ATVEF and DDE: Keys to new television services

By now, most television engineers have seen examples of interactive TV — applications that mix conventional television with specially designed Web pages.
Publish date:
Social count:

By now, most television engineers have seen examples of interactive TV — applications that mix conventional television with specially designed Web pages. As broadcasters and networks move toward a Web-based look and feel, it will become important for engineers to understand how the technologies behind these creative tools work. In this month's column, we take a look behind the scenes at one of these technologies: declarative data essence (DDE). The screen shot from “Family Feud” below shows a typical DDE application. The tools used to construct the frame, the text and other images on the screen are adapted versions of well-known Internet protocols such as unidirectional HTTP (UHTTP). ITV displays the TV picture as part of a Web page.

You may be familiar with “Family Feud.” In this show, contestants try to guess the most common responses the audience has given to a particular situation or phrase. In the ITV version, Web technologies and tools such as style sheets are combined with DDE triggers to change the behavior of the ITV content as the game progresses. When the viewer starts to play the game, the entire game is downloaded into the DDE-compliant box. As the show continues, the DDE triggers affect the behavior of the game. For example, these triggers may reveal answers or move on to the next question.

DDE started out as ATVEF, an open specification for creating and delivering interactive TV developed by the Advanced Television Enhancement Forum ( Once the ATVEF specification was substantially complete (ATVEF 1.1r26), the organization turned the document over to SMPTE for standardization. The result was a more complete specification of ATVEF: DDE-1, standardized as SMPTE 363M.

How does DDE work? The answer to this question comes on several levels. First, let's look at how DDE information is transported.

DDE-1 supports two transport modes, Transport A and Transport B. Transport A is used when a return channel is available, and Transport B is used for one-way broadcasting. Applications may be written to support both Transport A and Transport B. This lets the sender take maximum advantage of the transmission channel being used. It is important to note that DDE-1 does not specify a particular transport mechanism. It is a content format. However, DDE is designed to be transported using conventional vertical blanking interval (VBI) technology and new digital transport.

The original ATVEF specification included a model of how to transport ATVEF using NTSC. Transport A uses line 21 of the VBI, following the Text-2 (T2) service specified in EIA 608 (recommended practices for line-21 services). The encoding allows URLs and related data to be encoded in line 21 without interfering with closed captioning and other data that share this line. Transport B uses an IP-over-VBI binding in conjunction with North American Basic Teletext Specification (NABTS) coding. In either case, since the FCC mandates pass-through of line 21, DDE can pass from the source to the viewer unimpeded. It can be encoded on tapes and passed over satellite links as well. (Note that the FCC does not mandate that T2 be passed.)

Figure 1. Block diagram of receiver using DDE. Courtesy of Mike Dolan, industry technical consultant.

The model receiver shown in Figure 1 takes video (at the top of the diagram) and audio (at the bottom of the diagram) from the output of the tuner section of the receiver. Both of these signals pass through the DDE components and on to the screen and speakers of the television or monitor. Triggers and Internet data are delivered using the VBI. The Internet data is stored in a 1 megabyte data cache or buffer. Audio delivered as Internet data is played out via a PCM audio decoder and is mixed with real-time audio coming from the traditional TV channel. This mixed audio is then fed to speakers.

It may be easiest to describe the video chain starting at the output. An on-screen display (OSD) renders Internet content for display on the TV or monitor. The OSD receives inputs from two sources. The first source is an image renderer, which receives its input from various image decoders. In this way, images such as JPEG, BMP and others are mapped onto the screen. The other input to the OSD comes from the HTML4 and ASCII renderer, which derives its input from the HTML4 decoder. The CSS-1 decoder guides the manner in which HTML data is displayed. The HTML4 decoder receives and decodes HTML. This process is very similar to that of the Web browser in your computer. CSS stands for cascaded style sheets. Style sheets describe how documents are presented on screens or in print. The World Wide Web Consortium (W3C) has actively promoted the use of style sheets on the Web since 1994. By attaching style sheets to structured documents, authors can influence the presentation of documents without sacrificing device independence. In other words, CSS provides a page layout into which various HTML display elements can be incorporated, thus providing precise layout control (a feature infrequently used on the Web today).

Finally, triggers that control the timing and presentation of ITV content are received and decoded by the trigger processor. From there, they affect retrieval of information from the Internet, start or suspend scripts, and can affect the behavior of the HTML decoder.

Given that an ITV-capable box can act like a conventional television or an ITV device, it is useful to further explore how triggers affect the operation of these devices.

Enhancements are ITV content added to conventional TV programming. They consist of a sequence of HTML documents. The first HTML document of an enhancement is always instantiated by a trigger. Subsequent documents within an enhancement are instantiated as a result of navigation from the current document initiated either by a trigger or viewer selection.

ITV boxes have two basic states: enhancement active and no enhancement active. Figure 2 shows how a model receiver behaves through a sequence of viewer navigation or triggers. Note that to get from the no-enhancement-active state to enhancement-active state, the box must receive a new enhancement trigger. From that point, the box can change state based either on new triggers or user navigation.

Figure 2. Diagram of DDE enhancement states. Courtesy of John Barkley, NIST.

As time goes on, you may find yourself more involved in ITV issues in your facility. If you operate a conventional television station, you probably have already done what you need to do to ensure that ITV will work properly. However, if you find that ITV Transport B works well but Transport A does not, check to see that CC TEXT2 passes through your station. Some broadcasters strip TEXT2, which is not strictly required for transmission of closed captioning. This will cause a failure of ITV Transport A. In some stations, networks have provided ITV decoder boxes so that broadcasters can view ITV content as it is broadcast. You can use these decoders to troubleshoot problems throughout your facility.

Since ITV content and triggers are encoded in line 21, editing of source tapes can destroy portions of the ITV content, just as it can destroy closed captioning. Also, some programs that rely on a return path can experience significant problems if the latency between the viewer and the Internet is slow.

Brad Gilmer is president of Gilmer & Associates, Inc. He is executive director of the AAF Association, and technical moderator of the Video Services Forum.

Send questions and comments to: