Three competing speech recognition and translation technologies are vying to participate in the ‘TRANSTAC project,’ which will take a soldier’s English, recognize the speech, translate it into Pashto and then spit it out using text to speech. All of this happens in a matter of seconds using wireless Internet and servers somewhere else in the world to do the heavy lifting.
Then, a few seconds later, doing the reverse to complete the communication loop. NIST says that human translators can tire, but these devices obviously don’t.
You can see one of the entries is using a Google Nexus One phone outfitted with a push to talk type of speaker/mic system and likely running Android (above, right).
The demo isn’t complete without watching the video with splices of Star Trek clips and standard corporate communications music tracks, below.