VoiceXML - Full Duplex ?

[vickk7]

Hello,

I am trying to develop an "audio-karaoke-like" soft for phone users.

The server plays an audio file stored on its hard drive through the
phone, and the phone user sings on the music. The voice has to be
stored on the server, (in order to be mixed with the music...)

I think I need something like a "Voice XML Gateway". But I cannot
find if VoiceXML supports full-duplex. Can it play an audio file while
recording an incoming stream (voice) ?
And is recording time limited or the generated audio file size limited ?
Can I record a user singing during 5 minutes ?

Thanks.

Vickk

 

[Joseph Kesselman]

VoiceXML "describes the human-machine interaction provided by voice
response systems". It's a control file for those interactions, but does
not actually carry any audio data within the XML file. (And it
shouldn't; binary data is not something XML is particularly good at
handling.)

See http://www.w3.org/TR/voicexml20 for the spec. There are some
tutorials at http://www.voicexml.org/tutorials/intro1.html; the very
first intro slide has a block diagram which shows that the audio is
processed alongside the control scripts rather than through them.

VoiceXML can request that a full-duplex "bridge transfer" be
established. But that's the only instance of "duplex" in the VoiceXML
2.0 Recommendation.

 

[vickk7]

Hi Joseph.
Thank you for your answer.
I knew that audio data is processed alongside the control script.
If the VoiceXML can request a full-duplex "bridge transfer", I
think I can manage to develop my "karaoke service". I suppose
the audio data going through the "bridge" is already in a
"understandable" format as .wav or .au . That's fine ! Then I
only need to catch and store the incoming audio.
Thank you !
Vickk.