[X3D-Public] Integrating Speech Synthesis into X3D Systems

[John A. Stewart]

Len;

I think speech is an interesting area; FreeWRL did use the apple osx  
"say" command, and via the FreeWRL midi interface, one can manipulate  
any soundclip with professional midi tools.

I'm not focusing on the "say" interface in FreeWRL, as there are some  
nuisances with using it, and it is not platform independent.

I had prototyped a speech interface where I  could say "Viewpoint"  
"Left" etc, but the issue there was that my accent did not match the  
algorithm! If I tried to say things like a Californian, I had better  
luck, but overhearing me one night my wife was left wondering where  
the good looking blond-haired blue-eyed surfboard carrying guy  was...


Anyway,  fun stuff. JohnS.


On 24-Feb-10, at 7:33 PM, Len Bullard wrote:

> What would be the impact on the utility of real-time 3D if SSML were
> integrated into simple X3D avatar authoring utilities?
>
> http://www.w3.org/TR/2010/PR-speech-synthesis11-20100223/
>
> For dispatch systems, user interfaces should include animated  
> warnings.
> Voice synthesis would be perfect because then it is authored for  
> rewriting.
> Rerecording is always more expensive than rewriting, so even if the  
> eventual
> piece is replaced by a human reader, when developing the speech,  
> speech
> synthesis can do for the written text what midi does for musical
> composition.  Better development, better lifecycle, better  
> maintenance and
> finally, better upgrades and adaptation of the original speeches  
> into new
> contexts.
>
> len
>