The PC-CE Connection
By Philip De Lancie - Posted Mar 1, 2003

Metadata standards promise easier media interchange, but competing proposals may undermine potential benefits.

March 2003|It's been some 20 years now since the introduction of the compact disc ushered in the era of media as data. And while still images and video have lagged a bit behind audio, the overall trend toward digitization has proven consistent and irresistible. Consumer electronics (CE) devices such as digital cameras, DV camcorders, and DVD players are all part of the picture, evidence that consumers are eager to partake of the benefits promised by digital media. And because bits have become the common language, the personal computer readily serves as a bridge between capture and presentation. Armed with editing and authoring software, photo-ready printers, and recordable/rewritable CD and DVD drives, even non-professionals now have unprecedented control over the manipulation and presentation of content.

There is a lot of overlap between the two fields—consumer electronics and computing—that are converging in this digital media phenomenon. Even so, the two industries sometimes operate as if they are in different worlds, and progress has been slow in implementing universal standards that can facilitate a smooth user experience. One big step forward in recent years was settling on FireWire (aka IEEE 1394 or iLink) and USB as de facto standard interfaces for data transfer. But raw media file interchange is only part of what's required for seamless interaction between CE devices and computers. What's been missing is a standardized approach to the creation, storage, transfer, and display of metadata, information about the "presentation data" that may be passed along with it through each stage from capture to presentation.

Keeping track of contents
The lack of an overarching metadata standard hasn't been much of a problem for highly structured formats such as DVD-Video or CD-Audio, where specifications mandate the inclusion of information about the contents. But with more open-ended formats such as DVD-ROM or CD-ROM on recordable or rewritable media, a user can store pretty much any kind of digital file. Even if playback of those files is supported by a given CE player, it can be challenging for users to figure out what's on the disc when they want to play it back.

One common use for CD-Rs, for instance, is to load them up with MP3 music files transferred on a PC. Playback of such files is now supported by many models of set-top CD and DVD players. But Pieter van Zee, senior architect with the Digital Imaging Group at Hewlett-Packard in Palo Alto, California, points out that unlike the 10 or 20 songs typically found on a normal CD-Audio disc, a CD-R may fit upwards of a hundred MP3 files. He also notes that with the growth of digital cameras and retail services that offer photo discs, JPEG files on disc are also becoming widely available, and set-top player support for additional formats like JPEG is growing.

"Unlike CD-Audio, Video CD, DVD-Video, and other CE-focused disc formats," van Zee says, "there isn't a widely adopted standard for how to organize, access, and play MP3 and JPEG content on a data disc. This means that CE players cannot provide a great user experience that starts up quickly when the disc is inserted and provides user-friendly navigation and playlists. This inhibits widespread use and growth of market demand." Perry Solomon, president and CEO of Alera Technologies in Van Nuys, California, a vendor of recordable/ rewritable drive solutions for CD and DVD, agrees that the lack of a metadata standard is a big problem. "As more and more consumers are storing their memories digitally," he says, "navigating through the collections of mixed digital file types like music, photos, and videos on a CE device such as a DVD player becomes a tedious and difficult task."

______________________________________________________________
OSTA'S LIST OF BASIC REQUIREMENTS

_______________________________________________________________________

The problem, says Michael Aldridge, lead product manager for the Windows Digital Media division at Microsoft, is that there are a variety of inconsistent ways that CD and DVD players read digital media collections (photos, audio, video). "Each interface for finding media, and the viewable information available—like playlists, music metadata, and folders with photos or videos—varies depending on what each device supports," he says. "Many players do not support playlists or advanced playback options, like selecting photographs by date or music by genre. In many cases, this creates confusion for the user about how to find the music, photos, or videos they want."

The difficulties crop up not only with disc formats, but also with memory cards. "Many devices have the same need for consistent access to and playback of digital media content storage on memory cards," van Zee says. "Often this content is created on the CE device and then needs to be transferred to other CE devices or to a PC. For example, digital cameras allow consumers to easily create hundreds of photo, video, and even audio files, but they organize their content in diverse ways, so it is very hard to offer the consumer a digital player that recreates the experience captured by any particular camera."

Competing approaches
Just as there is general agreement on the fact that there is a problem, there is also seeming accord on the need to develop what Aldridge describes as "a standard way for PCs to structure digital media on physical formats and for consumer electronics devices to read these discs. This will eliminate compatibility issues, make startup times for data CDs and other physical formats shorter, and give consumers a consistent, easy navigation experience no matter what format of device they use."

Unfortunately, consensus on the solution breaks down when it comes to a specific implementation. Thus, it looks as if two competing efforts to address the problem will come to market, potentially creating another set of inconsistencies to befuddle consumers. On the one hand, the Optical Storage Technology Association (OSTA) has developed an open specification called "MPV" (for "MultiPhoto/Video") that is based on XML (eXtensible Markup Language). Microsoft and Panasonic, meanwhile, have announced their own metadata file format called "HighMAT."

