Re: Very Nice Tool
> When you couple this program with an
> easy to use interface (like
> perlbox-voice) it is extremely nice. The
> voices are every bit as clear as
> anything in the Microsoft toolbox, and
> the only thing going for Linux.
I'm using festival from a Perl/Tk application. It's wonderful to be able to plug in this kind of functionality so easily. I shudder to think how much of a hassle it would have been (both for me and for potential users) to do the same thing in the world of proprietary software. Woo hoo!
Re: Difficult to configure
if you are using a redhat variant....
if you have a crappy soundcard like i do, festival does not
work on it. it will talk too fast.
a slightly unbearable fix for this is to become root and
cd /usr/share/festival/voices/english/kal_diphone/festvox. then
look for a line which says something like:
(Parameter.set 'Duration_Stretch 4.5)
change the value at the end to slow or speed up the default voice...
it is still chipmunky though and i have found that changing this voice does not also change the voice in Gnopernicus which i was hoping it would. My mom is blind and i was trying to enable the speech function with this method.
Very Nice Tool
When you couple this program with an easy to use interface (like perlbox-voice (http://perlbox.org)) it is extremely nice. The voices are every bit as clear as anything in the Microsoft toolbox, and the only thing going for Linux.
Re: Difficult to configure
> The directions are scattered about the
> site's pages. Instructions concerning
> configuring voices are scattered between
> several site sections (manual, download
> addendums, etc.) Troubleshooting
> instructions seem to assume nothing
> beyond faulty directory pointers.
I must admit that the documentation on the Festival site has improved greatly in the last few years. My only remaining hurdle to use of it at this point is their lack of a good troubleshooting area. Festival speaks too fast to understand on my system.
This is really very cool, its text to speech capability is useful and output is understandable. Had grief in trying to get the java support working in 1.4.1 though.
Festival does text to speech
Festival takes a little effort to get working, but is pretty easy if your sound works and you install the rpm's from the festival web site. There are several voices available with the package. I tried all of them. The default one is decent. The one I like best is ked_diphone. You can change the default voice in /usr/share/festival/lib/voices.scm in the default-voice-priority-list directive.
You can make festival read text kernel docs to you with:
for f in *.txt
do festival --tts $f
Difficult to configure
The directions are scattered about the site's pages. Instructions concerning configuring voices are scattered between several site sections (manual, download addendums, etc.) Troubleshooting instructions seem to assume nothing beyond faulty directory pointers.
An open, cross-platform journaling program.
A scientific plotting package.