[asterisk-users] Replacement for Allison

[Matthew Rubenstein]

On Thu, 2008-01-24 at 13:51 -0500, Steve Prior wrote:
> Matthew Rubenstein wrote:
> > 	Is anyone else interested in creating new voices for Festival (the
> > voice synth bundled with Asterisk) that might not be as good as
> > Allison's recordings, but are better than the current Festival voices?
> 
> If you try to do live voice synth for prompts you'll probably run into 
> something I encountered - the background and speechbackground (I also 
> use Lumenvox in addition to Cepstrel) only take a file.  That means 
> unless I'm missing something you can't have a TTS prompt that can be 
> interrupted like a recorded file.  The workaround of TTS to a file and 
> then play the file sounds to me like it would introduce delays and 
> besides it's ugly.

	Since Asterisk has a preset collection of prompts, the voice synth can
be used to generate those files when installing Asterisk (or whenever).
The difference from Allison's prompts is that when a new app calls for a
new prompt, the synth can generate its new file instead of hiring
Allison to do it, and all the prompts sound consistent. And there are
ways to synthesize arbitrary new prompts at runtime just in time to be
picked up by the apps that can play only a file.


> That's why I posted a suggestion to the recently created 
> asteriskideas.org that background and speechbackground be enhanced to 
> take an app in addition to a simple file.  If you agree, then please 
> vote for it.

	I think the flexibility you described is important anyway.


> Steve
-- 

(C) Matthew Rubenstein