RE: [misterhouse-users] Speech issues (Linux/Festival)

[Keith W. <ke...@pl...>]

On Thu, 1 Nov 2001, Bruce Winter wrote:

> > Same here,
> >
> > I use:
> > set $VTxt_festival qq[(utt.save.wave (utt.synth (Utterance Text
> > "$parms{text}"))
> > "/mh/www/http_server.wav" 'riff)];
> > > In mh/lib/Voice_Text.pm, the SayText command that sends the text to
> > festival
> > > was hard-coded to 'hello world', so that was all my Audrey would say.
> > Fixed
> > > by replacing the "hello world" with $parms{text}.
> > >
> > > Now, the file that my remote machine is getting is the 'old' .wav file
>
> Silly me.  So any preference between these 2?
>
>   set $VTxt_festival qq[(utt.save.wave (SayText "$parms{text}")
> "$parms{to_file}" "riff")];

This version has the server speaking the same phrase twice, the Audrey
still speaking the 'previous' phrase.

>   set $VTxt_festival qq[(utt.save.wave (utt.synth (Utterance Text
> "$parms{text}")) "$parms{to_file}" "riff")];

The server speaks the phrase once, the Audrey speaking the 'previous'
phrase.

I also notice some interaction with the mute check box.  If I use the
function to "list recent phone calls", Audrey and server will attempt to
speak the list and the listing on the browser is "unformatted".  If I mute
wav-over-web, the listing is normal and speech from Audrey or the server.

I'm running 2.60 on Linux (Redhat 7.1), Festival 1.4.1, Audrey Final+Shell

Keith

> Bruce
>
>
> ________________________________________________________
> To unsubscribe from this list, go to: http://sourceforge.net/mail/?group_id=1365
>

-- 
Keith Webb
ke...@pl...
Cycle Cobb - Promote Cycling in Cobb County