Processing steps are:
Input - text string (of words in lexicon)
uttmake - makes XML wordseq
uttclass - looks up words in dictionary
uttcomp - build IP from components
ProXML scripts:
- assign default durations
- run Klatt rules
- calculate durations on segments
uttmbrol - MBROLA output
A document describing the implemented ProXML language is on its
way. I should be able to send it out this afternoon.
Mark
P.S. The first sentence? "They say it's a lie"