Amsterdam, Netherlands - A number of available H.264 codec solutions were on display at the International Broadcasting Convention that began here last week, indicating the gains the technology has made since last year's show. Little more than a curiosity 12 months ago, H.264 is now positioned as a possible successor to MPEG-4, though unresolved intellectual-property licensing issues could derail its momentum.
Since its formal ratification by the International Telecommunication Union in May, H.264 has attracted the interest of a range of broadcasters, service providers and consumer electronics companies, according to sources close to H.264 licensing negotiations. Carriers and satellite operators including DirecTV Inc. are determined to use the real-time encoding scheme for their next-generation video services, possibly starting in 2004, the sources said.
Unlike MPEG-4, H.264 includes a built-in Internet Protocol adaptation layer, so the H.264 protocol can be mapped onto any fixed IP, wireless IP, storage or broadcast network. That's why both telecom and consumer electronics companies appear ready to support H.264, which some call MPEG-4 AVC and others JVT (for Joint Video Team).
Serious interest from one mainstream satellite operator has set off a bidding war among suppliers of H.264-compliant silicon, several suppliers said.
Because it may be a common technology on a range of systems, some see H.264 as an enabler of next-generation convergence, allowing consumers to exchange and transfer multimedia files among mobile phones, camcorders, recordable DVD disks, home servers and PCs. But unresolved issues related to the licensing of H.264 intellectual property could pose a problem. "Ideally, we should have had [H.264 intellectual-property licensing] by today," said Rob Koenen, president of the MPEG Industry Forum. "My hope is that the H.264 licensing scheme would be simpler than MPEG-4's. There are legitimate concerns in regard to the use fee."
Ken McCann, chairman of Digital Video Broadcast's Audio Visual Coding module and a director of ZetaCast Ltd., said that "finding a viable business model" for licensing was the biggest roadblock facing H.264.
As licensing issues drag on, Microsoft Corp. is maneuvering to take over the codec market with its proprietary Windows Media technology, observed Rick Doherty, president of the Envisioneering Group.
Meanwhile, a large U.S. satellite operator has reportedly threatened to choose another encoding/decoding technology and take its business elsewhere if the H.264 licensors don't get their act together.
Despite those uncertainties, which could be resolved by the end of this year according to the most optimistic forecasters, a bevy of companies rolled out H.264 technologies at IBC, including Harmonic, Heinrich-Hertz-Institut, Ingenient Technologies, LSI Logic, Modulus Video, SandVideo, SeaChange, Tandberg TV, Texas Instruments and UBVideo.
Also at IBC was Vanguard Software Solutions Inc., a little-known Los Altos, Calif., startup with a majority of its R&D engineering team based in St. Petersburg, Russia. Vsofts, as the company is known, made waves here with its H.264 Baseline Profile codec algorithm, capable of encoding standard-definition content in software on a PC with a 3-GHz Pentium 4 processor. In a demo that mixed H.264 with MPEG-4 AAC-plus-SBR audio coding, Vsofts allowed visitors to pick a video file, pair it with audio and encode it in real-time.
No company had previously demonstrated a software-only real-time encoding capability that allowed standard-definition pictures to be shown on a PC platform without an acceleration board. Using its patent-pending H.264 codec, Vsofts' encoder "can run 40 times faster than the Joint Video Team's reference code, without degrading the picture quality," said Thierry Fautier, vice president of marketing at Vsofts. While the JVT reference code may be able to encode only one frame per second, the Vsofts encoder can encode the same reference sequence at 30 frames/s, the company said.
The software-only H.264 codec "is very important," said the MPEG Industry Forum's Koenen, "because any new codec that will be broadly used needs to get first established in the Internet world, capable of running on a PC platform."
Though Vsofts' current software codec is based on H.264's Baseline Profile, the low-latency, fast-response solution is suitable for such broadcast applications as news cameras or studio video manipulation equipment, Fautier said. Vsofts, he said, is in "talks with several chip companies," adding that its codec can be customized to various silicon platforms.
LSI Logic Corp. showed an H.264 Main Profile-compliant real-time standard-definition video technology platform at the booths of partners SeaChange International, a provider of video-on-demand systems, and Modulus Video Corp., a video-over-DSL company. LSI Logic's encoder platform is based on three FPGAs and two Pentium 4 processors running at 2.5 GHz.
Bob Saffari, senior director of marketing and application engineering of advanced video products at LSI Logic, stressed the difference between H.264's two profiles-Baseline and Main. The Baseline Profile is "a far less complicated encoding scheme, more suitable for videoconference types of applications," he said. Designed for broadcasting, the Main Profile supports interlace, and adds B frames and more-complex tools such as context-based adaptive binary-arithmetic coding. With its VLE4000 real-time encoder, VLD4000 real-time decoder and VSA2000 stream analysis tools, LSI Logic is prepared to provide customers with its H.264 Main Profile Level 3-compliant tools necessary for "proof-of-concept trials, early deployment and infrastructure build-up" that must begin later this year or early in 2004, Saffari said.
Texas Instruments Inc. demonstrated its Digital Media Development Kit at IBC, and showed H.264 Baseline and Main Profile encoding and decoding capabilities on its DSP-based platform. TI has been working with H.264 software algorithm developers such as UB Video Inc. and Ingenient Technologies Inc. TI's development kit features a 600-MHz DM642 DSP-based digital media processor and comes with a PCI-based emulator, video camera, application software and utilities, and evaluation board with 32 Mbytes of external SDRAM, 4 Mbytes of flash, composite and S-Video I/O, S-Video I/O, a VGA output port and an Ethernet port to enable streaming media. The comprehensive kit is intended "to allow video engineers to become more familiar with DSP, while making it easy for DSP engineers to develop a video system," said marketing manager Thomas Brooks.
See related chart