Microsoft has just announced another upgrade to the Kinect SDK and among the new features is support for Windows 8 and access to more sensor data.
The new SDK is welcome, although it is becoming difficult to keep up with the changes - fortunately, everything seems to be backwards compatible.
Microsoft is making a point of encouraging people to upgrade by saying they have nothing to lose.
However, before you do, here is one reason why you might not want to.
The most important single change is that you now have to use Visual Studio 2012, which many regard as a gray boring replacement for 2010. You can still use the free Express version but it has to be Visual Studio Express 2012. Note: Express 2012 will run on both Windows 7 and Windows 8.
Microsoft has pointed out that the SDK will work with Visual Studio 2010 or 2012. The system requirements are detailed at:
So there is no need to spend time considering the upgrade - just do it.
The big news, if you live in China, is that with the release of the new SDK you can now go and buy a Kinect sensor. Given the number of potential developers China could offer the Kinect, the rest of us had better start working faster.
Microsoft promises to add seven more countries: Chile, Colombia, the Czech Republic, Greece, Hungary, Poland, and Puerto Rico, very soon.
According to the Windows Kinect Blog, the key improvements to the SDK, from the developer's point of view are:
- Data from the sensor's 3-axis accelerometer is now exposed in the API. You may not have noticed that the Kinect has an accelerometer but now you can start to think up clever uses for it. Adding dead reckoning positioning for the sensor and general orientation information could be used to improve mobile data gathering.
- Extended-range depth data now provides details beyond 4 meters although at a lower accuracy.
- Color camera settings, such as brightness and exposure, can now be set explicitly by the application, allowing developers to tune a Kinect for Windows sensor’s environment.
- The infrared stream is now exposed in the API. Yes this means you can now use the IR camera directly. What could you use it for - how about a see in the dark application?
- You can now change the IR status faster so that you can support multiple overlapping sensors
The developer tools, specifically Kinect Studio, have also been updated. There is now a German speech recognition language pack and you can track skeletons using multiple sensors.
The new Windows 8 compatibility mode is clearly a big deal for Microsoft, but it is difficult to gage how immediately useful this will be. You can only create desktop applications, so really this isn't much of a story as the difference between Window 7 desktop apps and Windows 8 desktop apps is small.
A slightly odd addition is the ability to use the Kinect with a copy of Windows running on a virtual machine - Hyper-V, VMware and Parallels. Given speed is a major problem for most Kinect applications, it isn't clear that running on a VM is going to be useful.
Ben Lower of Microsoft adds:
This is for two primary purposes:
a. Using a linux or OSX machine as the primary dev box -> now these developers can create & spin up a Windows VM and still develop Kinect for Windows experiences using their primary dev machine
b. More testing options -> VM support means developers have more options for automated testing as well as testing across different machine configurations