Desktop video conferencing is moving into the mainstream as enterprises look for ways to save money, reduce their carbon footprint and better utilize their workforces. At the same time, many organizations are investing in high-definition, room-based video conferencing systems, which raises questions about how to extend HD support to the desktop. Do you have high-quality Web cameras? What are the PC requirements? How about the bandwidth demands?
HD requires cameras that can support higher resolution, but it is more complicated than that. The main number manufacturers use to market their products is the resolution when the camera is in a still/photographic mode and not in motion video mode. In general, as the video resolution goes up the frame rates drops dramatically. What you need to look for is high resolution, such as 960×720 or 1280×720, along with a frame rate approaching 15 to 30 frames per second (fps).
Additionally, a glass lens is preferred over a plastic one because it minimizes distortion and will provide more image detail and clarity. With HD the quality of the lens is significantly more important than with standard-definition video because the goal is to have the highest quality source image with the least amount of optical artifacts. The video codecs won’t waste cycles encoding and decoding irrelevant pixels.
Desktop video will likely be used at home and on the road where the lighting is typically not as good as in an office setting. The camera’s gain control should be able to increase image brightness and remove shadows from a face without overexposing the image. Another important feature is auto-focus because desktop users tend to move around in front of their PCs and many cameras have a limited distance range where the focus is accurate, especially in situations with less than optimal lighting. What’s more, everyone will appreciate the fact that remote users won’t have to constantly place their fingers in front of the camera to focus.
Core infrastructure requirements
In terms of PC support, H.264 is the video codec that most standards-based conferencing endpoints support. Before HD, the typical resolution for desktop video conferencing was either QCIF (176 x 144) or CIF (352 x 288), with some applications also supporting a second video stream of presentation data. A single processor system (2.0GHz CPU) could support this level of video quality for both send and receive. HD 720p (1280 x 720) is roughly nine times more resource intensive than CIF — there is simply nine times more data to encode and decode.
What these increased processing requirements you need at least a dual-core CPU (2.0GHz). One core encodes the video while the other decodes, leaving enough CPU to process audio and other PC tasks. A typical dual-ore system will have enough horsepower for receiving HD at 30 fps, but not quite enough to encode or send HD video at 30 fps. Most dual core systems will top out at 15 to 30 fps. Achieving 30 fps of encoded HD video requires a quad core machine – two cores for encoding, splitting the workload in half, and one core for decoding.
Although it is not necessary to have the most advanced processor in your PCs, to achieve the highest level of performance, product vendors will optimize their code using the advanced instruction sets offered on the latest generation processors as a standard development practice. In terms of memory, 2GB should be sufficient for most applications.
On the bandwidth side, most video conferencing endpoint vendors recommend 1.5Mbps for HD video. In practical use, HD requires a minimum of 768Kbps for very good quality, which becomes excellent at 1.5Mbps. In comparison, H.264 CIF video provided very good quality at 256K to 384Kbps. The difference is the amount of information with the higher resolution HD stream.
Although HD video to the desktop doesn’t require anything new for audio devices, it’s still helpful to review the best practices in this area. Many PCs and Webcams have integrated microphones, which are usually not recommended because they tend to be poor quality and pick up background noise, keyboard use and don’t have echo cancellation. However, there are some new notebooks and Webcams available that have array microphones and have echo cancellation and noise reduction capabilities built-in. As these are new to the market, you should test them with your specific conferencing application to see how they perform.
Experience has shown that you can’t go wrong with either a headset or one of the commonly available USB combination speaker/microphone devices. The inherent design of a headset eliminates echo, and the USB devices have built-in hardware echo cancellation. The most common problem observed with audio devices is the selection of the wrong device for recording (microphone) and playback (speaker) in the conferencing application.
When external audio devices are connected to a PC, Windows detects multiple devices and the conferencing application doesn’t necessarily choose the correct one. This problem is apparent with USB devices with integrated echo cancellation. If the same device is not selected for recording and playback, echo cancellation doesn’t work as well on some applications because it doesn’t know what audio is correct and what is echo.
The requirements above are general guidelines and you should verify the requirements of your specific conferencing application. The quality that HD brings with it ushers in the potential for new levels of acceptance and productivity for desktop video conferencing. If you have not experienced one of the new applications, HD deserves a try.
Wallman is director of product marketing at Radvision.