As described by Felix Nemirovsky, general manager of the OMSG division of Oak Technology in Sunnyvale, California, MPV provides a searchable database that supports a rich presentation of personal collections of content. The specification—the result of a year-long industry-wide development process involving not only OSTA, but also the International Imaging Industry Association (I3A)—is currently supported by at least 21 companies including Alera, ArcSoft, Kodak, ESS Technology, HP, Imation, LSI Logic, Oak Technology, Philips, Plasmon, Rimage, Roxio, Samsung, Sonic Solutions, Ulead Systems, Verbatim, and Zoran. At January's CES show in Las Vegas, OSTA representatives sought to add to the list and heighten MPV's profile by holding private meetings with interested CE manufacturers, media representatives, and industry analysts.

The 1.0 version of the specification, which may be found on the OSTA Web site (www.osta.org/mpv), was released in November 2002 and may be accessed and used without licensing or royalty fees. Also accessible through OSTA's site is the MPVTools SDK 1.0, an implementation of MPV reading, writing, and playback capability. Available at no cost as source code for general use in commercial products, the SDK is in the C language and is suitable for firmware deployment.

"MPV is an open specification," says van Zee, who was OSTA's editor for the specification. "It will greatly enhance the ease-of-use and success of creating and organizing content on PCs, burning it on CDs or transferring to memory cards, and enjoying it on CE devices like DVD and music players that are adding support for new digital media formats. MPV also enhances the interoperability of content created on CE devices with other CE devices and with PCs. MPV-enabled disc burning software and MPV-enabled CE devices and software players are in development for release in 2003."

According to Nemirovsky, the display of the information in the MPV XML file itself will depend on the user interface implementation of the "rendering device" (player). "However, the navigation and presentation will be the same on all devices. The device just reads the MPV file and parses the XML code to understand the metadata structure."

From location to presentation
According to HP's van Zee, five kinds of information are stored in an MPV "album/playlist": location, identity, metadata, relationships, and presentation control. "Location," he says, "is one or more filename(s) or URLs where the content is located. These are file system-specific, allowing MPV to represent content located on media with multiple file systems. MPV supports the use of long filenames created by consumers." Content may also be identified independently of the filename using identity information consisting of what van Zee refers to as "one or more digital signatures. This is useful if the filename has changed or the file has moved and the application wants to rediscover the current filename. It is also useful in Internet applications to avoid moving files unnecessarily."

Relationship information allows related content to be identified and grouped. "For example," van Zee says, "the same image in high-resolution, screen-resolution, and thumbnail sizes can be grouped together. This makes it easy to access the most appropriate content representation and can dramatically enhance playback performance by using small-sized representations." Metadata, meanwhile, is an open category for information about the content. "MPV defines a few basic metadata values, such as title, creator/author, dates, formats, and so forth," van Zee says, "but software and device manufacturers can also define their own metadata."

The fifth information type, presentation control, reaches into the realm of simple authoring, allowing the organization of content into playlists and the definition of presentation effects such as transitions and background music for a slideshow. "Pre-rendered presentation ‘shows' can also be provided," van Zee says, "allowing sophisticated software applications to produce whizzy playback experiences while maintaining access to the content from which they are made." To make it all work, van Zee says, "software applications that burn CDs and DVDs must add an MPV control file to the disc. Typically, this will be generated automatically as part of the disc burning process. The required changes are within the software applications—no OS or PC hardware changes are needed."

As for devices, Solomon says that MPV is compatible with "CDs and DVDs, memory cards, hard drives, and the Internet, including Internet-connected mobile phones." He adds that there is "little or no cost to the manufacturer in incorporating MPV support. DVD player manufacturers will be implementing the MPV player software in the firmware of the player and no change in firmware is required for cameras. At Alera Technologies, we have already started to include OSTA's MultiAudio feature set as part of our DVD-R Quad Cruiser 4X DVD-Recorder."

A proprietary alternative
Given Solomon's assertion that MPV does lots and costs little, there wouldn't appear to be a strong need for an additional, competing approach. But Panasonic and Microsoft don't share OSTA's view that XML is the best basis for a simple content indexing system. "MPV is higher-cost XML-based," says Tony Jasionowski, group manager at Panasonic Technologies in Secaucus, New Jersey. "HighMAT is a lower-cost solution and easier to implement."

Aldridge adds that OSTA did not design their specification as collaboratively with key consumer electronics companies as HighMAT's developers did. "Hence," he says, "few if any consumer electronics companies have announced support to date. HighMAT was under development before OSTA announced their plans, and our approach has been to work hand in hand with the industry to find a solution that works well both for the PC and CE device makers."

