In recent years, international standards for digital video compression have broadly divided into two major streams: ITU-T and ISO/IEC.
The standards set by ITU-T focus mainly on telecommunications applications, such as H.261 for ISDN Video Phones and H.263 for POTS/IP Video Phones.
ISO/IEC primarily addresses storage and broadcast applications, such as MPEG-1 for VCD and MPEG-2 for DVD/DTV.
After the year 2000, MPEG-4 and the jointly developed H.264/AVC standards became centered on internet video applications.
This article will discuss the chip development and applications of MPEG-4 and H.264/AVC, as well as the future of security and surveillance products. It is worth mentioning that the Short Video Header feature of MPEG-4 part 2 supports the H.263 baseline, indicating a convergence of technologies from the two major organizations.
The MPEG-4 video compression technology was established by the Moving Picture Experts Group (MPEG) under ISO/IEC. The committee was formed in 1988 and has developed standards such as MPEG-1 and MPEG-2, which have won Emmy Awards and have been successfully applied in industries related to VCD, MP3, DVD, DTV, and HDTV, fostering rapid development in computer and consumer multimedia industries in recent years.
MPEG-4 is formally numbered 14496 by ISO/IEC, with video codec technologies specified in part two (14496-2: Visual Part) and part ten (14496-10: Advanced Video Coding). 14496-2 is subdivided into various Tools, Profiles, and Levels to support a wide range of multimedia applications, including digital interactive television, mobile video phones, multimedia storage, multimedia messaging, distance education, and remote video surveillance.
The 14496-10 part, known as H.264 or AVC, is currently the most advanced video coding standard internationally, jointly developed by ITU-T and ISO/IEC in May 2003. Around late 2001, the ISO/IEC MPEG working group decided to collaborate with ITU-T VCEG to form a Joint Video Team to develop the H.264 standard (formerly known as H.26L), as its performance surpassed that of the MPEG-4 standard. Therefore, H.264 is also referred to as MPEG-4 Part 10, MPEG-4 AVC, or JVT. The standard boasts several unique advantages, including high compression ratios, high-quality video, error resilience, and enhanced network transmission capabilities, far exceeding the performance of other existing standards. The H.264 standard has defined three Profiles: Baseline Profile, Main Profile, and X Profile, suitable for video phone/conference/mobile applications, broadcast video applications, and streaming/mobile video applications, respectively.
Besides inheriting the advantages of increased compression ratios over older standards, MPEG-4 also accommodates bit rates ranging from 5 Kbps to 2 Gbps, capturing the benefits of early standards ITU-T H.261/263 (low bit rate encoding) and ISO/IEC MPEG-1/2 (high bit rate encoding), while supporting object-based coding and synthesized image compression suitable for advanced interactive functionalities and special video production. Its error-resilient coding technology and fine-grained scalable coding technology also make it applicable to the rapidly changing bandwidth scenarios of the internet.
Due to the advantages of MPEG-4, it has been selected as the multimedia coding standard by international standard organizations such as ISMA, 3GPP, M4IF, and WMF.
Regarding H.264/AVC, with the formal approval of the standard and the publication of licensing conditions, it marks the completion of the first step toward technological implementation. While observing the global industry's acceptance of this standard, it has already attracted interest from several major broadcasting companies, service providers, and consumer electronics companies.
In Japan, six television broadcasting companies, including NHK, TBS, NTV, Asahi TV, Fuji TV, and Tokyo TV, have decided to adopt H.264 as the video compression standard for mobile digital television, starting mobile digital television services before March 2006. Japan is expected to become a demonstration area for mobile digital television. DirecTV has also tentatively decided to choose H.264 encoding for its next-generation video service, which may launch in 2004. The DVD Forum conditionally accepted the H.264/AVC standard as one of the video compression technologies for its HD-DVD system in February 2003.
The DVB in Europe is also developing a standard for broadcasting DVB-T television signals to handheld devices, with H.264/AVC as its core video codec technology. The DMB standard promoted by Korea also adopts H.264/AVC as its core video codec technology. In the 3GPP context, H.264/AVC was chosen as the next generation's optional codec at the end of 2003.
For example, H.264/AVC can leverage its high complexity and performance characteristics to provide new application domains and business opportunities. It is already imminent to transmit TV broadcast quality video at 1 Mbps over transmission channels. Therefore, entertainment broadcast services based on ADSL (like Chunghwa Telecom's MOD) will become more competitive.
The implementation of video codec technology can be categorized into three methods: software, firmware, and hardware. The software implementation involves developing MPEG-4 Codec functionality through programming on general-purpose CPUs (mainly RISC), typically seen in PCs, laptops, or PDAs, and sometimes as value-added development on specific embedded systems or IA devices.
