Will MPEG-4 Soar or Bore?


After a delay of more than a year, digital media's next full-fledged standard, MPEG-4, is poised to make its entrance. But the digital industry is offering mixed reviews: Supporters said the standard will command center stage for video and interactive content distribution, while detractors countered that the show is pretty much over.

When the ISO/IEC Moving Picture Expert Group started work on MPEG-4 in the mid-1990s, the idea was to create the first digital-media standard geared for interactive and Web content. It evolved with flexible video presentation features tailored to fit varying bandwidth and the ability to mix video, audio, text and animation.

The first version of MPEG-4, which focused on video and audio presentation, was released in 2000. It can deliver video at bitrates as low as 5 kilobits per second and can adjust throughput on the fly to fit available bandwidth. While MPEG-2 needs about 3 mbps of throughput to deliver high-quality video, MPEG-4 can do the job at about 1 mbps, according to backers.

An advanced version that adds the interactive elements was released earlier this year, and another set of enhancements — including key intellectual-property management and protection elements — is expected in the middle of next year.


But all is not necessarily well with MPEG-4. Some major players have sounded strong notes of caution, saying that it took too long to develop. With newer codecs offering potentially better performance, the worry is that MPEG-4 may be relegated to the "Island of Misfit Codecs."

One thing is certain: MPEG-4 content has yet to flood the market. Industry insiders blamed that primarily on the crucial licensing pool, which governs who is paid for the MPEG-4 technology used to support the standard, and how much they receive.

Administered by the MPEG organization, the licensing pool has been delayed for more than a year because of pricing squabbles. It now looks like it will be formed some time in January.

Consequently, many applications developers have hesitated to use the technology until they know the licensing fees they will have to pay. That's one reason RealNetworks Inc. — maker of the most-used streaming media player on the market — doesn't yet support MPEG-4 in its player.

"It's sort of in an 'about-to-be stage' as opposed to the 'released stage,' " said RealNetworks general manager of products and systems Ben Rotholtz.


The delay with the MPEG-4 codec also has given other proprietary codecs — including streaming media formats developed by Microsoft Corp. and RealNetworks — the time to leapfrog ahead.

As evidence, a recent ExtremeTech.com codec review found that the MPEG-4 video presentation came up short, compared with the proprietary codecs used in Microsoft's Windows Media Player 8 or RealNetworks' Real-Player 8. That takes on added weight, given that an overwhelming majority of video content on the Web is formatted for one of these two platforms.

Although Microsoft does support the MPEG-4 video specification in its player, it is one codec among many — and not the best of the lot, according to Michael Aldridge, lead project manager for Microsoft's Windows Digital Media. He said the Windows Media Video 8 codec unveiled earlier this year is about 40 percent more efficient in delivering video than the MPEG-4 version supported in the player.

The software giant is working on a wireless MPEG-4 application with Japan's NTT DoCoMO, but in more mainstream applications, the technology hasn't proven as valuable.

"Certainly we believe in the value of standards for content interchange, which is one of the things people look at the value of MPEG-4 as having," Aldridge said. "But the struggle it has had has been that it is not as complete a solution as other technologies, nor is it state-of-the-art in terms of being able to deliver the highest quality at the lowest bandwidth — which is obviously critical, certainly when you are evaluating the viability of distributing content."

Real is working on an MPEG-4 implementation to make sure it covers all of its codec bases. But Rotholtz also found that the standard may have missed the moment.

"One of the things that happens with standards is by the time they get out there, there are other technologies that have so dramatically superceded them, that they are kind of footnotes in history," he said.


Though it's aimed at covering a wide swath of digital-media usage, MPEG-4's complexity may also prove its undoing. Content creators may choose to implement only select pieces of MPEG-4, making product interoperability an issue.

"It's a 600- to 700-page specification," Rotholtz noted. "There is a lot of stuff inside of that specification that is more sort of academic, as opposed to practical delivery or usage of MPEG-4."

Interoperability is important, but Microsoft has found that content providers place more emphasis on getting the highest quality and lowest bandwidth, rather having a solid digital rights management and security system and gaining the greatest reach. Microsoft's technology can now deliver DVD-quality video at 500 kbps and reaches 350 million users, Aldridge said.

"The content owners are already voting with their feet to a certain extent, because the state of the art and what technology can offer today is already delivering what they need, and I think MPEG-4 is still struggling to pull together all of the pieces," he said.


Aldridge said that although Microsoft supports MPEG-4's video specification, it has no plans to add the interactive-presentation elements.

"We believe there are parallel technologies that have developed as a result of the Internet revolution that have sort of obviated the need for these complex features that are part of the MPEG-4 standard," Aldridge said. "Few people want to abandon the Web standards such as [HyperText Markup Language], and invest in a completely new system that can essentially accomplish the same thing."

Rotholtz agreed, noting that while interactive capabilities often sound great, it's unclear as to whether consumers really want them.

"The big question with all of this is it has got to be pretty straightforward for people to in any kind of mass author for it. And consumers have to demonstrate that, 'Yep, that's the way that I want to work with video,' " he said.

Microsoft and Real's assessment of MPEG-4 may be grim, but there are just as many proponents who argue that the technology holds plenty of promise.

One company developing MPEG-4 based applications is iVast Inc. The Santa Clara, Calif.-based software developer has fielded an MPEG-4-based content platform allowing content owners to convert multimedia to the format, deliver it and then play it.

