Kinect for Windows SDK And Runtime Now Available

As promised, version 1.0 of the Kinect for Windows SDK is now available from Microsoft and beyond the usual bug fixes, etc, is actually a lot of new features that take advantage of the new hardware advantages of the motion tracker’s latest model.

There is only one Kinect that can be hooked to an Xbox 360 at a time, but developers know that a Windows machine is capable of handling much more. And so there is now support built in for four different sensors at one time. This will mean a lot to 3D modelers who will use the hardware to create new items. Having a perfect view from two or more sides to compile together into an object will be very useful indeed.

How the Kinect interacts with the user is also at the forefront in the new runtimes. Speech recognition is improved through a new acoustic model that can hear voices better and thus provide more accurate understanding of what it hears. It also features the newest Microsoft Speech v11 components in the SDK and binaries. And the device can now see closer than before, which is also represented in the routines to make it easier to recognize things like hand gestures 16 inches away.

Even though they say that the Kinect for Windows is not really about gaming, the new skeletal tracking system is certainly going to help in that regard should game companies want to jump into the fray. One of the problems that the previous version’s developers faced was knowing which user was actually being “seen” As you can imagine, this can cause a lot of hardship on the programmer to try to make that determination themselves. Fortunately for them, Microsoft realized this and has answered this issue.

There are other improvements across the entire scope of the API too numerous to mention here. Probably things that developers probably weren’t aware were broken or missing that they’ll come to appreciate later on. One highlight is that since the SDK can be used commercially, MS was nice enough to include a full installer that can be included with an application so that users don’t have to go through a big hassle to get a program running on their system.

And for those who are new writing for the Kinect or unaware of how to do something, there are new and improved samples that will provide instruction and even the actual code to handle a variety of tasks from the sensor like where sound is coming from, how to process color and depth from the camera, motor controls and the user tracking we mentioned above.

All in all, it sounds like some great stuff and MS isn’t through. They say that is only the latest in what will become 2-3 updates a year to improve and add new features as they are warranted. I’ll say one thing. They sure seem dedicated to this course. And as I said before, I can’t wait to see what people do with it.