Speech To SSML

Voice Browser Working Group,

Another interesting SSML topic is speech-to-SSML.  Prosodic data can be output by speech recognition components along with recognized text or XML content.  There have been some recent developments in that regard.

At the Speech Prosody 6th International Conference, a paper PROSOTRAN: a tool to annotate prosodically non-standard data (http://www.speechprosody2012.org/uploadfiles/file/sp2012_submission_156.pdf) describes a technology, PROSOTRAN, which "generates multi-tiered transcription that account for the way the three prosodic parameters vary over time in a given utterance."



Kind regards,
Adam Sobieski