Eric M Russell (6/4/2012)
I understand that the goal is to translate voice commands into web search or iPhone commands, but I don't see why it's necessary to steam the actual voice over to one of Apple's servers. Not only is that a privacy issue, but I would expect it to result in latency or loss of service issues as well.
A 3rd party could implement a client side voice recognition solution that does the same thing but only better. They wouldn't have to invest in server or operational infrastructure; just sell the widget for $5 a pop.
There is really not that much processing power on the phone. Voice recognition systems like Apple or Google have vast arrays of sound matching samplings and do considerable processing of the signal. Even a notebook, which is vastly more powerful than an iphone, does not match that performance.
-- FORTRAN manual for Xerox Computers --