Xbox, Listen! Kinect Audio Explained
Posted on Tuesday, March 1 @ 18:27:55 Eastern by Nicholas Tan
Of all the touted features of Microsoft's Kinect, its audio recognition is among the least used, especially when it comes to driving the gameplay of a video games. Even the novelty of being able to bring up the Kinect Hub dashboard using voice has worn off. But developers are hard at work trying to cultivate more tools and programs to incorporate voice commands seamlessly into games. Voice seems so intuitive and natural that it would seem wasteful to exclude sound.
The Kinect has a microphone array that is essentially an "open mic" that has speech recognition ideally within the Kinect play space. It also cancels out any echo and video game sounds so that it doesn't end up feeding back into the sound system; effectively, noise reduction. Along with sound suppression to eliminate room tones and hums, this makes player voices clearer in an online conversion. The system's sound position tracking also pinpoints the speaker from, say, a group of players to recognize a specific person's voice commands (as opposed to someone who might be imitating your voice to cheat).
Most of the time, voice chat using Kinect is not passed over by players in favor of an actual headset. On the flip side, the speech recognition is much more magical, particularly when it works right out of the box. By leveraging Microsoft's Speech Technologies, the speech recognition contains lexicons for supported languages, which includes English, UK English, Spanish, and Japanese. (Soon to come: French, Canadian French, German, Italian, and Australian English.) It also supports custom pronunciations for developers to use if they want use 'to-mah-to' instead of 'to-may-to'.
One special note: Since the Kinect tends to listen all the time, it will sometimes reject a voice command to make sure that it doesn't react to any kind of random shout-out; of course, that can lead to frustration. So don't be surprised if a game cues you to say something, having the Kinect listen at the right time. And if the game
doesn't recognize your voice commands, a game that has a character say "Huh?" or an animal like Kinectimals can help reduce frustration. Hey, it's a cat; of course it ignored you.
Until speech recognition becomes as sophisticated as Avina in Mass Effect, you should expect any speech recognition in games to require carefully separated voice commands, unique voice commands to reduce false positives, and plenty of confirmation screens. Voice recognition is still fairly rudimentary right now, but look for it to be one area of innovation that will be mastered within two console generations when I'm... oh, I just made myself sad...
FOLLOW ME ON TWITTER. YOU KNOW YOU WANT TO.
comments powered by Disqus