Panther and Vicki
Reading this forum and various other forums, I see that there will be a new high quality voice added to Panther called Vicki. I am unable to speak due to a stroke at the age of 13 and I use Apple's text to speech to speak with friends over the phone using the Fred voice (and now that I own an iSight, as of last week, it would be nice to use text to speech to communicate with people I video chat with). Needless to say that I am very much interested in Apple's text to speech abilities. I saw that someone wrote that not even Vicki compared to voices that AT&T uses. The person left the URL of the demo to hear the AT&T voices. Another person had a link to a demo of Vicki. They are right, AT&T's voices sound better than Vicki or any of the other text to speech voices that Apple uses. Some were speculating that Vicki's huge file size means that some time down the line that Apple will be upgrading their text to speech abilities either in one of the incremental upgrades of Panther, or maybe it will mean a new iApp.
I would like to hear other people's opinions on where Apple's text to speech abilities, or rather, technology is going.
I would like to hear other people's opinions on where Apple's text to speech abilities, or rather, technology is going.
Comments
http://www.rhetoricalsystems.com/
http://festvox.org/
http://www.cstr.ed.ac.uk/projects/festival/
http://verdens.navle.no/musikk/fitter.vicki.aiff
Vicki sounds like a slower, slightly smoother version of Victoria, and neither of those sound as lifelike as the AT&T voices to me.
Hope this helps. It's not really that great, it just seems to reduce the occurrance of some of the strange noises you hear when the computer is talking.
EDIT: Beat me to it!
Originally posted by Placebo
but it's almost to the AT&T level.
Hardly. I've heard other voice-synthesizing software that sounds FAR better than this. Apple has made just marginal improvements to the voice system. Most of the same old problems in Macintalk are still present. It just sounds like Apple's using a higher sample rate or something of the like. Bleh.
Vicki likes Radiohead.
FestVox version 2.0: Jan 2003: new in this release
Better clunits general voice support
Support for CMU Sphinx and SphinxTrain to build acoustic models for labeling
DOCBOOK version of the documentation, with more general backgfround documentation
Initial support for Mac OS X
configure support to match Edinburgh Speech Tools
Interesting, I wonder if there work will be better then Vicki, or is Apple using this open source? I hope they start using more and more open source to save money and build up a good rep.
Anyone have any samples of other, better, solutions?
http://www.naturalvoices.att.com/demos/
Try making that white voice sound street.
Much fun.
If I remember (lol), I'll probably set that to default when Panther comes (even though i don't use Speech a lot anyway...)
If they could just get rid of those weird "pops" in the voice, it would sound truly impressive - especially for purely synthesized speech, which the AT&T technology is not.