Thanks Stefan,
My intent is actually to optimize the second part of non-static
requests. This test does indeed could used pre-generated voice files
(per the cached remark), but the intent is to identify the best way to
handle the dynamic parts.
Is this the best way to handle dynamic tts parts?=20
-D
> -----Original Message-----
> From: ast...@li...=20
> [mailto:ast...@li...] On=20
> Behalf Of Stefan Reuter
> Sent: Tuesday, February 28, 2006 1:59 PM
> To: ast...@li...
> Subject: Re: [Asterisk-java-users] Text-to-speech and=20
> gathering data (fastAGI)
>=20
> Darren,
>=20
> it looks like you are always using TTS to generate the voice=20
> files for your application even if they are rathe static. I=20
> would only run the TTS once (outside of your AGI scripts) and=20
> then use the generated voice files. If you have dynamic parts=20
> (like greeting the user with his name) I would only render=20
> those parts at run time. This should eliminate the delay and=20
> take away some load from your Asterisk server.
>=20
> =3DStefan
>=20
> Darren Hartford wrote:
> > Hey all,
> > Continuing my mad plan for using asterisk-java & fastAGI...taking=20
> > text, convert to voice, and then gathering data back.
> >=20
> > The below works, but there is a 3-sec delay for the=20
> conversion (which=20
> > could be cached for future use, but focusing on the=20
> one-offs for now):
> >=20
> > =3D=3D=3D=3D=3D=3D[code]=3D=3D=3D=3D=3D
> > String texttospeech =3D "Please enter your favorite number=20
> > between 10 and 100 followed by pound sign.";
> > int soundhash =3D texttospeech.hashCode();
> > String sounddir =3D "/var/lib/asterisk/sounds/tts"; =20
> > String textfile =3D sounddir + "/ttsTEST-" + soundhash=20
> + ".txt";
> > String wavefile =3D sounddir + "/ttsTEST-" + soundhash=20
> + ".wav";
> > String gsmfile =3D sounddir + "/ttsTEST-" + soundhash + =
".gsm";
> > result =3D exec("System","echo '" + texttospeech + "' > " +=20
> > textfile);
> > System.out.println("text save result: " + result);
> > result =3D exec("System","text2wave -F 8000 -o " +=20
> wavefile + " "=20
> > + textfile);
> > System.out.println("text2wave result: " + result);
> > // result =3D exec("System","sox " + wavefile + " -r=20
> 8000 -c1 " +
> > textfile);
> > result =3D exec("System","sox " + wavefile + " " + gsmfile);
> > System.out.println("sox result: " + result);
> > =20
> > String captureddata =3D "";
> > captureddata =3D getData("tts/ttsTEST-" + soundhash);=20
> > =3D=3D=3D=3D=3D[/code]=3D=3D=3D=3D=3D=3D
> >=20
> > Any recommendations to get this streamlined for one-off text=20
> > conversions, or any pitfalls?
> >=20
> > TIA,
> > -D
>=20
|
