Scalable video coding (SVC)

SVC (Scalable Video Coding) is a video conferencing architecture based on scalable video coding. Here’s how it works:

1. The client forms an SVC stream by compressing video stream layers by layers and sends it to the video conferencing server. The number of layers in a video stream to be sent depends on the width of the connection channel and bandwidth.
2. The server processes obtained SVC streams from all the clients and removes extra layers without recording.
3. The server returns processed video streams to each client.
4. The client forms a conference layout.

SVC technology allows transferring several substream of different quality in one stream. As a rule these streams are basic and secondary ones. The basic stream is transferred in standard quality, while the secondary one – in the enhanced quality, for example, with higher frame rate or video resolution.

And it’s all available in automatic mode. For example, when changing the video layout, the window that was the main one, and then reduced in size, begins to receive a smaller thread. This is how the server reacts to changes in the window’s resolution. The same applies in both directions: when the window is zoomed in, the video quality increases.

SVC enables a video conferencing server to adjust the video stream to the varying characteristics of the endpoint: CPU capabilities and bandwidth. The server sets the specific stream type for each device: endpoints with high bandwidth decode the whole stream, while endpoints with low bandwidth or CPU (mobile phones and tablets) receive only the basic stream with lower data transfer rate. In other words, “weaker” participants don't influence video quality other participants receive during the conference.

For SVC support there is no need for a powerful, expensive server, just an ordinary computer with a modern processor. The technology is so optimized in stream encoding that the number of participants in a video conference can reach several thousands people.

In terms of network load, a set of threads differs from a single thread (as in MCU) by 20-30%. There is no redundancy, as SVC multilayer allows selecting a thread very precisely for the resolution chosen by the user.

Without SVC all multipoint conference participants receive a video stream of the quality satisfactory for the device with the weakest characteristics. From now all multipoint conference participants will receive the image quality maximum possible for the equipment and network channels they use.

In TrueConf solutions SVC technology was adjusted to meet requirements of VP8 video codec. We want you to enjoy our video conferencing solutions and take full advantage of visual communication.

That’s why we’ve developed scalable video coding technology, which allows a server to easily change data transfer rate by modifying resolution or frame rate. In other words, each participant of a multipoint video conference receives the best video his/her device and network channels can provide. SVC introduction became an effective alternative to the technologies implemented in MCU (Multipoint Control Unit), which use total video conversion for each device. This process requires high CPU facilities and causes high pricing of the MCU server.

SVC-based systems have all advantages of mixing architectures and at the same time are deprived of multiplexing systems’ faults. They are cheap, cross-platform and easily scalable.

You can also learn more about other video conferencing architecture types on our website.