Bigus Dickus
Regular
Additionally at least Google Voice and Siri draw from a near constant internet connection to leverage a larger pool of data for voice recognition. That is only available on Kinect when using Bing search while connected to the internet, at which point accuracy improves dramatically.
Regards,
SB
Bing search needs to parse any possible spoken word. What I describe, even though I wrote "complete library", can actually operate with a rather tiny subset. What really would need to happen is implementation of the other crucial parts of the nuance engine... editable and customizable library, training, learning. Many people already achieve nearly perfect results with kinect voice commands, and I have my doubts that MS has suddenly become the technological leader in speech recognition software. But I have no doubt there is serious unlocked potential and room for improvement, and I believe MS has in making Kinect work well developed a nice hardware solution for a housewide speech recognition solution.
Look, I use speech recognition all day for my job, and it isn't perfect in ideal situations and it never will be, and won't be close for a long time still.
Didn't stop "good enough" from being an absolute game changer in my field, and it shouldn't stop MS from vastly expanding its scope of vision, product abilities, and market appeal.