The United States established its national standard for terrestrial broadcasts of high-definition digital television, known as ATSC (for Advanced Television Systems Committee), in 1996. The European Union settled on its standard, Digital Video Broadcast-Terrestrial, or DVB-T, in 1997. Japan developed its Integrated Services Digital Broadcasting-Terrestrial (ISDB-T) in the 1980s and adopted it in 2003. But China just finalized its digital television standard in late 2006, beginning transmission with last summer’s Beijing Olympics.
Being late in this particular game is not necessarily a bad thing. It allowed China to take advantage of advances in information-coding technologies that make digital television in China—unlike that in the rest of the world—work well even in bad weather. These technologies mean that China’s digital television can be viewed on the go; it won’t break up even at 200 kilometers per hour—you can watch a broadcast on a cellphone while sitting on a high-speed train. (The United States is only now trying to retrofit its digital-television standard for mobile reception.)
Development of what is formally called GB20600-2006 (GB stands for guo biao , ”national standard” in Chinese) began in 2000 at China’s DTV Technology Research Center, in Beijing, established by my company, Legend Silicon Corp., of Fremont, Calif., and Tsinghua University, also in Beijing. The research team studied the digital-television standards of various countries around the world—including DVB-T and ATSC—looking to adopt the best features of each and improve upon them whenever possible. Their goal: to deliver free-to-air HDTV to the Chinese people in time for the Beijing Olympics.
Digital-television standards differ in two important aspects. One is their capacity in terms of bandwidth, which limits just how much information broadcasters can send. Second, some standards allow not only televisions sitting in living rooms to decode the signal but also ones that are moving fast, such as those on an automobile or a train. And, of course, each standard uses modulation schemes and encoding technologies that reflect the state-of-the-art technology at the time it was developed.
The U.S. ATSC standard uses the 8-VSB, for Vestigial Sideband Modulation Scheme, which means the payload bits are mapped to eight different levels in the vestigial sideband. This scheme meets the needs of a large-screen digital television that’s in a fixed location. In 1996, developers of the standard considered it more important to accommodate a high-definition full motion-picture broadcast (equivalent to a payload bandwidth of about 19 megabits per second) to a stationary digital television screen rather than support mobile television reception. Given that developers were working in an era before smartphones and built-in rear-seat car entertainment systems, this mobile reception capacity didn't seem relevant.
The DVB standard of Europe, on the other hand, uses COFDM (Coded Orthogonal Frequency-Division Multiplexing), which compromises on the bandwidth available to send picture data. This results in a lower digital-television resolution, but it can also support mobile reception.
China’s GB20600-2006 standard can fully support both fixed and mobile systems and send high resolution images, thanks to newer modulation schemes and information-encoding technologies. The standard supports maximum payloads of nearly 24 Mb/s compared with 19 Mb/s in the United States, and mobile reception at speeds greater than 200 km/h compared with similar speeds supported by DVB.
The core technology behind China’s standard is Time Domain Synchronous-OFDM, or TDS-OFDM, invented by Legend Silicon and developed jointly with Tsinghua University.
Any digital-television broadcast signal, from the moment it leaves the transmission tower, is bashed around as it propagates in the atmosphere. Tall structures such as buildings will obstruct, deflect, and weaken this signal; trees and foliage in the signal path will dissipate its strength. If the signal encounters a body of water, the ripples on the surface will interfere with the transmission. The ultimate result is a weak and unpredictable television signal, varying in direction as well as strength. And this is just for a fixed television receiver. If the television unit itself is moving at rapid speeds, then the reception situation gets worse, with severe fade-ins and fade-outs.
The challenge of any digital-television standard—and hence any digital-television receiver that supports such a standard—is to embed in it the technology to smoothly put the bashed-up television signal back together and turn it into a high-quality picture. That would be easy given an infinite amount of money to put into technology at both the transmitter and receiver end, but the commercial success of any standard depends on the affordability of the technology.
To understand how OFDM technology works, imagine a large, high-bandwidth digital payload—for example, a live football-game feed—embedded in the digital-television transmitter signal as it passes through the atmosphere. If this payload is packaged and transmitted as is, without any preemptive corrective mechanisms as it moves over large areas, it will lose strength, the amount of loss increasing as the distance between the transmitter and receiver grows. For mobile reception, it gets even more complicated because the distance continually varies.
Next, the atmosphere through which the signal travels is in a constant state of flux, with both slow and rapid electromagnetic changes that directly affect the progress of a digital-television signal through the air. At the same time, the movement of large and small vehicles on the roads and freeways contributes to atmospheric variations. All these phenomena stretch and twist the payload, as if it were a piece of cloth.
Finally, the multiple reflections from tall structures, trees, and foliage spread the signal over time, bouncing it back and forth and varying the arrival time at its destination. This can create an effect like that of artificially speeding up or slowing down the video/audio playback.
OFDM technology minimizes such debilitating effects on the television signal, rearranging the transmitter signal before it is transmitted in anticipation of the atmospheric effects. When the signal reaches the television receiver, a matched OFDM circuit in the television set can recover the payload—in this case, the live football game—with high fidelity.
In this rearranging, the system organizes the large digital payload: Rather than transmitting the entire payload in a single carrier of one frequency, as is done in FM radio and ATSC systems, it breaks up the payload and distributes it into several smaller carriers of different frequencies, hence the name Frequency-Division Multiplexing. In such multicarrier transmission, each individual carrier is smaller in bandwidth, so the impact of the atmospheric degradation on each individual carrier is also smaller.
It’s not quite that simple, however. Whenever a large number of such subcarriers are grouped together, you create a new problem: that these subcarriers will interfere with one another.
