message from Alan W Black <[email protected]> to festival-talk
= = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
Luisina Violante wrote:
Hi,
I've created a new Spanish voice in festival with clunits, following the
instructions described in this link: http://festvox.org/festvox/x3082.html.
Now, I wanna create a new voice (identical to the one already created)
using diphone selection instead of phone selection. I wouldn't like to
create a diphone database from scratch, recording the sentences containing
the diphones. I wonder if there's any way to do that, using the already
recorded wavs and taking advantage of the work done until here.
The technique in
http://www.festvox.org/bsv/x3144.html
Can work quite well depending on the coverage in your general database,
the technique can be extended (by hand) to select (and reject) the
diphone it finds, and get it to find other ones.
This is quite a good technique for getting a very small voice.
Alan
Thanks in advance,
Luisina
= = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
= University of Edinburgh's Festival Speech Synthesis System =
= http://festvox.org/festival Sent Via [email protected] =
= To unsubscribe mail [email protected] =
= = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
_______________________________________________
Festlang-talk mailing list
[email protected]
https://lists.berlios.de/mailman/listinfo/festlang-talk
