View this page "mobicents-tts-ra"

[amit.b...@gmail.com]

Click on http://groups.google.com/group/mobicents-public/web/mobicents-tts-ra
- or copy & paste it into your browser's address bar if that doesn't
work.

[Oleg Kulikov]

It should be an extension of media ra.

mediara.play("file:/welcome.wav");
mediara.play("You are welcome");

On 19 май, 15:16, "amit.bhay...@gmail.com" <amit.bhay...@gmail.com>
wrote:
> Click onhttp://groups.google.com/group/mobicents-public/web/mobicents-tts-ra

[Ivelin Ivanov]

Not following. what would mediara.play() do without being associated
with a media session or media connection?

TTS is responsible for converting text to speech. It seems fine to me
to have the speech written in a file before being played via the Media
RA. Eventually it may make sense for the TTS to stream the speech, but
doesn't look like an immediate problem to solve.

Ivelin

[Amit Bhayani]

In addition I will extend the TTS RA in future where user can add different Voices and Lexicon. As of now its only available with US English lexicon and three fixed voices. Do we want to blot the Media RA with all these details of TTS?

Amit.

[Ivelin Ivanov]

I suppose these can be initialization properties for a new TTS session.

[Ivelin Ivanov]

what is the lifecycle of a TTS session? I can imagine synchronous
requests for TTS conversion and asynchronous events with URL (or SDP
in the future) to the file where the result is available. Asynchronous
events can also provide updates on the progress if its a large text
file or there are any intermediate errors.

What else can happen in the lifecycle of a TTS session? Possibly a
session initialized with given lexicon and Voice can be reused
multiple times by feeding new text.

Ivelin

[Oleg Kulikov]

Anyway, speech or voice are media data. The extesion of media ra with TTS related events seems to me more transparrent and will be easy used. From other hand media ra is deisgned to not only for RTP processing but to work with all kind of media oriented sessions. TTS may be considered as special endpoint for MGCP or special context for Megaco and developer may use same api to record speech to a file or for rtp streaming.

 

 

2007/5/21, Ivelin Ivanov <ivelin.atan...@gmail.com>: