Hi, Is there a way to return a list of *phonemes* (not a list of words) when using x-webkit-speech in the Chrome browser?
Maybe there's an API that's been published somewhere that I've missed, (I just started trying out x-webkit-speech recently after researching lots of other speech-recognition technology eg: http://swik.net/voice+Software?popular) This new x-webkit-speech API seems very well-done, and the fact that it works in the browser is important. I'd love to see a list of any methods and properties which might be available to a web app programmer. Thanks! ==== Here's an example application: Say I am interested in trying to use x-webkit-speech speech recognition for teaching English. Many times, I already know *exactly* what the student should be saying (eg, I displayed the "target" sentence as text, and simply asked the student to read same the text out loud back to the browser, to test the student's pronunciation). The student might mispronounce some of the words (or the speech-recognizer might not recognize them even if the student pronounced them more-or-less correctly :-). Rather than return the words the speech recognizer *thinks* it heard, I would like to return the originally displayed "target" sentence - only now with some words in red (ERRORS: the student seriously mispronounced these words) and some other words in yellow (WARNINGS: the student kinda mispronounced these words). But to do this properly, I really should focus internally on *phonemes*, not on words. I assume somewhere in the inner workings of x-webkit-speech, the utterance is represented as a list of phonemes (before x-webkit-speech converts that list of phonemes into words, to display to the user). Is there an API where I can access this list of phonemes? Thanks.
_______________________________________________ webkit-help mailing list webkit-help@lists.webkit.org http://lists.webkit.org/mailman/listinfo.cgi/webkit-help
