Unified Communications products like Microsoft Office Communications Server 2007 make it pretty simple to integrate the video experience into a user's daily routine, requiring only off-the-shelf Web cameras to layer on the new communications channel.
But what quality of video are you really getting with this kind of integrated solution? Will it meet your needs and expectations?
In my tests of Office Communications Server, I learned from Microsoft's Quality of Experience Monitoring Server that video calls use Microsoft's RT Video codec. By default, I found person-to-person calls had a 352-by-288 resolution at a frame rate of 14 frames per second--when the call is placed over a LAN.
Qualitatively, the video picture looks fine in the small Office Communicator box that is normally shown on the screen. But when blown up to full screen size, I could see some slow transitions and artifacts, and I could definitely tell that the lip synchronization of video and audio was not that great.
The video quality is certainly not up to the standard of high-definition audio we get from Office Communications Server, which uses a wideband RT Audio codec on fast network connections--and sounds excellent and clean. But again, the video quality is not too bad on a small screen, especially if you don't come to the game expecting the best quality.
On the other end of the spectrum, there are some really fabulous high-definition video alternatives out there that also rely on software rendering--not hugely expensive dedicated A/V rendering hardware. But these software solutions come with their own kind of costs.
Take for example the HD video experience offered by GIPS (Global IP Solutions)--which has HD video capabilities in both its two-way VoiceEngine products and multiparty ConferenceEngine line--and uses both its own proprietary LSVX codec as well as standard codecs like H.264. Global IP Solutions first demoed at the Fall VON conference in 2007, and I got to see it up close in person last week at the company's offices in San Francisco.
In my demo, the video stream--at 30 frames per second--had a resolution of 960 by 720. This translated to a truly stunning picture--so clear that I was literally able to count the bricks in the side of a building half a block away when we pointed the HD video camera (a pretty high-end Sony HD camera, by the way--not some Webcam) out the window. And the lip synch between audio and video was practically perfect, making it much easier to carry on a conversation without getting distracted by slightly out-of-sync behavior.
The company claims it can scale up to a full HD picture as well.
Of course, the tax in this case is computational. During the demonstration, the quad-core server doing the rendering on my end of the call was clocking in at a hefty 55 percent overall utilization--something that would be even higher for full HD. The company claims to have done significant work to optimize its rendering for Intel processors, and it claims testing on AMD platforms will also be done in the coming weeks, with the expectation that rendering performance will at least be in the same ballpark.
GIPS sells its own products, or you might find its technology in other products. For instance, I know that Toktumi is working on integration with GIPS' REX softphone (which I will be reviewing soon), and yesterday, RADVision announced that it will be using GIPS codecs and features from the VoiceEngine platform as well.