High-definition video development is heating up, fueled by the availability of better display and storage technology, as well as the looming deadline for digital-TV transmission in North America. Recent advances have enabled new video capabilities in equipment ranging in size from cell phones and MP3 players right up to video walls and billboards. Video system designers, lured by the promise of wider screens and higher resolution--and offered support through technology that is already in place or is being deployed--now must fully investigate the issues of implementing HD.

It is important to understand what HD involves, since a variety of display formats can be called high-definition, and some or all of them may be inappropriate in certain applications. TV, with its new digital formats, is an important application, but every type of display is a potential target for upgraded technology. The real question for the developer is not so much whether HD is required, but how to achieve the best possible quality, given the system's display, bandwidth, storage constraints and cost constraints.

Any changeover in technology involves issues that may not be immediately apparent to system developers. Here are a few do's and don'ts to remember as you undertake the transition to HD.

Do

• Be aware of the greater system requirements of the HD format. Decompressed 1080i60 video has six times the data of decompressed SD video, so in raw terms the system has to provide 6x the processing throughput and memory. Also, the advanced codecs that support HD achieve greater compression by employing more memory, I/O bandwidth and processing, pushing system requirements further.

• Focus on audio as well as video, since it figures just as highly in perceived quality.

• Consider the display size. HD is only effective with displays measuring 40 inches or more diagonally; with smaller displays, viewers can't tell the difference between HD and SD. And compression algorithms are lossy, so the decompressed image is less well-defined than the original.

• Realize that programmability allows the system to be scaled and redesigned to meet requirements for different regions and markets.

• Keep in mind that it is important to be able to interoperate with legacy systems. Systems that exchange video within home networks will increasingly require transcoding among standards.

• Analyze the trade-offs in quality, bandwidth and cost. A system-on-chip using DSP can offset some of the cost with the right combination of flexibility and optimized video peripherals and acceleration. Processors that integrate both DSP and RISC cores can partition performance. Programmability provides versatility and scalability for system implementation, and comprehensive software platforms, with audio-video APIs and standard codec algorithms, can speed development.

Don't

Assume that digital video is the same as HD video. The general public may be a bit confused about the difference between HDTV and DTV, but the entire spectrum of digital video applications is enormous, and HD is only a portion of it.

• Consider only one format. HD encompasses a range of display resolutions. The table shows some of the more common DTV formats, though there are others that might be used as well. While most commercial TV broadcasters in North America are implementing 1080i60, other formats may be more advantageous for your application.

• Underestimate market complexity when choosing a processor. All of the ITU/MPEG standards offer variations in implementation, and the changeover from MPEG-2 to H.264/MPEG-4 AVC codecs will be significant in the next few years. Competing standards such as WMV9/VC-1 and China's AVS will also claim their share of applications. Systems such as set-tops may have to deal dynamically with any number of standards and variations, as well as interface to entertainment and gaming consoles and transcode and transrate video for home computer networks.

• Forget that system integration is lowering the cost of HD rapidly, so a system that is not cost-effective today may be so in the next generation. The system developer has to decide whether to support HD now, at today's higher cost, or stay with SD for a year or two until HD component costs are lower and HD demand has increased.

• Let HD hype make system decisions for you. HD is not required for all applications. The real question is how to achieve the best possible video quality for the target display, within system bandwidth and storage constraints and that ever-present factor: cost.

By Jeremiah Golston (j-golston @ti.com), chief technology officer of the Streaming Media Division at Texas Instruments Inc., and Gene Frantz (genf@ti.com), principal fellow at Texas Instruments

    See related chart