Electronic Design

Software Development Kits Speed MPEG-4 Product Designs

These kits simplify the introduction of high-quality, streaming media to the Internet.

The term streaming media is on everyone's lips today. While content providers are eager to boost revenues, potential users hope to download movies of their choice from the Web instead of driving to a local video store. Such visions can now be realized due to MPEG-4, a rapidly emerging media-interchange technology format.

Quick to recognize the potential of this technology, e-Vue Inc. of Iselin, N.J., is introducing a suite of MPEG-4 software development kits (SDKs). With these kits, designers will be able to equip their products with the basic MPEG-4 capabilities.

To understand MPEG-4's importance, it helps to look at how it emerged. Until now, proprietary formats such as Real, Microsoft, and QuickTime have hindered the prompt realization of a ubiquitous streaming network. Each of these formats produces different versions of a streaming-media technology. Although there are tools that bundle all three, there are no standards in place to deliver the content.

What's essential for any interoperable network is a common-media-interchange format. Representative of this is the way the Internet has built upon its early principles of development. Expanding upon TCP/IP, which arrived in the 1980s, made possible file transfer and e-mail systems. Also, HTTP is considered the foundation for developing browser technologies, a process begun in the mid-1990s. Since these open standards became global, the networks and interconnections tying everything together became interoperable. They achieved this despite differences in carriers and equipment.

Some eight years in the making, MPEG-4 is poised to furnish just such a solution for streaming media. It's well on the way to becoming a broad, common international standard for packaging media as well as for transporting that media over arbitrary networks. In addition, MPEG-4 will soon be able to travel across all networks.

"The only way people will get anything done with MPEG-4 is with software development kits," says John Lynch, COO/CTO of e-Vue. "With the MPEG-4 standard, there is a need for everyone to deliver interoperable functionality to the marketplace in a similar time frame. And e-Vue is providing the codec and security software-development kits that will allow people to do that."

MPEG-4 supports powerful new features for both content creators and users. These enhancements include audio, video, image, 3D, animation, and interactive capabilities. Besides the traditional audio and video coding supplied by MPEG-1 and MPEG-2, MPEG-4 offers vastly increased range and functionality.

Beginning with content, the kits enable designers to encode media during the "create" phase. Next comes the "deliver" stage, where content is transmitted over any of the familiar networks. Finally, the content is forwarded to those who "consume" (Fig. 1).

A designer may be dealing with any or all of the three different networks, which include cable/satellites, IP, and wireless. Under "create," the first step is to encode the audio/video/image media into MPEG-4-compliant "elementary streams." The next task is to "author" these streams by combining them to create an .MP4 file. After this is the "publish" phase, in which security and network-specific information are added.

The kits let designers take advantage of MPEG-4's advanced technologies for each of these steps. For instance, encoding may use rectangular "frame-based" video encoders or "arbitrary-shape" encoding.

Similarly, the authoring step can package only the audio, video, and image information into one file—or also include 2D, 3D, animation, and interactivity information. The publishing step can be skipped altogether if security and networking aren't required. Although these are all nice options to have, there are a lot of them. So an obvious question might be, "what's important?"

Fortunately, over the past several months there has been a focus on deploying the most important components of the MPEG-4 standard first. Audio, video, image, and security are the top priorities. While having a global standard for audio, video, and imagery is wonderful, equally significant is the ability to protect this content. Security in MPEG-4 is known as intellectual-property management and protection (IPMP), a benefit afforded by the new SDKs.

Using e-Vue's SDKs, designers can take advantage of MPEG-4's concept of "objects," a feature that wasn't included in earlier MPEG standards. Each of the various media objects are dealt with independently. Audio objects, image objects, video objects, and even security objects can be added one at a time and composed in the player to create the final scene. This means that different parts of the scene can be encoded and transmitted separately, then brought together later in the player. MPEG-4's flexibility can be tapped using the development kits, allowing virtually any number of objects in a scene.

Another benefit of MPEG-4 is that different codecs can be used for the different media types. An example is how MPEG-4 uses discrete cosine transforms (DCT) codecs for video objects, whereas the "VTC wavelet" codec is used for still-image objects. To further illustrate the concept of objects, let's begin with two images. One, a video sequence of a newscaster, has been segmented into an arbitrarily shaped (nonrectangular) video. The second image, an object of a boat on a river, is also being sent to the player. Once it receives both objects, the player puts them together to create the final scene. This is a trivial example considering the many different audio, video, image, 3D, animation, interactive, and security objects that can be used.

To help manage these options, the kits employ the important "profiles" initially being used for media types and system components. Media profiles describe the features for the audio, video, and image codecs. For example, a simple video profile is what we're most familiar with. It's a conventional rectangular video "frame." System profiles define the level of "object processing" required (e.g., 2D, 3D, animation, interactivity, and security). Profiles enable designers of cell phones, set-top boxes, PCs, PDAs, and similar devices to choose the subset of features that they need. Industry consortiums such as the International Streaming Media Alliance (ISMA) and the Wireless Multimedia Forum (WMF) are actively selecting the simple MPEG-4 profiles for use in their markets.

"The object framework provided in MPEG-4 is fundamentally important, but it will not be fully exploited in initial deployments," Lynch says. "What is important now is quality and security."

Also, an infrastructure that can fully benefit from all the flexibility available in object-based media doesn't yet exist. Initial MPEG-4 profiles will therefore make it possible for simple and high-quality audio, video, and images to be securely delivered over networks. They also set the stage for more advanced versions of MPEG-4 in the future.

Security is essential because no content providers are going to put anything of quality on the Internet if they're likely to be deprived of the revenue that rightfully belongs to them. Anyone familiar with Napster understands this.

With IPMP, MPEG-4 takes important steps to create a common framework for sharing media in a secure way. IPMP comprises four components:

• Encryption: Encryption of a bit stream is designed by the creator so that the contents cannot be decoded by a reader without knowing the original key used.

• Key management: One of the most difficult issues is figuring out a way to protect the key. Key management is a young and quite challenging technology. If a key is lost, someone may obtain unauthorized access to the media.

• Policy management: This pertains to the rules that the content provider establishes for the various ways the content user can experience the media. Customers may be permitted to hear it, see it, save it, play it on different devices, share it with friends, or enjoy some mix of all of these.

An example of a very simple policy is one for set-top boxes and TV sets, where the viewer simply pays per view to see the media. The content is viewed only once, and it cannot be saved. That's a simple policy that could be loaded into a PC. A customer can pay per view or pay to download the material and the right to see it several times. Or, in a super-distribution model, everyone can share copies of the content and acquire rights to experience it according to their own preferences.

Policies can become quite complex. As a result, languages have been developed to express policy information.

• Watermarking or fingerprinting: These are really two aspects to a single technology, used in two quite different ways.

With watermarking, an indelible mark is placed within the media to represent the content owner. If the media is a Disney movie, for example, Disney imprints a watermark identifying the copyrighted content as Disney property.

Fingerprinting, on the other hand, is indelible information that's added to the content. Its purpose is to identify the users who have played the content. Every time the content is viewed, the user's identity is embedded in the media. Fingerprinting is useful for tracing stolen or inappropriately used content.

With regard to the elements listed above, the primary focus is on deploying the encryption/decryption and key management that protects the key (as well as simple policies that let users download and play content several times). Policy management languages and watermarking are more advanced aspects that will be introduced in later generations of MPEG-4 products.

The e-Vue MPEG-4 toolkit embraces many elements, including audio, video, images (2D, 3D, and others), interactive objects, and security. But audio, video, image technology, and security will be deployed first.

For this to happen, what's needed are encoders and decoders, encrypters and decrypters, and key management to build in the necessary security. The new software development kits offer all of these advantages. They deliver the software source code, documentation, and examples to MPEG-4-compliant codecs so developers can get their products to market quickly. In addition, e-Vue provides an IPMP/DRM (digital rights management) kit. Finally, there's an MPEG-4 file-format development kit. The file format is the "streamable container" that MPEG-4 uses in all three levels of MPEG-4 (creation, delivery, and consumption).

If you build content-creation products, you'll want encoder, security, and file-format kits. When developing servers, you may already have most of what's necessary to configure the final product. But you'll need the MPEG-4 file format, which is the container for all of the different media objects, equipping the server to read MPEG-4 files.

In players like set-top boxes, TVs, PCs, and PDAs, the required development kits are the file format (MP4FF), audio decode, video decode, image decode, object-descriptor (OD) decode, and IPMP decode (Fig. 3). The binary format for scenes (BIFS) decoder is an advanced system feature that enables 3D animation and interactivity. It will appear in later versions of MPEG-4.

Summing it up, Lynch says, "We are not focusing on the full range of MPEG-4 because that's not what's needed now. What's essential now is MPEG-4-compliant audio, video, image, and security, plus the file format which is the packaging container for MPEG-4."

Price & Availability
The e-Vue software development kits will be available this quarter. They cost from $0.25 to $1.00 per kit per unit shipped in volume quantities.

e-Vue Inc., 33 Wood Ave., Iselin, NJ 08830; (732) 590-0102; fax (732) 452-9726; Internet: www.e-vue.com.

Hide comments


  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.