We are in the midst of a revolution with high-quality digital video delivered to mobile devices for the first time. Reliable transmission of video over cellular and land mobile radio channels presents many challenges that are just now being overcome. After the first video compression algorithms were perfected in the 1990s, the next big obstacle was bandwidth. Even highly compressed video requires several hundred kilobits per second, and until recently, only fixed terminal satellite channels had the bandwidth to pass these bit rates over the air.
But with the advent of 3G cellular, 4.9 GHz and broadband 700 MHz systems, there are now wideband channels capable of delivering digital video content to mobile devices. But the mobile radio channel is considerably more hostile than the fixed terminal satellite channel (e.g., DirecTV), and advances in modulation and coding were needed to ensure error-free communications on fading mobile radio channels. The dominant approach today is orthogonal frequency division multiplexing (OFDM) with trellis-coded modulation and interleaving.
Digital video is used in a variety of applications, including security cameras, video conferencing and video subscription services such as Qualcomm’s MediaFLO. In this article, we will examine techniques for audio and video source encoding, list some of the popular digital video transmission standards, and introduce a few of the digital video services available to mobile devices.
Audio and video source encoding. Those working in the field of information theory distinguish between channel encoding, which is the use of controlled redundancy to detect and correct errors introduced on the communications channel, and source encoding, which is compression of the original analog or digital information to be transmitted over the channel.
There are two types of source encoding: lossless and lossy. Lossless source encoding compresses the signal without losing any of the original information. One example of lossless encoding is Huffman encoding, which is used in fax machines. Because most of a fax is white space, it is more efficient to scan each line of the fax and describe the run length of white space rather than describe each individual pixel as white or black. Another form of lossless source encoding is the Lempel-Ziv algorithm, which exploits redundancy in strings of ASCII characters. The ZIP software utility found on most personal computers uses the Lempel-Ziv algorithm, and many software programs, including Adobe Acrobat, use some form of Lempel-Ziv lossless compression.
The written word — and data files in general — must use lossless compression because a single missing bit can render the file unreadable. Audio and video files, on the other hand, depend on human perception so lossy compression techniques can be used without any perceived loss in content. For example, analog television uses the principle of persistence of vision to update the picture at a rate of just 30 frames per second. Similarly, a motion picture is a series of still frames presented at 28 frames per second. The human brain can perceive neither the discontinuity between frames nor the black bar that is displayed briefly during the television vertical blanking interval.
Digital video compression standards typically specify techniques for compressing both audio and video because most video content includes sound. There are numerous standards for compressing audio, including low-bit rate vocoders used by Project 25 and cellular radios that typically operate between 4 and 13 kb/s. Audio that accompanies video is usually of higher quality. For example, the MPEG-1, Audio Layer 3 standard, known as MP3, employs a variety of bit rates, but the most popular for two-channel stereo is 128 kb/s. In contrast, audio CDs, which do not compress the audio, operate at 1.4112 Mb/s for two-channel stereo. A detailed description of the techniques used to compress audio is beyond the scope of this article, but the subject is treated in some depth by Ken Pohlmann in Principles of Digital Audio, Fourth Edition.
Video compression involves two categories of techniques: spatial and temporal. All video is a sequence of frames, each of which is a still image. Spatial compression treats each image separately and compresses it without reference to previous or succeeding images. However, real video has a great deal of similarity between adjacent frames; temporal compression exploits this characteristic to further improve the compression ratio.
To illustrate the need for image compression, consider a 640- by 480-pixel color image that consists of 307,200 pixels. The color of each pixel is represented by a combination of red, green and blue. For high-quality images, each of these colors requires at least 8 bits of resolution (256 levels) for a total of 24 bits per pixel. It follows that the number of bits required to display the image is 24 times the number of pixels, or 7,372,800. For full-motion video, the frame rate is 30 frames per second, so the raw information rate required to transmit such a video is 221,184,000 bits per second. (Good luck finding a radio channel with this much bandwidth.) Using video compression, the same 640 by 480 image can be represented by about 800,000 bits, roughly a 10:1 compression. MPEG-2, the video compression technique used in DVDs and digital television, typically achieves a 50:1 compression ratio using both spatial and temporal techniques.
The most efficient image compression algorithms exploit limitations of the human eye and brain. Nearly all video compression methods in common use apply a discrete cosine transform (DCT) for spatial redundancy reduction. Other methods, such as fractal compression and discrete wavelet transform (DWT) are areas of active research, but are rarely used in practice. For more information on compression algorithms, see Digital Video Compression by Peter Symes.
Video compression and transmission standards. In addition to MPEG-2, digital video compression standards include MPEG-4, a newer standard optimized for video transmission over the Internet. Other transmission standards include the following:
- ATSC Digital Television
ATSC is the digital television standard for North America. It uses 6 MHz television channels, 8-Vestigial Sideband (8-VSB) modulation, and a gross bit rate of 19.4 Mb/s. The ATSC standard has been criticized for employing a modulation and coding technique that is not well-suited to mobile television (in contrast to OFDM).
- Digital Video Broadcasting (DVB)
DVB is a family of standards adopted by European countries for digital television emanating from terrestrial towers and satellites. The DVB standards use OFDM, which is the same modulation technique used in Wi-Fi and WiMAX. DVB-T is the terrestrial broadcast standard for stationary receivers, while DVB-H is the terrestrial standard for mobile devices. DVB-S2 is the newest standard for direct broadcast satellite to fixed terminals while DVB-SH is the standard for video from satellites to mobile and portable devices.
- Integrated Services Digital Broadcasting (ISDB)
Japan chose its own set of standards for digital video broadcasting that fall under the ISDB family. They are similar to DVB and use OFDM.
Several new digital video services are being launched in the U.S. and Canada. Qualcomm’s MediaFLO provides digital video service to cellular phone subscribers through AT&T Mobility and Verizon Wireless. The MediaFLO waveform is a proprietary version of the 6 MHz-wide DVB-H standard. The 50 kW ERP signal transmits on television Channel 55 (716-722 MHz) and a handful of transmitter sites are required in each metropolitan area. The subscriber cell phone includes a special 700 MHz receiver for demodulating and decoding the MediaFLO signal. Digital video services over normal cellular phone channels also are available, but these systems have been less popular to date.
Other digital video services broadcast over satellite using the DVB-SH standard. These include ICO-G, which launched its first geostationary satellite in 2008 and will offer service in 2010, Terrestar, and Mobile Satellite Ventures.
Jay Jacobsmeyer is president of Pericle Communications Co., a consulting engineering firm in Colorado Springs, Colo. He holds BS and MS degrees in electrical engineering from Virginia Tech and Cornell University, respectively, and has more than 25 years of experience as a radio frequency engineer.