------------
# Rate      -100 to 100     0           -10 to 10        0
# Volume    -100 to 100     0            0 to 100        100

GenericRateAdd          0
GenericVolumeAdd        100

# (These values are multiplied by 100, because DotConf currently
# doesn't support floats. So you can write 0.85 as 85 and so on.)

GenericRateMultiply     10
GenericVolumeMultiply   0

Some information on the requirements of the SIGKILL on SD would also be
useful as it is currently not implemented the client will die but the
server will continue to speak the queued messages. When SD calls a
SIGKILL should the server immediately stops and purge the current
speech? 

The next few weeks will be putting together an installation script to
make things easy for users to install.

The system is a little bit rough around the edges but if anyone would
like to test the system for its current performance then let me know by
e-mail.

Tom 














reply via email to

[Prev in Thread] Current Thread [Next in Thread]