Firmware development emphasizes using DSP (Digital Signal Processor) as the core, employing the unique multiply-accumulate instruction sets of DSP, combining high-level and low-level programming languages, and downloading to the DSP platform for execution. This solution boasts short development timelines, easy programmability, and decent transmission speeds.
The most cost-effective solution is undoubtedly the hardware approach, completing video codec tasks through dedicated circuits and connecting all these dedicated circuits using pre-designed components to form a complete workflow. This way, efficiency can be maximized. To expand the scope of applications and feasibility, video codec firmware based on DSP or specific processors provides programmability and quick market entry, which will capture early or niche markets. Lastly, considering power consumption, pricing, and other factors, hardware design will dominate the market.
With the evolution of video codec standards over time, notable breakthroughs have been achieved in the past decade. H.264/AVC has already outperformed MPEG-2 by twice, but the increase in complexity necessitates more advanced designs and manufacturing processes. Regarding applicable ranges, the profiles defined in H.264/AVC offer a glimpse into its capabilities. As various applications gain market acceptance, the superiority of H.264/AVC technology will be realized in multimedia communication and storage applications. Its applications can be categorized as follows:
Regarding the development of MPEG-4-related projects worldwide, several examples can illustrate this. iVAST intends to port the iVAST Experience Player. Philips' next-generation multimedia processor, Nexperia, primarily focuses on next-generation digital STBs and recorders.
Additionally, NTT DoCoMo has launched the FOMA project, mainly applying Mobile Video Phones. Europe, the US, Japan, and Korea have plans or collaborations underway for the application and technical research of MPEG-4. In Korea, Samsung Electronics has decided to collaborate with Serome Technology to develop MPEG-4-based multimedia digital content playback solutions.
MI was the first to commercialize MPEG-4-based Web Cameras in early 2002 and has signed supply contracts with businesses in Japan and mainland China. Infiny Systems has developed MPEG-4-based DVRs for security use, which will be mass-produced soon. Symbion System has developed a 16-Channel DVR using MPEG-4 technology and is preparing for mass production.
Ilryung Telesys is currently collaborating with ETRI (Korea Electronics and Telecommunications Research Institute) to develop a Home Gateway integrating MPEG-4 architecture for home security devices. The third information technology has developed an MPEG-4-based multi-party video conferencing system. MPEG Solutions has introduced an MPEG-4-based PDA, actively competing in the market. MclickPlus, OnTimetek, and Samsung Electronics have video phones and related MPEG-4 technologies, while netNtv, netCODEC, Serome Technology, and MPEG Solutions are developing MPEG-4-based internet broadcasting solutions.
Japan has announced several MPEG-4 hardware solutions, with Matsushita, Fujitsu, and Hitachi presenting MPEG-4 Video ICs at ISSCC 2002. Toshiba Corp. has released MPEG-4 video codec chips, and NEC Corp. has marketed MPEG-4-related chips in collaboration with US company PacketVideo Corp., providing MPEG-4 chips needed for the FOMA project. Moreover, Panasonic, Sharp, JVC, and Sony have also released solutions. Software companies in Europe and the US, like Microsoft, PacketVideo, Solid Streaming, DiamondbackVision, and UBVideo, have already provided codec software. Microsoft’s Windows Media Technologies can offer MPEG-4-related coding and streaming, based on personal computers with Windows OS.
PacketVideo has received Intel's investment and is collaborating to develop a mobile phone platform based on StrongARM. Motorola, NEC, Mitsubishi, Sanyo, and Casio are also adopting PacketVideo software for various mobile devices. UBVideo is a leading company in technology, especially concerning 14496-10 AVC, already providing solutions. MPEG-4 hardware solutions can offer high-quality real-time encoding, applicable to numerous non-PC Information Appliance (IA) products, categorized into two main applications: low-cost/BW (Bandwidth) portable/mobile devices and high-performance products for entertainment use, such as next-generation STBs, HD DVD players, surveillance DVRs, and streaming servers.
Considering the current and future market demand forecasts, the low-cost/BW approach will initially emerge in products using DSP-based solutions due to significant specification variations and market entry design considerations. The current application products mainly include DSC (Digital Still Cameras), followed by mobile phones and PDAs. However, looking ahead at market competitiveness, product integration, and low power consumption, ASIC solutions and even SoC will likely become market mainstreams, with main vendors including TI, Toshiba, Panasonic, Zoran, Luxxon, and UB Video.
蘊藏許多助人的知識與智慧。