The most human speech I have heard to date is the MicroSoft Speech
engine. It actually uses fragments of recorded human speech which are
spliced together in real time to form the words. The inflection can be a
bit odd, and really pretty funny some times, but most words are really
pretty natural.
Don't confuse JFW talking install with this. That is just a wave file of
Earl Harrison's fine voice telling you about Dopctor Jaws and all of
that.
Jim Snowbarger
-
Visit the jfw ml web page: http://jfw.cjb.net
