For IT managers, the words “video conferencing” can be frightening.
Design guides call for 1.5 to 15 megabits per second of dedicated bandwidth per participant, with tight limits on latency, jitter and packet loss. Not all networks can meet those high barriers to entry, especially when international circuits or public services form part of the wide area network.
Yet people hold video conferencing sessions using consumer-grade software, DSL and cable Internet connections, and inexpensive web cameras. They may not be thrilled with the quality, but they’re collaborating successfully and getting work done.
Many organizations can find a middle ground. With some attention to detail, IT managers can improve video and audio quality without expensive engineering.
The quality of a video conferencing session depends on many factors, and the network is only one of them. Good room design, proper lighting, light control and attention to detail in areas such as sound quality all add to the video conferencing experience.
Even when users run conferencing sessions from their desktops, notebooks and tablets, a little bit of education and some inexpensive equipment and room adjustments can make a big difference.
For example, using an external microphone rather than one built into the device (probably too near a noisy fan and filled with dust) improves sound quality and makes everyone more understandable. Giving PC users USB headsets with built-in microphones reduces distractions and adds to overall comprehension even more. Teach users how to set up room lighting, to avoid undraped windows in the background and to pay attention to the audio and video signal they’re sending.
Video conferencing traffic tends to be easily identified at the LAN edge, usually by some sort of differentiated services packet marking. Call setup using H.323, Session Initiation Protocol or a proprietary protocol isn’t the issue; it’s the real-time video and audio traffic that needs attention.
Prioritize outbound real-time traffic on the wide area network (WAN) using edge routers and firewalls. Within a building, employ quality of service (QoS) engineering to separate and prioritize voice and video traffic.
Any prioritization has to work both ways. When WANs are built on public services such as the Internet, this calls for configuring consumer-grade DSL or cable modems in branch and home offices to enable QoS prioritization and send voice and video packets to the head of the queue. For private multiprotocol label-switching networks, the service provider must take part in ensuring prioritization or bandwidth allocation across the WAN.
Finally, focus on the video conferencing hardware and software configuration. Most conferencing tools have a setting to automatically adjust audio and video quality to compensate for network conditions; this should be enabled. In cases where network bandwidth is perennially in short supply or of low quality, setting the system to always use a lower speed gives a more predictable performance, which leads to higher user satisfaction.
Where multiple CODECs for video and audio are available, experiment with different CODECs rather than simply taking the defaults. The defaults are usually configured for a bandwidth-rich environment rather than a network-constrained WAN. Sites with diverse network characteristics usually require different settings, not a one-size-fits-all approach.
When network characteristics don’t meet high-end specifications, an immersive telepresence experience may not be possible. But a little attention in human factors and network engineering definitely can improve quality and user satisfaction at a very low cost.