Scalable Video Coding (SVC)

Overview

ITU-T H.264, or AVC video coding, is fast replacing the earlier video standards in the end-to-end chain of video content delivery, personal entertainment, video surveillance, and visual communications. Scalable Video Coding (SVC) is an amendment to H.264 that is aimed at creating an efficient scalable representation of video that can be consumed according to the capabilities of diverse networks and end-terminals.SVC offers temporal (frame-rate), spatial (resolution) and quality (SNR) scalabilities that are built on an AVC-compliant base layer. This ensures backward compatibility and re-use of technologies developed for AVC.

Applications for SVC include

• Mobile Broadcast - SVC has been provisioned for new mobile broadcasting standards such as ATSC-M / H. This standard employs the spatial scalability aspect.
• Video surveillance - applications that require multiple resolutions and frame-rates across storage, LAN, and WAN accesses benefit from the spatial and temporal scalability aspects of SVC.
• Multi-point video conferencing - can leverage SVC across wired / wireless networks with diverse end-terminal capabilities. In addition to resolution and frame-rate scalabilities, quality scalability is also needed to adapt the stream to the available bandwidth that will be time varying. SVC offers the possibility of eliminating additional end-to-end latency currently incurred due to complex transcoding on an MCU.
• Video streaming and video-on-demand - applications over internet can eliminate the need to encode at multiple resolution, frame-rates, and bit-rates.
• Home servers can have a single representation from which different versions can be derived without requiring transcoding 

Closed systems and unequal error protection based approaches can benefit from SVC readily. The benefits of SVC will increase as different network elements, such as routers, become SVC aware, sending only the necessary portion of the bitstream over to the next router based on available bandwidth assessed or based on the capabilities of the end-terminals.

Ittiam SVC

Ittiam has developed a rich portfolio of intellectual property (IP) around the H.264 standard over the past many years. This IP includes highly differentiated algorithms in the areas of motion estimation, mode selection, rate control, transcoding, low latency solutions, and error resilience / error propagation control. These IPs are embodied in Ittiam's Reference C model which forms the basis of implementations of H.264 on various platforms.

These IPs for H.264 and other video coding standards, along with additional platform-specific IPs such as multi-threaded designs, multi-core architectures, and optimizations have resulted in efficient codecs realized on multiple programmable platforms (e.g., TI processors such as OMAP3, DM6467) and cores (such as ARM Cortex A8 / A9). These codecs have been licensed to several customers in the areas of entertainment devices (personal, in-flight, in-car), video confere ncing equipments, surveillance netcams / DVRs, and transcoding devices.

Leveraging this extensive video compression and platform implementation expertise, Ittiam has developed a reference C-model for the SVC encoder and decoder. These C models are suitable for directing to implementations on a diverse set of platforms. The SVC encoder and decoder C models can also be used as a starting point for further IP development and differentiation by licensees.