Seeing is believing or should we upgrade just for an S?

Posted by:

Until recently I did not really see the value to add video to business calls and business meetings. What made me change my mind was statistics saying that video is the most trusted media, we simply don’t trust text or a voice as much as a video. Somewhat correlated was realizing that most of us want human help when we have a problem, in other word we don’t look primarily in the intranet for FAQs or articles but for experts, human knowledge and talent.

Yesterday I visited a session on advanced video capabilities that IBM will ship in their next version of Sametime.  The session came down to two acronyms that affects all of us pretty much but hopefully without too many of you even notice.  I am talking about SVC and AVC (Scalable Video Encoding and Advanced Video Encoding)

Today Sametime video relies on AVC H.264, that is pretty good but not good enough for you!  From what I understand AVC means that the Sametime server looks at the connecting clients in a meeting/session and then transcodes the signal to the least common denominator client requirement and then sends that video signal to all connecting clients. This requires a lot of server power and limits the result for the meeting/session members that have a good link to the meeting server if there is a mobile with a poor link joining in. The server transcoding also makes it impossible to show more than one video stream in the meeting at once.

With SVC H.264 (which is actually an extension to AVC) the video is instead separated in three dynamic layers from the beginning and the server does not need to do any transcoding and each connecting client automatically chooses the best video-mix for themselves.

The layers are

  • Temporal
    • Fram rates 30 fps, 15 etc
  • Spatial
    • Resolution, VGA, HD, iPhone, 1024/768 etc
  • Quality
    • The fidelity of compression, how close a reconstructed frame is to the original input

For example conference room TVs can request a HD spatial signal with maxed out frame rate and quality while a mobile connecting to the same meeting can use something much “smaller”. The same device might always require the same Spatial signal but automatically change to lover temporal and quality rates when connecting over for example a slow 3G or wifi link.

SVC can also handle bandwidth/link changes during the video session gracefully adapting the video quality automatically down and then up again as soon as the bandwidth gets better again. In addition to that the SVC signal can switch between servers and we can for example have local SVC video servers closer to the end users in for example in Asia.

Another advantage with the lack of transcoding is multi video meetings; the server will be able to show several meeting participants videos at once. Sometime that might be good, it will be interesting to see how this will change the meeting cultures. Can we expect phone meetings with mandatory video to avoid unfocused attendees?

If you want to read more in detail on SVC I recommend Wikipedia

http://en.wikipedia.org/wiki/Scalable_Video_Coding

Perhaps you wonder when we can help you get this important S. First IBM needs to release Sametime 9 (sometime after the summer sounds likely but I wouldn’t be too surprised if it is released earlier than so) and then your Sametime Standard servers needs to get to the new release. ST9 is smart so clients without support for SVC will still be able to join in but with the AVC limitations still there. That’s how we infrastructure geeks want stuff so we can deliver new value without necessarily needing to touch all clients. The Sametime mobile apps will of course get SVC support right after IBMs release!

0

Add a Comment