Run a command line and play the output file.
It is mainly for testing purpose. It may not fit for extensive use (but it may). Once you have chosen your TTS engine, write a native module!
- In modules.conf, add asr_tts/mod_tts_commandline.
- Run make, make install.
- copy the msvc project from freeswitch-contrib in grmt\mod_tts_commandline for windows
- add project to Freeswitch.2010.sln and build it (sorry, I didn't test MSVC 2008)
- In freeswitch/conf/autoload_configs/modules.conf.xml, uncomment <load module="mod_tts_commandline"/>.
- Edit conf/autoload_configs/tts_commandline.conf.xml.
- Reload configuration by reloadxml (or restarting FreeSWITCH)
- Add tts_commandline to dialplan (or phrase, etc.) like this:
Using festival's text2wave:
Using festival's server client:
(may be needed adjust festival's default sample rate on server)
Using Cepstral's swift:
Using espeak + mbrola :
Using svox pico (from the Android project - supports it-IT, fr-FR, es-ES, de-DE, en-US and en-GB languages:
Write your own here...
Notes on svox pico
- build from MSVC 2010 solution file freeswitch-contrib in grmt\svox pico\svoxpico for windows (again sorry I didn't do MSVC 2008)
Linux (at least tested on x64 centos - includes fixes from below and even more fixes from debian-sid and some more fixes to make it compatible with MSVC 2010):
- freeswitch-contrib in grmt/svox\ pico/svox/pico
- sh ./autogen.sh
- make && make install
or alternatively :
Here is a sample of cmd that work on centos 5 64bit (This code includes a fix so it works ok on a 64bit system)