We’ve worked on merging the native code base of both windows and osx to one project. This should make it easier for us to line up the features between the windows & osx native code: new features / bugfixes should be deployed faster in the future

These are some of the key points in this release:

• OpenNI support on windows. You can choose to use the MS SDK or OpenNI for you AIRKinect projects on the windows platform.
• General bugfixes: fixed some issues with wrong joint mapping & data corruption.
• Multithreading: We’ve integrated boost threading on the windows platform too now, the kinect device now runs in a seperate thread in windows too.
• Support for alternative depth camera’s using OpenNI (ex: Asus Xtion Pro) on both windows and mac.
• Support for settings the camera angle & near mode when you use the MS SDK version.

if(device.capabilities.hasCameraElevationSupport) {
	device.cameraElevationAngle.value = 27; //value between -27 & 27
}

var settings:KinectSettings = new KinectSettings();
if(device.capabilities.hasNearModeSupport) {
	settings.depthEnableNearMode = true;
}

Using nearmode, you will not have skeleton tracking, only user tracking. This is a limitation of the MS SDK.

There’s one catch: we had to create separate ane files: if you want to use the ms sdk driver on windows, you will use airkinect-2-core-mssdk, in the case of openni you will use airkinect-2-core-openni. On osx it doesn’t matter which ane you use, they both have the OSX version compiled.

Same as before:

• You can find the core library (with the ane) and the documentation at http://as3nui.github.com/airkinect-2-core/
• The examples are at https://github.com/AS3NUI/airkinect-2-examples
• You can report bugs in the ane on https://github.com/AS3NUI/airkinect-2-core/issues
• You can report bugs in the examples on https://github.com/AS3NUI/airkinect-2-examples/issues

Happy coding!

With our experience on developing AIRKinect 1, and the feedback we got from the community, we went to work on AIRKinect 2.0. The API is completely different from AIRKinect 1.x, but we hope it will be a lot easier to implement.

There are also a few new features which made it into this update:

• extended point cloud options: setting density, defining point cloud regions
• changing mirroring of camera images & skeleton data while the device is running
• User center-of-mass information
• Basic user management in the ane (before you needed the extended libs to do this)
• shortcut getters to get easy access to some general joints (head, leftHand, rightHand, …)
• compatibility with OSX Snow Leopard (10.6)
• infrared camera support (OSX)
• joint orientation information (OSX)
• multiple kinect support (Windows)

We looked at other builtin API’s (Camera, Accelerometer, …), and try to mimic their way of usage. This way, working with the AIRKinect API should feel very familiar for Actionscript developers.

First of all, you can check if there’s a Kinect device available on your system:

if(Kinect.isSupported())
{
}

Within that if statement, you can get a reference to a Kinect camera, apply some settings and event listeners and start the device. For example, if you want to access the depth camera stream, you would use:

kinect = Kinect.getDevice();
kinect.addEventListener(CameraImageEvent.DEPTH_IMAGE_UPDATE, depthUpdateHandler);
var settings:KinectSettings = new KinectSettings();
settings.depthEnabled = true;
kinect.start(settings);

We won’t go through the entire api here, as this would take us too far. We’ve created new examples, where you can see the new API’s in work.

We’ve also created a short screencast, showing you how you can setup your project (Flash Builder 4.6) to use airkinect 2 & how you can use some of the basic api’s (camera access & skeleton tracking):

Let’s wrap up this post:

• You can find the core library (with the ane) and the documentation at http://as3nui.github.com/airkinect-2-core/
• The examples are at https://github.com/AS3NUI/airkinect-2-examples
• You can report bugs in the ane on https://github.com/AS3NUI/airkinect-2-core/issues
• You can report bugs in the examples on https://github.com/AS3NUI/airkinect-2-examples/issues

Happy coding!

So big thanks to the IntelliJ team we couldn’t be making exciting innovations without great IDE’s!

