Historically Microsoft published a really good whitepaper on RTAudio and QoE, but information on how the RTVideo Codec works has been sparse. But at a presentation today at Interact 2008 more details were revealed (and we where promised that more details will appear, as video gets more and more attention).
Besides all the interesting stuff about High Definition video in next versions, I just briefly wanted to touch on a few things that applies to the current versions -
P2P Video Calls with VGA quality requires Dual Core (And Quad for HD in wave 13)
So you probably know that you can change from CIF to VGA quality on P2P video calls (Not conferencing) by changing a registry key. They are documented in the release notes for OC 2007 and referenced here -
To enable VGA video in a point-to-point video call, users need to set the following registry keys on both endpoints -
[HKEY_CURRENT_USER\Software\Microsoft\RTC\Quality]
"MaxAllowedSendVideoSize"=dword:00000002
"MaxAllowedReceiveVideoSize"=dword:00000002
And restart Communicator.
But what is new to me is that the hardware requirements are actually Dual Core processors for optimal quality. It is not enforced, so VGA quality will be started with one processor, but may fall back to lower quality due to lack of CPU resoucres. In wave 13 Microsoft is talking about enforcing this requirement, so VGA will not start if you are not running from a Dual Core. In terms of HD (720P) conferencing the current requirement is going to be Quad Core processors and again the requirement will be enforced.
Roundtable quality requires Dual Core
Dual Core Processors are required for Roundtable Panorama Video but again it is not enforced, so Roundtable and Panorama Vidoe quality will be started with one processor, but may fall back to lower quality due to lack of CPU resoucres).