Comments on the Text to Speech "algorithm"

[Chris Brannon]

Tim Cross wrote:
> However, I believe I'm in the minority and the demand is for
> natural sounding voices that are used mainly at natural speaking rates and
> that it will get harder to get the older style.

I wasn't going to weigh in on this thread, but I don't like the
"human-sounding" voices, either.  To me, they sound like a mockery of the
real thing.  Human speech is not a mere concatenation of utterances.
It contains emotion, and it is delivered in some context.
Those two factors really don't apply to machines.  At some level,
synthesized speech is always going to be flat and mechanical.
If you want to hear an amusing example of this phenomenon, and you have
Windows Vista, try reading some lewd sentences with the Microsoft Anna
voice.  The construction of lewd sentences is left as an exercise for
the reader.  Be creative!  Anna sounds fairly human, until you put some
provocative phrases in her mouth.
Myself, I'll stick with the mechanical-sounding voices, but to each his own.

-- Chris