Design suggestion: The server just for synthesis
msp at altlinux.ru
Fri Nov 12 06:29:12 CET 2010
I (maybe better "we") need well-structured set of tools with clear
purpose and having features exactly required. Production quality is the
main strong condition.
You are offering way of easy minor fixes, perhaps it is possible, but it
is the way with ambiguous future.
Let see once again the example with large text reading software. Using
patched spd just for TTS tasks is something like using gun just for
And, finally, you wrote it is easy to make all connections in spd
completely independent. I am not sure spd with such modification will
stay usable in usual way.
>> On 4.11.2010 20:22, Michael Pozhidaev wrote:
>> what do you think about creation an new separated server dedicated only to speech synthesis?
> well I still didn't get why you need a "new separated server" and how
> this server will interact with SD
>> 1. The server has no any audio output.
>> 2. It can accept connection, receive portion of text and send audio data back to the client.
> This what SD can do with some modifications
>> 3. Connections are completely independent.
> This is also can be done pretty easily in SD by modifying server core
>> 4. SABLE(SSML) markup emulation for TTS without their support.
> IMO, since SD API supports SSML this functionality shall be added to SD
>> 5. Clear configuration format and plug-in structure.
> Speechd mailing list
> Speechd at lists.freebsoft.org
Michael Pozhidaev. Tomsk, Russia. E-mail: msp at altlinux.ru
Russian info page: http://www.marigostra.ru/
More information about the Speechd