OFDM deals with this new problem with a series of techniques. First, it selects only subcarriers that cannot interfere with each other—that is, they are mutually orthogonal. Second, the digital bits of the payload are distributed among these multiple subcarriers. It’s best not to pack these bits too close to one another but to leave a certain amount of space between adjoining groups of bits. OFDM fills this guard interval with special-purpose bits that prevent one group of payload bits from interfering with an adjacent group, acting like a fence wall. This fence prevents what little atmospheric spreading and stretching still remain from causing a collision.
However, any bit slot allocated to the guard interval is a bit slot that’s not available for the actual payload bit. So the guard interval is an overhead that doesn’t help to increase the payload bandwidth but instead reduces it.
Moreover, because the key to OFDM is in arranging payload bits into multiple subcarriers, and because the guard interval bits are also in the frequency domain, there’s a lot of complex processing that must be done on the receiving end to sort it all out. The high performance of the OFDM technology does come at a cost.
This type of OFDM technology is widely used not only in Wi-Fi and in the emerging WiMax standards but also in almost all digital-television broadcast standards worldwide, including Japan’s ISDB-T, the Qualcomm MediaFLO standard, and South Korea’s Terrestrial Digital Multimedia Broadcasting (T-DMB). It is not part of the United States’ ATSC system.
Even with OFDM technology and additional stages of information coding, it is still a big challenge to reliably receive a digital-television signal at high vehicular speeds. Unlike the satellite broadcast, where there is always a line of sight from the satellite transmitter to the television receiver on the ground, in any terrestrial broadcast there is usually no clear line of sight between the transmitting tower and the television receiver.
When the mobile television receiver begins to move, this problem of not having a line of sight only becomes worse. At high mobile speeds, the Doppler effect adds another element of difficulty. The Doppler phenomenon means that the effective speed experienced by a mobile television receiver vis-à-vis the transmitted digital terrestrial television signal is no longer just the speed of the receiver itself but less than or more than the actual speed of the car. To attain smooth reception at such varying speeds is a big challenge for any mobile television standard.
China’s GB20600-2006 solves this problem partly by how it deals with the guard interval bits. Besides acting as a digital fence, the receiver uses the guard bits—a known sequence—as reference information when it corrects for signal distortion.
There is an important difference between the guard interval bits in China’s GB standard and those used in other OFDM systems. In traditional OFDM, the television receiver also processes these guard interval bits in the frequency domain. It turns out that this choice of frequency domain processing doesn’t work so well when the receiver is in motion—the bits just take too long to decode. So China’s standard instead fills the guard interval with pseudo-noise bits that can be processed in the time domain, which is significantly faster—hence the moniker TDS-OFDM, for Time Domain Synchronous. In COFDM it takes 100 milliseconds for a receiver to lock in on a signal; in TDS-OFDM it takes only 5 ms. The rest of the payload—that is, those bits that are not guard-interval bits—are still processed in the frequency domain, as in COFDM, thereby making the TDS-OFDM a hybrid time/frequency domain processing technology.
In addition to the dramatic improvement in mobile reception speeds with TDS-OFDM, the GB20600-2006 standard uses the most advanced information-encoding technology that’s commercially feasible today on a worldwide basis, known as low-density parity check (LDPC). The LDPC coding technology enables the GB20600-2006 to broadcast across a larger geographic area because the robust encoding scheme protects the payload against degradation.
The success of the new standard was spurred by the readiness of the broadcast infrastructure and the consumer mobile TV entertainment devices in the market at the time of the Beijing Olympics. According to media reports, more than 840 million television viewers in China watched the Olympics opening ceremonies. This gave CCTV, China’s only national television network, a perfect opportunity to show off the new national HDTV standard on buses, trains , and in public squares and amphitheaters.
China formally launched digital terrestrial television broadcasts using the new standard in eight cities for the event, including six host cities and the cities of Guangzhou and Shenzhen, starting with one high-definition program and six standard-definition programs on two channels.
With a television penetration into over 96 percent of households, China is home to over 400 million television sets. Unlike in the United States, where cable television dominates, fewer than 30 percent of Chinese households receive cable. China plans to turn off the analog terrestrial television broadcast in 2015. The country has set itself the goal of converting the households that still rely on only the free-to-air terrestrial television signal—more than 70 percent—into digital television households. China will likely subsidize the cost of conversion by making low-cost converter boxes available.
In late December 2008, China’s State Administration of Radio, Film, and Television announced that China would invest 2.5 billion yuan (US $366 million) to build a national digital TV network based on the GB20600-2006 standard.
In January 2009, Chinese broadcasters began a two-year launch of between 30 to and 40 digital high-definition and standard-definition programs, broadcast mostly by CCTV and local satellite TV stations. At the same time, all those who have a stake in the success of the new standard are expecting a total of 37 large and medium-size cities to launch the first phase of the digital-television service, relaying the CCTV’s HD programs while also broadcasting the SD programs. In the second stage of a massive rollout, 360 prefectures, medium and large cities, and 2861 counties will begin broadcasting standard definition programs on central, city, and county TV stations.
By any measure, this mobile-friendly terrestrial digital HDTV standard is reaching the average consumer in China. While consumers are still anticipating the digital-rich media revolution in other countries, in China it has already arrived.
About the Author
Raj Karamchedu is the director of product marketing at Legend Silicon Corp. He has been a high-technology product and business management professional since early 1994, having started his career at Cadence Design Systems as a signal processing/digital-communication-system design engineer. Until July 2005, Karamchedu was a senior product marketing manager for HDMI products at Silicon Image. His previous product marketing roles were at Philips Semiconductors, Systemonic, and Chameleon Systems. He is the author of It's Not About the Technology: Developing the Craft of Thinking for a High Technology Corporation , published by Springer (2005). He can be reached at email@example.com.