DTV's Future: The Advanced MPEG-4 AVC Codec

As broadcasters and multichannel operators seek to deploy multiple high definition television channels, expand their digital channel offerings, and develop additional two-way interactive services, one factor stands in the way of their realizing profits from these services: increasing bandwidth costs. There is, however, a solution to this problem. New compression systems promise to dramatically lower the bandwidth required to deliver DTV channels and new business models at a significantly lower cost and with less disruption to legacy services.
Current SD and HD digital video is based almost entirely on MPEG-2, an almost 10-year-old standard that has nearly reached the limit of its video compression efficiency. Soon it will be supplanted by MPEG-4 Part 10/AVC (Advanced Video Coding), a new and open video compression standard that was ratified in 2003 by both the ITU and the MPEG committee.
From the beginning, MPEG-4 Part 10/AVC (the ISO standards designation) and ITU-T H.264 (ITU-T standards name) was developed specifically by and for television broadcasting, whether it’s via terrestrial, cable, satellite, or Internet delivery. It leverages the existing MPEG-2 transmission equipment to deliver the new types of bandwidth-hungry media to a single user or thousands of users simultaneously.
By using the same transport stream protocol and modulation techniques as MPEG-2, AVC is immediately deployable by broadcasters and distribution plants alongside their current programming. Simply exchanging your MPEG-2 encoder for an AVC encoder enables programs to be sent in half the bandwidth.
That’s because AVC codecs have been shown to outperform MPEG-2 and the current MPEG-4 ASP (Advanced Simple Profile) by at least a factor of two, thus requiring 50% or less bandwidth or storage capacity to deliver the same video quality. This means that instead of having to transmit HDTV at 19Mbps and SD at 4Mbps, engineers can now get equivalent HD picture quality at around 8Mbps, SD at 2Mbps, and DVD-quality video at less than 1Mbps. Even greater efficiency can be expected in the future.

A New Compression Standard
AVC is already a mature standard. It includes all of the capabilities of MPEG-2:
-- Progressive and interlaced scanning
--4:2:0, 4:2:2 and 4:4:4 chroma sampling
-- 8-, 10-, and 12-bit-per-pixel resolution
-- Well defined licensing
In addition to having been proven superior to MPEG-2 and MPEG-4 ASP, it has shown better performance in at least some applications to Microsoft’s Windows Media 9. In contrast, the technology within Microsoft’s Windows Media 9 has just begun the SMPTE standardization process, where it is known as VC-9. Furthermore, as part of standardization, Microsoft has only recently begun the licensing process by submitting it to MPEG-LA, the same entity responsible for AVC licensing.

Moving Forward
At Modulus, we’re initiating trials of our first AVC compression product, the AVE-SD, a realtime SD compressor that accepts SDI video and audio on its input and delivers fully compliant AVC Main Profile Level 3 video in MPEG-2 transport stream packets over UDP or DVB-ASI outputs. The computational power of our FPGA accelerator (over 100 billion operations per second—BOPS) is far beyond that of systems based upon one or several DSPs, each with at most five BOPS. That power allows our AVE-SD to more fully exploit the many encoding tools included in the AVC standard.
At NAB2004, we showed HD video compressed in realtime to 6.6Mbps. Constant bit-rate SD video being displayed was captivating at 1.5 Mbps, and we have only just begun to scratch the surface of what’s possible with the technology. In our labs, we are proving that MPEG-4.10 AVC has even more performance potential as we extend the computational power of our system.
In the real world, Path 1 Network Technologies, which provides technology to send broadcast-quality MPEG video over IP networks, has successfully demonstrated a MPEG-4 AVC-based transport system with Modulus Video, using Path 1’s Chameleon vidXwan technology platform. This allows new program distributors to send digital video (SD and HD) and other data over traditional phone lines as well as over fiber-optic cable, or a secure Internet connection, in a highly cost-effective way.
In the near future, a next-generation MPEG-4 AVC codec will also enable distributed broadcast content over IP networks utilizing a low-bandwidth live backhaul service and mobile reception with cell phones, PDAs, and specialized pagers. These new types of business models were not possible before the advent of MPEG-4 AVC and its complex software algorithms/hardware solutions. In addition, delivery of video over DSL IP networks will enable telcoms and cable over-builders to get into the video delivery business at a realistic financial investment.
It’s clear that a number of businesses can and will benefit. Whether, however, the MPEG-4 AVC will primarily be used for extending the coverage area of TV stations, delivering low bit-rate, realtime video to next-generation handheld PDAs, or simply increasing the capacity to TV receivers within the existing coverage area, remains to be decided. Basically, the market will come to its own conclusions, but it’s clear that MPEG-4 AVC can facilitate it all.