Scalable Video Coding (SVC Video)

• Maintain high quality to and from remote conference participants over the public Internet, despite problematic connectivity, packet loss and limited bandwidth
• Deliver high quality desktop to desktop conferencing even if network conditions or client capabilities are limited
• Achieve superior connectivity between MCUs in distributed networks, even if network paths are prone to errors
• Deliver superior high quality video in a mixed SVC/AVC network environment 
• Reduce costs with a simple, streamlined MCU approach

SCOPIA H.264-SVC Frequently Asked Questions

What is SVC?

Scalable video coding (SVC) is a technique that enables a video stream to be broken into multiple layers of resolution, quality and frame rate. 

What are the main benefits of using scalable video for video conferencing?

The main benefit of this technology for video conferencing is to improve error resiliency and therefore video quality across networks that are prone to packet loss. This is particularly important where available bandwidth cannot be guaranteed or across unmanaged networks like the public Internet. 

Is SVC a standard?

The extension to the H.264 video protocol that defines how devices encode and decode multi-layered streams is a ratified standard and is called H.264-SVC. Devices using H.264-SVC can benefit from improved error resiliency when communicating with each other. However many other components which are required to allow full interoperability between SVC enabled devices have not yet been fully standardized. These include such things as support for SVC in SIP signaling and RTP transport and how error protection mechanisms are handled. These components are currently in the process of design and ratification but until then these components might be implemented differently by each vendor. 

How is SVC better than existing error correction schemes such as IPLR or Polycom’s LPR?

The quality of video streams using standard video codecs such as H.264 without any error correction schemes typically degrade dramatically when there is as little as 1% packet loss. Current error protection schemes such as IPLR or Polycom’s LPR can help protect the quality of the video steam up to approximately 5% packet loss. However these schemes use error resilience techniques that can be categorized into two main groups: protective coding and correction codes. In protective coding the encoder (using standard H.264) does not exploit all the redundancy in the stream for compression resulting in much more resilience to packet loss but much less efficiency in compression which degrades quality. Using correction codes (e.g. Forward Error Correction) the encoder can protect the stream without degrading the quality but it dramatically increases required bandwidth and/or latency.

Exploiting the layered structure of SVC, RADVISION is able to protect the stream without adding substantial amounts of bandwidth (like in the protective coding approach), while keeping very high quality (like in the correction code approach).

This dramtically improves the quality of the video over unmanaged networks. H.264-SVC has been shown to provide high video quality even over networks with as much as 50% packet loss while requiring only a nominal increase in bandwidth (less than 10%). 

What are the main drawbacks of using Scalable Video Coding-based solutions today?

The main drawback of using scaled video today is that they are not interoperable with standards based video conferencing devices today. In order to deploy SVC devices into today’s heterogeneous networks requires either a SVC to AVC gateway or a SVC enabled MCU to connect the two worlds. There are significant differences in performance between these two solutions. Gateways can introduce delays and degradation in quality to standard based video conferencing devices and since today this is the vast majority of devices deployed this solution does not favor today’s environment. SVC enabled MCUs maintain high quality support for today’s standards based room systems and telepresence solutions while allowing interoperability with SVC devices. 

What are RADVISION’s plans regarding Scalable Video Coding?

RADVISION is introducing scalable video coding support to its infrastructure and desktop products. RADVISION believes introducing SVC into the current environment as an evolution rather than revolution. H.264-SVC is an important improvement to video coding just like improvements were made from H.263 to H.264. In the next couple of months, expect to see RADVISION release desktop to desktop conferencing using SVC for mixed desktop and room solutions.

How is the RADVISION market approach unique?

One approach for deploying SVC is to create a SVC “cloud” where all devices in the cloud are SVC enabled. To interoperate with standard devices (e.g. H.264-AVC), gateways are used to allow AVC devices into the SVC cloud. This approach requires a gateway connection for every each device which limits scalability and results in double transcoding for the standards devices – once coming into the SVC cloud and once going back out. This results in a lower quality, higher latency video experience.

The RADVISION approach does not degrade the video quality of the existing video devices (such as HD room systems and telepresence) and will provide the best video quality possible to each device. 

What level of error resiliency can I expect from RADVISION’s SVC implementation?

The following graph shows the PSNR (video quality) vs. packet loss rate when using standard H.264 and RADVISION’s H.264-SVC implementation. It shows that when using standard H.264, there is rapid degradation in the video quality with as little as 1% to 2% packet loss. However, when using RADVISION SVC the video quality remains high even at packet loss as high as 20%.



The following screenshot was taken from a real test using SCOPIA infrastructure. There is a clear advantage to the RADVISION SVC implementation over standard H.264 in networks with packet loss:

Will RADVISION’s solution interoperate with standard-based rooms systems and telepresence devices?

Yes. The key to RADVISION’s approach is support of H.264-SVC in the SCOPIA infrastructure allowing interoperability with RADVISION’s SVC enabled SCOPIA Desktop with other standard video codecs like H.263 and H.264. Different devices can be connected to the same conference where each device will get the best video codec according to its capabilities. Devices that do not support H.264-SVC will be connected using H.263 or H.264 while SVC enabled devices will be able to utilize the benefits of SVC.
 
Is RADVISION’s SVC implementation a standard? Will it work with other SVC enabled solutions?

RADVISION will implement standards based H.264-SVC codecs to the SCOPIA platform and is committed to full interoperability testing with other SVC devices once they are available.