Design suggestion: The server just for synthesis

[William Hubbs]

Hi Hynek and all,

On Fri, Nov 12, 2010 at 04:43:51PM +0100, Hynek Hanke wrote:
> On 12.11.2010 07:38, Michael Pozhidaev wrote:
> > Yes, removing TTS modules from Speech Dispatcher can significantly reduce
> > current complexity. Having TTS modules both in Speech Dispatcher and in
> > new server we are discussing is something strange.
> 
> Completely agreed. Then that puts the thing into a new light
> and is even more along the lines I was thinking about for
> the future.
> 
> I don't actually think we are talking about some entirely new
> server. It's much more about refactoring the existing Speech
> Dispatcher technology into three pieces, re-using its knowledge,
> its design etc.
> 
> So if we are to commit that this is the way forward for the Speech
> Dispatcher project and part of the Free(b)Soft project, we need
> to work together on finishing TTS API, setting up a roadmap, setting
> up a technical specification and the work on its implementation
> under the review of the existing project reviewers. That way would
> guarantee that the work is towards the same goals and it is spent
> in a useful way.

I have to admit that I'm completely lost here. I don't understand why we
need another layer, I don't even quite understand the purpose of the
TTSAPI.

Currently, the flow is from the client (the program that wants to
speak), to the SD server, which prioritizes the messages and sends them
to the output module for the appropriate synthesizer. The output module
for the synthesizer handles all of the communication to the synthesizer,
and also plays audio if that needs to be done.

Hynek, can you elaborate more on what needs to be factored out of sd,
and what TTSAPI does? Couldn't we just add the functionality in TTSAPI
to speech dispatcher some how?

William

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: not available
URL: 
<http://lists.freebsoft.org/pipermail/speechd/attachments/20101112/ced5ec51/attachment.pgp>