Design suggestion: The server just for synthesis

[Michael Pozhidaev]

Hello, William!

> speak).  It handles the messages that the clients send and passes them
> over to the modules, which do the actual tts/audio processing.

Modules both for TTS and audio tasks must be in the one single
application and must be executed on the same machine. As I wrote to
Hynek, using separated servers can increase stability and extend area of
tasks  we will be able to do.

> I don't see the advantage of moving any of that processing to the server
> since it has to be handled differently for each synthesizer.

All modules we have for TTS engines will be moved into separate process
and we will be able to get raw audio data directly with it.

One example: with new server just for TTS tassks I can use the single
core for screen reading software and for text books reading
software (daisy or anything else), including speech generation not for
real-time listening but for packing into OGG. 

-- 
Michael Pozhidaev. Tomsk, Russia. E-mail: msp at altlinux.ru
Russian info page: http://www.marigostra.ru/