[MV] Re: Questions on customising VV

[T. Patrick Henebry]

>My worry is that, by training and re-training a single model in both 
>environments, I may be (in effect) corrupting the model;  certainly 
>it is redundant, irritating & time-consuming.  In other words, I see 
>the wisdom of having 2 voice models, and I'm all for it.
>
>All I was trying to do was / is figure out how to have a second 
>voice model without having to go through reading those damn' stories 
>again, et cetera, et cetera.

There is no way to create a second voice model without reading the 
training stories again. Reading the trading stories is essential to 
generating data the computer uses when creating a voice model.

I myself have multiple voice models to account for varying 
conditions. Although I never had a problem with corruption when using 
a single voice model I found the time involved in continuously 
retraining to be prohibitive. It was well worth the time for me to 
create a second voice model and switch between them as necessary 
(while being sure to run Audio Setup after switching).
-- 
----------
T. Patrick Henebry