Re: [Asterisk-java-users] Text-to-speech and gathering data (fastAGI)

[Stefan R. <sr...@re...>]

Darren Hartford wrote:
> Thanks Stefan,
> My intent is actually to optimize the second part of non-static
> requests.  This test does indeed could used pre-generated voice files
> (per the cached remark), but the intent is to identify the best way to
> handle the dynamic parts.
>=20
> Is this the best way to handle dynamic tts parts?=20

You are probably thinking of something like directly streaming the tts
generated voice files to the user instead of sequentially creating the
while voice file and then streaming it. As far as I know this is
currently not possible.
So the best approach that will currently work is
- use pregenerated voice files wherever possible
- try to minimize the dynamic generated part
- generate the dynamic files in the background and as soon as possible,
i.e. as soon as you know what you have to say create them in a new
Thread (not even through AGI but directly invoiking your tts engine) and
meanwhile play something else to the user to avoid dead air.

=3DStefan

--=20
reuter network consulting
Neusser Str. 110
50760 K=F6ln
Germany
Telefon: +49 221 1305699-0
Telefax: +49 221 1305699-90
E-Mail:  sr...@re...
Jabber:  sr...@ja...