Aldridge says that HighMAT discs will contain a metadata directory with information about the contents, such as the title of each file, the file type, and the total number of files on the disc. "HighMAT-compatible devices will be guaranteed to be able to read this metadata file and all of the media on the disc," he says, "allowing them to offer a more consistent navigation and playback experience to users."

Jasionowski describes HighMAT as "a logical format based on ISO 9660." Consequently, he says, HighMAT players and discs are "fully cross-compatible" with non-HighMAT players and discs. The HighMAT information itself will only be used by players whose LSI chips have the ability to decode compressed media data. Supported media formats include Windows Media Audio and MP3 for audio, JPEG for photos, and Windows Media Video, with MPEG-4 video support optional.

Jasionowski says that Panasonic expects HighMAT to become "the de facto standard for recording audio, video, and photo files for PCs and consumer audio-video products." At CES, Panasonic unveiled seven new products featuring built- in HighMAT support, including portable, single-disc, and five-disc DVD players, DVD/CD music systems, and a DVD/ CD home theater system.

Microsoft, meanwhile, has already begun incorporating the ability to create HighMAT directory files into a couple of its software products. "Microsoft included HighMAT support in Windows Media Player 9 Series," Aldridge says, "and in Windows Movie Maker 2, the digital video editing and publishing feature of Windows XP. We also hope other software vendors who create CDs will license this technology."

While Microsoft and Panasonic parent Matsushita are each leaders in their fields, their joint announcement of HighMAT in October 2002 didn't initially set off an avalanche of support. The only other company to lend its name publicly was Fuji Photo Film, and details on the nature of Fuji's plans were not released. One reason for the slow start may have been that publicly available information on HighMAT's specifics has been somewhat sketchy. "There will be a specification published soon," Aldridge says. "It will be available from the HighMAT Web site [www.highmat.com] for both PC hardware developers and CE device manufacturers."

Even without the specification having been made publicly available, Aldridge says that "a number of other companies from various parts of the industry are evaluating support of HighMAT." At CES, 11 such companies, including JVC, Pinnacle Systems, ESS Technology, and Sonic Solutions, announced their intent to support HighMAT in future products.

Coexistence or confusion?
While the existence of two approaches to the content indexing problem complicates efforts to streamline the consumer media experience, Jim Taylor, general manager of the Advanced Technology Group at Sonic Solutions in Novato, California, points out that HighMAT and MPV are not entirely redundant. "HighMAT and MPV overlap to some degree," he says, "but they are designed from different perspectives. HighMAT will work with a wide variety of playback devices, including very simple, very low-cost devices, using a small set of technologies. MPV covers a much broader set of formats and addresses a wider range of issues."

Taylor adds that HighMAT and MPV can "easily co-exist on the same disc." Sonic has announced plans to add support for both MPV and HighMAT to its DVD authoring tools and AuthorScript authoring engine, and other makers of desktop content creation tools may well follow suit. But makers of digital cameras and camcorders, operating under very different design constraints, could be reluctant to try to cover all bases in their devices.

So far there's little evidence that either camp is interested in settling on any approach other than their own. "If other companies introduce MPV support," Jasionowski says, "then consumers will have a choice of both HighMAT or MPV. But we at Panasonic believe that HighMAT will meet most of the consumer market disc playback needs. Panasonic does not have any plans to support MPV at this time." In a similar vein, Aldridge says that to the extent that MPV or any other format becomes "as broadly adopted as HighMAT, it is technically possible to have discs supporting other technologies. But HighMAT is the format we have chosen to support in Windows."

As for the MPV camp, Solomon, Nemirovsky, and van Zee all say that the limited information available so far about HighMAT makes it impossible to draw any conclusions about its merits. "One thing is certain," Nemirovsky says. "MPV is an open standard that anyone can learn about, while HighMAT remains proprietary to Microsoft, Panasonic, and Fuji." It's also certain that no one will benefit from a confusing patchwork of metadata implementations rather than a single, universal standard. But as of this writing, the former seems more likely than the latter.

COMPANIES MENTIONED IN THIS ARTICLE
Alera Technologies, LLC www.aleratec.com
ArcSoft www.arcsoft.com
ESS Technology, Inc. www.esstech.com
Hewlett-Packard Co. www.hp.com
Imation www.imation.com
Kodak www.kodak.com
LSI Logic, Inc. www.lsilogic.com
Microsoft www.microsoft.com
Oak Technology www.oaktech.com
OSTA www.osta.org
Panasonic www.panasonic.com
Philips Electronics www.philips.com
Pinnacle Systems www.pinnaclesys.com
Plasmon www.plasmon.com
Rimage www.rimage.com
Roxio www.roxio.com
Samsung www.samsung.com
Sonic Solutions www.sonic.com
Ulead Systems, Inc. www.ulead.com
Verbatim www.verbatim.com
Zoran www.zoran.com