Wow... That's some *quality *engineering. Very cool stuff.

Boosting this quality contribution up in the Hot Listing - declared quality by lucky760.

Adding video to channels (Engineering) - requested by lucky760.

So the next step would be to apply some sort of grammar reorganization to the translation. This is already being done by some text translation software so it shouldn't be too hard.

Ultimately, you should be able to grab phoneme samples on the fly from the speaker and reproduce the translated speach in the speakers own voice. This will be a really cool technology one day.

Comment hidden because you are ignoring dag.(show it anyway)

Cool technology - poor presentation.

Psh, Star Trek had this years ago.

Reminds me of the Google Wave robot Rosy (for Rosetta stone). They showed it working in the Wave demonstration all that time ago, but for text only.

Pretty cool. More work for Exchange admins. Yay

I love that they are working on this. It already looks good enough to get through basic conversations. It's been a while since I've seen the current state of voice recognition, and they seem to have made great strides. You can still see some mistakes, but the guy is talking at full speed, and the system is picking up a lot of it accurately. Just a few years ago you would have gotten nothing but garbage at conversational speed. Custom voice recreation would be cool, but not terribly important. I'd be more interested to see how translation programs are coming along on that pesky problem of idiom translation?