MPEG-4 – Advanced Video Coding

Few in the industry would disagree that H.264 will likely emerge as the winning standard for video compression for the next decade. Among the alternatives that have been proposed, the H.264 standard was developed inside the entertainment industry, and it provides the best combination of bit rate (half that of MPEG-2) and high video picture quality. Like MPEG2 before it, H.264 compression can be used for storage and distribution of all types of video, ranging from low resolution CIF formats for hand held mobile applications, to high definition (HD) formats for broadcast TV where the consumer can enjoy a “larger than life”¬†viewing experience.

H.264/MPEG-4 Part 10 or AVC (Advanced Video Coding) is a standard for video compression, and is currently one of the most commonly used formats for the recording, compression, and distribution of high definition video.

Among these applications, broadcasting of HD video has proven to be the strongest catalyst for adoption of H.264 where an extremely large volume of data has to be squeezed into existing channels for video distribution. Consider the problem-a 1920 by 1080 line of uncompressed 4:4:4 HD video has a real time data rate of close to 3Gbit/s, and a 2 hour HD video requires 2.69 Tera bytes of storage! MPEG2 can compress HD down to 19Mbit/s, but H.264 can compress the video even further to 8 – 10Mbit/s, and given the constrained channel bandwidths for satellite and IP networks, H.264 is becoming an essential part of worldwide HD broadcasting for the future. However, there are multiple technical and business issues that have to be addressed by the industry on the road to making H.264 (also referred to as the Advanced Video Codec – AVC) as ubiquitous as MPEG-2.

A number of technical challenges are involved. First of these is that H.264 achieves its low bit rate and high quality through increased computation complexity relative to MPEG2. Table above illustrates the algorithmic differences between AVC and MPEG-2, and as can be seen the calculations for motion estimation and entropy encoding are far more sophisticated in AVC. Where MPEG2 specifies only 16 x 16 motion estimation, H.264 provides a choice of seven adaptive block sizes. Intraprediction and interprediction modes in MPEG2 are fixed and limited, while in H.264 there are numerous choices for both. MPEG2 specifies fixed VLC entropy encoding, whereas in H.264 two types of entropy encoding are specified: Contextbased Adaptive Binary Arithmetic Coding (CABAC) and Context-based Adaptive Variable-Length Coding (CAVLC).

The industry is starting to resolve this issue through a new class of H.264 semiconductor chips, which provide a higher level of integration than typical general-purpose DSPs and FPGAs. These chips vary from fully programmable processors to hardwired and partially programmable ASICs/SOCs that are suitable for consumer H.264 applications. Fully programmable processors are the best match for professional broadcast applications for encoding. This is so for a couple of reasons: First, an H.264 encoder can achieve a low bit rate in a number of creative ways using the flexibility and adaptability of the H.264 algorithm. Hence an encoder based on a programmable processor can make incremental algorithmic refinements in the implementation, and over time improve video quality while continuing to lower the bit rate. Second, H.264, like MPEG2 before it, is an evolving standard.

To support the most demanding professional applications and deliver the highest resolutions and quality in content editing and post processing, the industry has already defined extensions of the H.264 standard such as those encompassed by the Fidelity Range Extensions (FRext). Fully programmable silicon solutions allow encoder manufacturers to conserve their software and hardware investments as they migrate their equipment to support FRext. With the help of new broadcast quality HD H.264 processor chips, including Telairity’s T1P2000 processor for broadcast-quality HD encoding, AVC system vendors can deliver reasonably priced equipment and motivate broadcasters and content aggregators to engage in the capital spending required to make their facilities fully AVC compatible.

The professional broadcasting side of the H.264 equation is just one part of a larger challenge. Even more demanding are the technical challenges surrounding the decoder chips that will be used in consumer products such as set top boxes, DVRs, and HD-DVD and BluRay DVD players. Unlike encoder solutions, decoder solutions do not have many degrees of creative/algorithmic freedom and must support the H.264 standard exactly. They also need to support the superset of audio and video compression standards that will be used to broadcast content over the next two to three years. The VC1 codec, for example, is supported by several studios and broadcasters, and thus decoders for consumer products must support VC1 along with MPEG2 and H.264. For the HD/AVC market to truly take off, decoder SoCs will need to be multiple-sourced, available in high volumes, and relatively inexpensive so as to enable HDTV set top boxes and other consumer entertainment products.

Even so, the industry is well on its way to overcoming the technical and business obstacles to making H.264 ubiquitous. The recent announcement of powerful HD camcorders from Sony and Panasonic that now support HD-AVC compression indicates the gathering momentum in AVC content creation and post production workflows for AVC video. Sports broadcasters such as NFL and ESPN have been a primary driver for HD and H.264 compression in 2006, and have been a big draw for consumer spending and advertising dollars. Several new high definition AVC codec chips, decoders and processors, were announced over the last year to address both professional and consumer applications, and AVC capable set top boxes are ramping in volume. Satellite service providers such as DirecTV have been able to substantially increase their HD programming by switching over to AVC transmission, and IP networks and DVB-H broadcasters have also adopted H.264. Assuming these trends continue, we can look forward to watching more and more content in high definition worldwide.