Kinect for Windows

What Kinect for Windows is

Kinect for Windows is a Microsoft product that packages the Kinect sensor — a camera bar that measures the distance to objects in a room — together with a software development kit (SDK) and runtime for Windows computers. It became officially available on 1 February 2012, and Microsoft set a suggested retail price of 249 US dollars for the hardware and software package ^[1]. Microsoft framed the release as opening the device to commercial use, that is, to organisations that wanted to build their own applications around the sensor rather than use it as a finished consumer product ^[1].

How the launch was organised

At launch, Microsoft said its distribution partners in twelve countries would begin shipping Kinect for Windows hardware ^[1]. Those countries were the United States, Australia, Canada, France, Germany, Ireland, Italy, Japan, Mexico, New Zealand, Spain and the United Kingdom ^[2]. The software side was handled separately: version 1.0 of the Kinect for Windows SDK and the matching runtime were made available for download on the same day ^[1]. A developer therefore did not have to wait for a boxed unit to arrive before installing the toolkit and starting to write code.

What the toolkit added

The release added a feature Microsoft called Near Mode, which let the depth camera see objects as close as 40 centimetres in front of the device ^[1]. A depth camera reports a distance for each pixel in its field of view, and it can only do so reliably once a scene is far enough away; anything closer falls outside the measurable range. Near Mode shortens that minimum distance, so hands, small parts or objects held just in front of the sensor stay in view. That matters for desk-based applications, where the user sits close to the device instead of standing across a room.

Speech was the second area of work. Microsoft said the release included the Microsoft Speech components V11 and an improved "far talk" acoustic model that increased speech recognition accuracy ^[1]. Far talk refers to the case where the microphone is some distance from the speaker's mouth rather than mounted on a headset, so the captured audio is quieter and carries more room echo; an acoustic model trained for that condition is what makes voice commands workable for someone sitting or standing a metre or two away.

Microsoft also said the SDK supported up to four Kinect sensors plugged into the same computer ^[1]. Several sensors on one machine let an application cover a wider physical area, or look at the same scene from more than one angle at a time.

Developers and pricing

At the time of release, Microsoft said more than 300 companies from over 25 countries were building applications with Kinect for Windows ^[1]. That figure describes organisations already working with the toolkit around launch rather than products finished and shipped. Microsoft also said it would offer academic pricing of 149 dollars for qualified educational users later in 2012, lowering the entry cost for universities and research groups ^[1].

Where it sits

What distinguishes Kinect for Windows from the sensor alone is the arrangement around it: a published list price, a toolkit numbered 1.0, a documented runtime, an explicit commercial-use framing and a separate education price ^[1]. A version number starting at 1.0 signals a product Microsoft intended to maintain and revise rather than a one-off release, and a fixed price plus distribution partners in named countries made it something an organisation could budget for and order through ordinary channels ^{[1] [2]}. The features singled out at launch — short-range depth sensing, speech recognition at a distance, and support for several sensors on one machine — describe the kind of applications the product was built for: kiosks, desks, classrooms and research setups where people interact with a computer without touching it.