Microsoft and Real's lukewarm response to MPEG-4 likely stems from the fact that they have a vested interest in their own codecs, said iVast vice president of marketing Kent Libbey.

While the economic slowdown has prompted some customers to cut back or slow their interactive content initiatives, Libbey said he's "hearing a whole lot more interest really on all of our fronts now — both in Hollywood as well as the MSO community, satellite business and more and more, in the enterprise and institutional market."

MPEG-4 has also gained a vote of support from The Internet Streaming Media Alliance, a nonprofit corporation formed to create standards-based interoperability guidelines for companies developing IP-based streaming-media products.

Members — including such heavy hitters as Sun Microsystems Inc., Apple Computer Corp. IBM Corp. and Cisco Systems Inc. — hope the guidelines will lead to better, more economical streaming-media applications.

MPEG-4's many features are the reason why it's anchoring ISMA's first interoperability specification, released in October, according to Tom Jacobs, director of engineering at Sun Microsystems Labs and president of the ISMA.

"The hope of what MPEG-4 represents is that in the way that MPEG-2 brought us digitized, low-cost, high-quality video for the home, MPEG-4 should be able to do that for us for the Internet," Jacobs said. "The bit-rates, the efficiency of the encoding algorithms, are getting better on a regular basis."

That Microsoft and Real are not strong MPEG-4 backers, and chose not to join the ISMA, is problematic. But Jacobs said that may change over time.

"There's a lot of reasons why companies like Microsoft or Real wouldn't support something like this right out of the chute," Jacobs said. "There's people with an early-on stake in proprietary systems, because that's the way they got into the business, and over time they will either be subsumed or build standards-based solutions."

Proprietary codecs may have an edge for now, but history has sided with standards. Witness the battle between MPEG-2 and General Instrument Corp.'s Digicypher codec. Though Digicypher initially won a lion's share of the market, over time broadcasters and cable operators changed over to MPEG-2.

"The people that hold the content out there are the people who hold all of the cards, as much as we might think of it any other way," Jacobs said. "If you are a Vivendi [S.A.] or [AOL] Time Warner [Inc.] or someone else, you would like to be able to choose your partners and know that they are not competing with you."

Claims that MPEG-4's video quality is inferior to newer codecs also need careful examination, according to iVast director of technology Ganesh Rajan. Video quality depends on the content itself, so Rajan said he is "very skeptical about these tests."

His company's own tests show little difference in video quality between MPEG-4 and the player codecs. Rajan added the MPEG-4 committee is working on a new set of tools to update the standard to current performance levels that should be released in the next year. It is also working on upgrades to transport efficiency that may be available by the middle of 2002.

MPEG-4 will soon take another significant step as it is integrated directly onto silicon, probably by the end of this year. Setting the standard in silicon makes it much easier for developers to aim MPEG-4 content for consumer devices.

"That really, to me, is the critical inflection point for MPEG-4," iVast's Libbey said. "When you add to the PC, set-top boxes, personal video recorders, enhanced DVD game consoles, PDAs, mobile phones and so on and so forth — all of a sudden the game changes completely. The ability to reduce very complex computation that is required for the decoding of MPEG-4 to a chipset, which is dollars or cents for that matter, is a key point that MPEG-4 has going in its favor."

As to the idea that other Internet languages can provide for sufficient interactive content, Libbey argues that is not necessarily an easier process because that tack requires programmers to be proficient in a number of languages to piece together interactive content, as opposed to using just MPEG-4.

"The advantage of being able to do that within a single framework – within a single authoring environment and most importantly within a single player device— that's a whole different story. You are not going to get that through a combination of Real, QuickTime and Windows Media," Libbey said. "Economies of scale and scope for MPEG-4 are far better."

MPEG-4 also has applications outside streaming media that are able to handle digital video for any delivery platform – from Internet to broadcast, cable and satellite. iVast is working on several deals with cable and satellite providers.

"They do get the fact that they not only get the immediate short-term gains of efficiencies by way of MPEG-4 compression versus MPEG-2, but they also understand it is a bridge to the future," Libbey said. "And it does open up all sorts of possibilities."


For the cable industry, MPEG-4 is not exactly a top priority. While it might offer significant bandwidth savings for either IP or traditional broadcast transport, other more pressing technology issues make MPEG-4 little more than a technical curiosity right now.

"I don't see that yet and I don't think we will see that for a little while, simply because we have just spent three years investing billions of dollars, and we are just now looking at how to monetize that infrastructure," noted Jupiter Research analyst Lydia Loizides. "MPEG-4 requires upgrades to that infrastructure, and I think people are holding off a little bit on that until some of the stuff we've been working on proves itself out."

Cable Television Laboratories Inc., the industry's technology development consortium, is just starting to examine MPEG-4 codecs and authoring tools, as well as surveying developers to see what kinds of hardware implementations they are cooking up, according to project manager of streaming media tech Yasser Syed. But until there is more of a market push with applications and some silicon hardware to work with, MPEG-4 will stay on the drawing board for CableLabs.

"A lot of this stuff is still on paper. We're still trying to see what starts developing and what is more commercially applicable at this point," Syed said. "We are still kind of at a transition point on it. It's coming from 'Wow, it's a neat technology.' to actual 'Here's some implementation issues and how do we get on hardware," and that type of stuff. We've just got to see how that pans out."