[Asterisk-Users] Re: Zork and Asterisk

[John Todd]

While I love voice synthesis, I think that you'd gain "legend" status 
only if you had Allison (or, I hesitate to say, some other voice 
talent) do the dramatic readings.  People respond so much better to 
real recordings - I tend to use synthesis only where it is impossible 
to have pre-recorded phrases, and since Zork is pretty much entirely 
"canned"....

Do you have any idea how many text blocks there are in Zork?  In 
sentences, kilobytes, or...?  She charges like $120 USD per hour (I 
think) but she can fit a LOT of talking into an hour if someone else 
is willing to do the editing to split up the continuous recording 
into blobs.  I'd put $150 or so towards the project if you decide to 
do it, but I can't put any time...

You should ping Gerd Graumann <gerd at LumenVox.com> at LumenVox to get 
some demo licenses for voice recognition, and convince him  ;-) that 
this is a good beta of the Asterisk/LumenVox partnership, because I 
think it would be an excellent example of how useful a good voice 
recognition engine would be in general applications.  If you make it 
avaialble at no cost, then people would essentially be trial-ing the 
voice recognition engine and seeing how nice it is for themselves in 
an Asterisk system.

JT


>That's not far off from my actual plans... I've had a server donated to
>me for the project and have had a couple of CLEC's offer incoming lines,
>as well as had a developer license for Cepstral's TTS engine donated.
>
>Another feature I want to add is meetme-like capabilities so that you can
>talk with your fellow players when you're in the same room, maybe even
>connect it up to a mud or a moo or something.
>
>If anyone feels like collaborating, feel free to contact me off-list by
>dropping the -asterisk from my address.
>
>Cheers,
>spd
>
>On Tue, 6 Jun 2006, John Todd wrote:
>
>>
>>  http://www.boingboing.net/2006/06/05/play_zork_by_phone.html
>>
>>  Let me preface this idea with one comment: I don't have the time to
>>  do this - I don't even have time to eat these days.  But someone out
>>  there has the cycles to do this... and it would be very cool.
>>
>>  OK, so now Zork is attached to Asterisk, but using the
>>  less-than-clear Festival engine.  There are beta tests of the
>>  LumenVox speech recognition engine out there which tie directly into
>>  Asterisk.  Allison Smith (the voice of Asterisk) would almost
>>  CERTAINLY do a great dramatic reading of all of the text blocks
>>  within Zork.  I see an excellent opportunity for a demo server on
>>  some CLEC who would love to get some $ by opening up a few DIDs to a
>>  huge recip comp traffic load.  Even if it's just available via IAX2
>>  or SIP, this would be one of those "legends of the Net" in the next
>>  few years...
>>
>>  JT
>>