embedded text to speech converter

[Kyle]

If eSpeak doesn't sound as good as you like, SVox Pico may work better 
on embedded systems than OpenMary. SVox Pico is the default Android 
speech synthesizer, and speech-dispatcher works with it somewhat now, 
and support should improve, as its module is rather new. Also, neither 
SVox Pico nor eSpeak require Java to be installed, although there is 
some Java stuff in SVox Pico's git tree, presumably for Android. It 
isn't needed for Ubuntu AFAIK. Take a look at libttspico0 and related 
packages on Ubuntu 10.10 and later.

You shouldn't need a speech system that is based on QT,. You should 
simply be able to link your QT application against the needed speech 
libraries and program your application to speak where necessary. You 
could connect to speech-dispatcher through its various backends, or 
simply link against the library for your speech synthesizer of choice 
directly. My personal recommendation is to use speech-dispatcher, since 
it provides an abstraction layer for a number of free and proprietary 
speech synthesizers.
~Kyle