Share via

Frequently Asked Questions

  • Article

Kinect for Windows 1.5, 1.6, 1.7, 1.8

This page answers frequently asked questions about using Kinect for Windows.

  • Audio
  • Driver Installation
  • Depth image artifacts
  • Recommended Hardware Configuration for Kinect
  • Terminology for Kinect and NUI

Audio

When running the MicArrayEchoCancellation sample, you should be streaming data before you start the capture loop. For example, you could play a music file on Windows Media Player.

Driver Installation

If driver installation fails, here are some things that you can try:

  • Unplug the Kinect. Reconnect it.
  • Make sure that the sensor is connected to a functioning USB port. For example, connect another USB device such as a mouse that you’ve already used on your computer.
  • If the Kinect is plugged into a USB 3.0 port, plug it into a USB 2.0 port instead.
  • Make sure the Kinect is plugged directly into a USB 2.0 port, and not through an external USB hub.
  • Since a Kinect requires at least 50% of the USB bandwidth available, make sure that the Kinect does not share the USB controller with any other devices.

Make sure to uninstall any previous driver (including any beta Kinect for Windows SDKs). To do this:

  • Unplug the Kinect sensor.
  • Select the Kinect drivers in Programs and Features in Control Panel and click Uninstall.
  • If you have non-Microsoft drivers, follow the instructions from the driver’s manufacturer or check your system’s registry to ensure than any previous driver has been removed.
  • Uninstall previous versions of this SDK driver in Programs and Features in Control Panel. Then reinstall it from the original SDK download package.
  • Plug in the Kinect sensor again.

Depth image artifacts

The depth stream images have periodic vertical line artifacts and also radial band artifacts. The magnitude of these artifacts is small enough that they won't interfere with coarse-level operations, such as skeleton or other object tracking, but they will affect attempts at precise measurement of object dimensions more significantly.

What is the recommended hardware configuration for intensive processing with Kinect?

For applications designed to make intensive use of Kinect Skeletal Tracking in addition to custom application processing, we recommend using the following PC configuration as a reference.

We tested it, and found that it can provide great performances in the most challenging tracking scenarios, while still providing room for applications to do additional processing and maintaining optimal frame rate.

  • Processor Type: Intel Core i5 Desktop Processor Series or superior
  • Clock Speed: 3.0 GHz or superior
  • Number of Cores: 4 or more
  • Memory: 4 GB DDR3 1333 or superior
  • Operating System: Operating System: 64-bit version of Windows 7 or 64-bit version of Windows 8

Terminology for Kinect and NUI

This section contains terminology that is used in the Kinect for Windows SDK.

  • beamforming algorithm
    An algorithm that determines the direction of the sound source in the horizontal plane.
  • CDRP
    Color, depth, and pixel registration.
  • depth stream
    The data that is produced by the depth image camera in the Kinect sensor. Each frame in the stream contains the distance, in millimeters, to the nearest object at a particular x and y coordinate in the depth sensor's field of view. Two depth data streams are available, one with a frame size of 320×240 pixels and another with a frame size of 80×60 pixels.
  • Kinect sensor
    A physical device that contains cameras and a microphone array. The sensor connects to a PC by a USB cable, although Windows Plug and Play does not identify it as a USB device.
  • latency
    The required time for isochronous data to travel from the USB isochronous device to the client application. Many factors can affect latency, such as the video frame format (uncompressed, compressed, and so on), the size of the video frame, the data rate, and the CPU speed.
  • natural user interface (NUI)
    An evolving model for human-computer interaction that is context-appropriate and adaptive. Because the NUI exploits a user’s existing skills and expectations, it is easy to learn. A NUI might incorporate speech, gesture, touch, or location, depending upon the application and the user’s environment.
  • USB isochronous interface
    An interface that supports four USB data flow types: isochronous, control, interrupt, and bulk. The Kinect sensor has an isochronous interface, which is typically used for audio, video, and other streaming data from cameras and similar devices. The interface provides constant bandwidth that is reserved exclusively for a particular device.
  • YUV stream
    A stream that contains digital video data in which the three components are Y’ (luma, which represents brightness) and two chroma values (Cb and Cr), which measure color difference. Because the human eye is more sensitive to differences in brightness than in hue, the chroma values are sampled at half the rate of the luma values, so a YUV stream contains two Y’ values for each Cb and Cr. Consequently, at 8 bits per pixel, YUV data is more compressed than RGB data because only 4 bytes are needed for each two pixels. YUV color provides 16-bit, gamma-corrected linear UYVY-formatted color bitmaps. For more information, see About YUV Video in the MSDN library.