MPEG-4 about two years away.......................................
asiansources.com
MPEG-4Posted December 5, 1997
Latest MPEG standard could open new markets
It will be a few years before most MPEG-4 products emerge, but the low-bandwidth multimedia technology should complement MPEG-2 capabilities.
Just as MPEG-2 technology begins to move into the mass consumer market as part of DVD, a successor standard with its attendant technology is being hammered out. MPEG-4 is the next effort of the Moving Pictures Experts Group (MPEG) and, while based on the older approach, it is quite different from MPEG-2 in several key aspects. As a result, several new applications and markets could emerge.
Several challenges must also be overcome before MPEG-4 can enter high-volume production. Nevertheless, companies such as Microsoft, Sun, Intel, Motorola, Toshiba, Philips and SGS-Thomson are engaged in developing the standard, as well as the products for that standard. Experts forecast widespread acceptance and implementation of the new technology.
"Eventually, it will be hard to buy a personal computer that's not capable of doing this," predicts Ralph Rogers, principal analyst for multimedia at San Jose, California-based Dataquest Inc.
Eventually in this case means at least several years. The standard will be finalized in two steps, with version 1 slated for final release in late 1998 and version 2 a year after that. Concerns about the difficulty of implementing all the features of MPEG-4 lie behind this two-step approach. The complete, frozen standard will not be ready until nearly the turn of the century. Products and applications will have to wait on that event, which means the MPEG-4 era has to take place some time after the year 2000. Version 1 products will show up before then, but even those will not appear until late 1998 at best.
Multimedia protocol
MPEG-2, both the standard and the technology that supports it, focuses primarily on how to handle the transmission of digital video. It is essentially a digital representation of an existing analog broadcast standard, and is based on transmitting complete scenes using 6MHz bandwidth.
MPEG-4 is different in several ways. First, it is not designed to be either just a video or an audio specification. It is an entire multimedia protocol, with standards for how to stream video, how to synchronize multimedia, and how to manage different data types.
"MPEG 4 is object-orientated," explains Paul Fellows, external projects manager for SGS-Thomson in Bristol, England. "It's basically concerned with objects in the same line as pixels."
SGS-Thomson is one of the world's foremost manufacturers of MPEG-2 semiconductors, and the company is active in defining MPEG-4 with an eye toward commercial implementation of the standard. SGS-Thomson is teaming up with other European partners in the European Union's Emphasis project to make reference hardware MPEG-4 designs using the SGS-Thomson Chameleon and Philips Trimedia processors. Philips is already demonstrating an MPEG-4 decoding algorithm running on its Trimedia chip.
The Trimedia chip decodes multimedia objects that the receiving terminal uses to reconstruct scenes. These objects can be video, audio, text, animation, or any other type of multimedia. This object orientation has several benefits. For one, this means that the best playback method can be used for a given data type without degrading any other. Text can be visually sharp while at the same time streaming video is being played back. Another advantage is that it is much easier to select or manipulate parts of the scene. That benefits multimedia production, but it also allows viewers to interact with broadcasting in ways not possible before.
"The audience participates or becomes part of the experience of the show," says Asaf Mohr, president and CEO of Cambridge, Massachusetts-based VDOnet, a leading supplier of real-time video software.
Computing demands
However, there are some drawbacks. One is that it is hard to predict the amount of computing power needed by the new standard. With MPEG-2, that figure was relatively constant, and hence easy to judge and plan for. With MPEG-4, the load could be much lighter if there were relatively few objects that were changing frame to frame, but it could also be much greater if there were many.
Just how much computing power will be needed by MPEG-4 also depends upon the how "shapey" the objects being encoded and decoded are. Fellows, of SGS-Thomson, reports that with proper care the burden can be lessened from 50 percent of total processor time to 17 percent if content developers choose objects carefully.
Another key difference with the older standard is that MPEG-4 does not die when bit rates drop below 1Mbps. The standard will be able to handle everything from megabits per second on down to a few kilobits per second. Because of this low bit rate capability, MPEG-4 can be used in such applications as video telephony, Internet broadcasting, or in wireless communications. It is such applications that are of interest in Europe, Japan and elsewhere among hardware vendors.
"If there are ASIC or any kind of hardware components that come out, they'll be targeted to low-end things, like say putting video in a pager, putting video in a cell phone," comments Jed Johnson, director of the streaming media business unit of Motorola, in Mansfield, Massachusetts.
Motorola's streaming media business unit makes software for providing real-time video over the Internet and company intranets. Microsoft also is active in this area, and it is Microsoft's backing of MPEG-4, along with a strong interest in the standard in Europe, that is likely to make the new approach popular.
One downside of the lower bit rate capability is that to maintain acceptable playback rates at such low speeds requires higher and more computationally intensive compression and decompression of the objects. Consequently, in situations where network congestion may impact bandwidth, such as the Internet, the load on processors could vary quite a bit.
Although Motorola does make MPEG-2 chips and is in both the pager and cellular phone business, the company has no announced plans for a hardware implementation of MPEG-4. Other hardware vendors such as Philips and SGS-Thomson also will not release details about specific products, but they do acknowledge that they expect that eventually MPEG-4 decoders will be a combination of software and hardware.
Not the end of MPEG-2
This is not to say that MPEG-2 will be swept aside in favor of MPEG-4. For one thing, MPEG-2 has a large head start and will soon have a large installed base. Dataquest predicts that there will be some 40 million DVD players, either in PCs or as standalone units, sold in the year 2000. That is significantly up from the less than a million in 1997.
Also, the two standards are complimentary. MPEG-2 excels at handling high-bandwidth broadcasting, while MPEG-4 is intended more for low-bandwidth interactive applications. For PCs and software-configurable standalone playback stations, the ability to handle MPEG-2 computing load also enables them to handle MPEG-4. In such cases, converting from one to the other may only require a programming change, so much of the installed MPEG-2 base might eventually be able to handle either standard. Having access to so many viewers is also likely to be a big boost to the adoption of MPEG-4.
It is even possible that a playback device would be called upon to handle both at the same time to take advantage of what each has to offer.
"Terminals could receive MPEG-2 current broadcast channels and also decode MPEG-4 streams from the Internet or side channels and enable videoconferencing or 3D games or teleshopping-applications where MPEG-4 has clearly some advantages," notes Isabelle Corset, head of the image and communication department at Philips Electronics Labs (LEP), in Paris, France.
While some have characterized MPEG-4 as a technology solution in search of a problem, there seems to be enough interest in it to ensure the new standard's success. And if not, there is always the next standard waiting in the wings-MPEG-7.
Electronic Components - January 1998 |
