MPEG Standardization Development

Digital video and audio coding technology has a long research history and international standards have also been established. It can be said that all of the effort involved with this finally culminated in the MPEG scheme. MPEG (Moving Picture Experts Group) is a byname of a joint technical committee subsidiary group of the International Organization for Standardization (ISO) and the International Electrotechnical Commission (IEC). Since the end of the 1980's, the MPEG-1, MPEG-2, and MPEG-4 standards were developed through the dozens of meetings and numerous between meeting discussions accomplished by e-mail throughout most of the decade. Standardization of the MPEG-7 system will follow the above-mentioned systems, and it will regulate contents description data formats that support video and audio content retrieval.
Up to now, our laboratories have contributed to the standardization activity in MPEG. We will contribute continuously in the future.

MPEG-1

MPEG-1 (January, 1988-November, 1992) Application CD-ROM, Video Karaoke Input SIF video (horizontal 352 vertical 240 pixels) Characteristics Editing/random access feasible

MPEG-1 standard was designed to store video data on CD-ROM, and later became a foundation for MPEG-2 and newer standards.

MPEG-1 is the byname of the first video coding standard completed by MPEG. It is a coding format for storage media such as CD-ROMs, and it compresses video data with half the resolution of SDTV both vertically and horizontally (SIF) down to approximately 1.5Mbps.
Using a technique called "motion compensation prediction," MPEG-1 first extracts frames with different image patterns from consecutive frames. There are two methods of motion compensation: one estimates from the previous frame and the other estimates from the previous and preceding frames (bi-directional prediction). Additionally, it classifies this selected information utilizing a DCT (Discrete Cosine Transform) technique based on the relevance to human visual perception. By omitting extraneous information using human visual peculiarity, drastic data compression is accomplished by means of MPEG-1.
Although MPEG-1 picture quality is inadequate for regular broadcasting due to the significant data reduction, and utilization of MPEG-1 is limited only to broadcasting data retrieval, MPEG-1 is the very foundation of the MPEG-2 encoding standards and those that followed.

MPEG-2

While MPEG-1 was a coding standard for storage media, MPEG-2 was designed with the aim of expanding the general-purpose coding scheme, making it applicable to various other media fields including broadcasting. Its standardization was advanced in cooperation with other international standardization organizations such as the ITU-R and the ITU-T. The MPEG-2 system manages a wide range of video input images, from SIF to HDTV. The data compression technique is based on the MPEG-1 format, and it also includes a system that efficiently codes interlaced images (which comprise a normal television signal). Furthermore, it introduced a hierarchical (scalable) coding function enabling a low quality partial reproduction of coded information to be completely reproduced as high-quality images.
Since MPEG-2 covers a wide range of applications, some of its functions are unnecessary depending on how it is used. Therefore, multiple subset standards are defined within the system and utilized by the application of two indexes: one is a profile index that indicates the degree of functionality, and the other is a level index that establishes object image resolution. For example, the above-mentioned bi-directional prediction cannot be accomplished in a simple-profile, but only in a main-profile. A main-level is able to handle video inputs up to the point of SDTV, and a high-level up to HDTV images. The MPEG-2 standard has been adopted for use in digital broadcasting worldwide.

Figure: Object-based coding system

MPEG-4

MPEG-4 (November, 1994-) Applications Portable videophone, internet video, studio production Input QCIF (horizontal 176vertical 144 pixels) to HDTV video Characteristics CG animation coding, coding of individual component elements May 1999 Version 1 (fundamental part) standardized. March 2000 Version 2 (function addition) to be standardized. Version 3 (studio production standard) is under examination.

While MPEG-2 was hardware-oriented, standardization of MPEG-4 was initiated endeavoring for a general-purpose coding method for multimedia data software processing.

In the beginning, the application of MPEG-4 was supposed to be limited to portable videophones or the Internet, applications consisting of a lower bit rate than that of MPEG-1. However, high-bit-rate applications are presently being developed including broadcast program production. New concept functions are involved with this such as coding schemes that impede transmission errors. CG animation image coding was also introduced in addition to MPEG-2 technical improvements such as advanced motion compensation. In addition to taking note of the coding efficiency, object-based coding also had been adopted. As shown in the figure, this technique first divides video images into composing objects such as a person and a background scene. The system then performs an individual coded transmission of each object, and the receiver synthesizes the data to reproduce the desired image. This enables more efficient compression and various video controls utilizing video component information that is very useful for program production.
In terms of low-bit-rate applications, there are home video cameras, Internet video distribution, and next-generation portable videophones, some of which are already in practical use. I inserted a table of product examples. There is also a vision of using MPEG-4 for mobile media digital satellite broadcasting.

Table: MPEG-4 product examples
Company (product name, etc.)	Description	Bit rate
Toshiba (MobileMotion)	In-house education/training, Internet advertizement	Max. 1.5Mbps
Hitachi (MEDIAHALL)	Non-linear editing for in-house education, Intranet distribution	Max. 2Mbps
Sharp (Internet Viewcam)	Recording to Smart Media	Max. 384Kbps
Matsushita (under development)	Internet video distribution
Microsoft (Windows Media Player)	Internet video	28.8 300Kbps
Toshiba/Fujitsu/Toyota/NTV, etc. (RadiVision)	Mobile satellite broadcasting service	Max. 256Kbps

MPEG-7

MPEG-7, its official name being Multimedia Content Description Interface, is the fourth MPEG standard. MPEG-1, MPEG-2, and MPEG-4 are compression coding standards for actual video and audio data. To the contrary, the MPEG-7 standard covers the content description data that is attached to multimedia data in various other formats including those of MPEG-1, MPEG-2, and MPEG-4. Content description data, sometimes referred to as metadata or an index, is a description of what is included within the data. To illustrate, if we imagine video, audio, or textual data that consists of multimedia contents as books in a library, then the content description data would correspond to a library card. In other words, by referring to this description data (a library card), retrieval of multi media contents (books) becomes possible. The MPEG-7 standardizes the format and description of data in the same manner as a library card system.

Standardization items
One of the standardization items I will explain here is the "descriptor."
The descriptor is a function that defines and forms the contents description data. In the context of the previously mentioned library card illustration: this data will indicate properties such as the "name of a book," "subtitle," "author," and "abstract." The Descriptor standardizes the meaning of each item and the data format that it should be described with (i.e. the "name of a book" is defined as a title that is written on the front page of the book, expressed in Japanese characters.) Although this library card illustration highlights the descriptor's "liberal-art" attributes, the descriptor for video and audio data involves more "technical" functions. For example, the size and motion of image subjects in a video are expressed numerically.
To have the above-mentioned descriptor adopted as a standard, it is essential to verify its efficient performance using actual video contents. The MPEG is pursuing the establishment of the standard while simultaneously conducting verification tests.

Prospects of contents distribution promotion
The MPEG-7 standard is scheduled to be approved as an ISO international standard in September of 2001. Helping to make this standard useful to broadcasting systems as well, NHK has given necessary suggestions and proposals based on its research results.
When MPEG-7 is diffused and its use becomes more common, contents with better description data attached will probably have a larger audience. This seems to bring about new possibilities for the contents industry.
Beginning with the standardization of MPEG-7, the marketing targets that will involve MPEG seem to be global, targets such as providing general technologies for multimedia contents circulation.

The MPEG is about to start working on a new standard, MPEG-21, in expectation of the formation of a video/audio contents distribution market. Although the standards parameters are not yet concrete, discussion is underway regarding copyright protection technology for contents circulation.