Download info is here at the forum announcement for AIRKinect 1.7.1 

One new feature will be the option to specify a density for the point cloud. This way you can lower the dataset – for example, only get each 4th point instead of all points – and get faster running results (although it runs fast in release builds). We’re also looking at adding RGB information to the point cloud, so you don’t see white dots, but you see the dots, matching the RGB colors from the camera.

We’d also like to add interactivity to the point cloud. For this, we’re adding a feature called “point cloud regions”, which enable you to specify 3D regions, where you can track the number of dots. This counting will happen on the C side of the extension, so it doesn’t slow down your actionscript code. With this point count, you could add triggers as soon as a certain amount of points is reached within a region. This way, you don’t need to do full skeleton tracking to create interactive applications.

I hope you like what we’ve got lined up!

The release has been tested with the XBOX hardware and version 1 SDK, but not with the run time installer which you can follow us on twitter for an update with the results. Thanks for the support everyone and we are really excited about version 2. Version 2 will  fill the gaps between Microsoft SDK and OpenNI and provide a great development environment for NUI.

Grab the compiled sources - https://github.com/AS3NUI/airkinect-1-release and do not forget to check out the download section for ready to install examples.

So without any further delay here is the link to our github page, hop on over and check em out!

https://github.com/AS3NUI

We’re completely rewriting the extension from scratch, to give you guys a faster & easier to use library. There’s going to be some new features aswell, such as the player masking I showed you earlier, and joint rotations. Joint rotations make it easier for you, to control the bones inside a rigged 3D model, or write pose detection algorithms.

I’ve uploaded a quick demo of this working, together with away3D, to keep you entertained while we continue our work

We’re finishing up a new release for AIRKinect, which will add coordinate translation to the OSX version, together with a new feature called “player masking”.

Player masking will enable you to extract the player’s RGB pixels. Think of it as a green-screen-without-green / background removal. There will be a new CameraFrameEvent type PLAYER_MASK, that will be dispatched when you enable player masking in the extension. This will give you a bitmap image, with the masked rgb pixels.

Check out the screencast below, to see a quick demo of this feature. We’re wrapping everything up right now, and will release this update on our github page very soon.

As promised today we are releasing AIRKinect v1.6 along with AIRKinect-Extended v1.0. We are really excited about both of these releases as we have added a lot of great functionality to the core extension along with a great collection of helpers and utilities in our Extended library.  There are some features still being ported to OSX but we are confident those things will be ported and supported across both platforms.  We can’t wait to see what the community can do with these tools! For Downloads and Documentation please head on over to the AIRKinect Public Repo.

Updates for AIRKinect include

• Skeleton to RGB & Depth Space Translation (windows only)
• Per Pixel Depth information via ByteArray (point cloud)
• Renaming & Restructing of Classes
• Updated to AIR 3.1
• Kinect-less Support. Proper Handling when No Kinect is connected.

AIRKinect Extended v1.0 includes

• AIRKinectManager for managing Skeleton Add/Remove/Update along with Camera Updates.
• 3D Regions
• Basic Velocity Based Gestures with Manager
• ManagedSkeletons with history
• PointCloudHelper for saving Point Clouds in standard formats.
• SkeletonSimulator for Recording and Playback of SkeletonXML
• UI
• UIManager to handle Cursor event dispatching
• MouseSimulator translates mouse into a cursor for testing or Web demos
• UI Components
• SelectableHandle (cursor snapping & timer based selection)
• Target (rollover behavior & timer based selection)
• HotSpot (Rollover based selection)
• SlideHandle (cursor snapping, slide progress & slide based selection)
• RepeatingSelectableHandle (cursor snapping & repeatable timer based selection )
• CrankHandle (cursor snapping & rotation based endless scroll)

 ***UPDATE***

Hey guys, thanks so much for all your comments but if you can move any questions, support, concerns, discussions over to the new forums site, http://forums.as3nui.com, it will make things much eaiser to manage. We will likely be shutting down comments on posts like this. Thanks again!