Where's the voice in VXML
So an interesting requirement popped up recently and I was somewhat surprised that the VXML specification did not cover it. Why can't I record the utterance made within a field?
Say I want to store the result of a field in both a text and audio format, there is no way to do this right now. As anyone knows speech recognition can be patchy at times, wouldn't it be nice to compare the interpretaion to the actual utterance?
Say I want to store the result of a field in both a text and audio format, there is no way to do this right now. As anyone knows speech recognition can be patchy at times, wouldn't it be nice to compare the interpretaion to the actual utterance?
posted by Brian OConnor at 5:58 PM
3 Comments:
You can't believe the coincidence of this, but the exact same requirement recently came up for me. Nice blog, good writing by an understanding mind in a rapidly changing field. How do I become a contributor?
Use the VoiceXML 2.1 "recordutterance" feature described in http://www.w3.org/TR/voicexml21/#sec-reco_reco
Thanks Ken,
I am guilty here of basing this comment on the available implemented platforms at the time of writing. Sloppy of me not to have reviewed the W3C specification.
Post a Comment
<< Home