Digital Television Station and Network Implementation GRAHAM A. JONES, JAMES M. DEFILIPPIS, MEMBER, IEEE, HANS HOFFMANN, MEMBER, IEEE, AND EDMUND A. WILLIAMS, SENIOR MEMBER, IEEE Invited Paper

The comparatively recent introduction of digital transmission provides superior video and audio quality and increased capabilities for new services. However, for many years broadcast stations and networks have been transitioning to digital systems for various aspects of production and distribution. The introduction of digital TV has required further major changes to broadcast stations and networks, with large capital investments. This paper describes some of the considerations in implementing digital systems for studios and transmission, and compares broadcast service models adopted in various parts of the world. Keywords—Advanced Television Systems Committee (ATSC), digital audio, digital video, digital television (DTV), Digital Video Broadcasting (DVB), high definition televison (HDTV), Integrated Services Digital Broadcasting (ISDB), multiplex, standard definition television (SDTV).

I. INTRODUCTION In the previous major change in broadcasting, from black-and-white to color television, existing black-and-white receivers could still display programs broadcast in color, and new color receivers could display legacy black-and-white transmissions. This backward and forward compatibility considerably eased transition arrangements for the industry and consumers. However, this does not apply for the change to digital television (DTV) transmission, where the digital and analog broadcasts are incompatible and new equipment is required both to transmit and receive the new services.

Manuscript received June 6, 2005; revised September 26, 2005. G. A. Jones is with the National Association of Broadcasters, Washington, DC 20036 USA (e-mail: [email protected]). J. M. Defilippis is with the Fox Technology Group, Los Angeles, CA 90035 USA (e-mail: [email protected]). H. Hoffmann is with the European Broadcasting Union, Geneva, Switzerland (e-mail: [email protected]). E. A. Williams, retired, was with the Public Broadcasting Service, Alexandria, VA 22314 USA (e-mail: [email protected]). Digital Object Identifier 10.1109/JPROC.2006.861003

Therefore, in the United States and elsewhere, most broadcast stations have been required to maintain their existing analog transmissions during the transition period, while adding new digital services in accordance with the DTV standard, as illustrated in Fig. 1. Arrangements for the DTV transition are being implemented with different emphasis in various countries, depending on network and station arrangements, government regulations, and market conditions. In the United States, recovery of spectrum associated with analog transmissions is a prime political and technical consideration. In Europe, the use of this “digital dividend” is not yet decided, but current indications are that it will largely be used to develop further broadcast services. The move to digital transmission is not the only task for broadcasters. This paper outlines some of the many implementation issues addressed by terrestrial broadcast stations and associated networks as they have transitioned to DTV, and indicates some of the approaches adopted. Detailed descriptions of the compression and transmission technologies, as well as implementation for digital cable and satellite delivery, are provided in the respective papers for those topics in this special issue of the PROCEEDINGS OF THE IEEE.

II. DTV EFFECT ON THE PROGRAM CHAIN The term DTV is generally applied to a system in which television programming is delivered digitally to the viewer at home. The various DTV standards provide the capability to deliver superior video and audio quality, eliminating the noise, distortions, and ghosting that can degrade analog television. DTV also enables new features and services that may include: • high definition television (HDTV) with higher resolution and wide-screen images; • 5.1 channel surround sound audio;

0018-9219/$20.00 © 2006 IEEE 22

PROCEEDINGS OF THE IEEE, VOL. 94, NO. 1, JANUARY 2006

Fig. 1.

System impact of DTV standard.

• multicasting—delivery of multiple standard definition television (SDTV) and/or HDTV programs in one broadcast channel; • electronic program guide; • data broadcasting content. A DTV channel may carry either HDTV or SDTV programming, or possibly both HDTV and SDTV simultaneously. The number of programs depends on the bandwidth (actually the bit rate) allocated for each program service. The abbreviations DTV, HDTV, and SDTV are often used imprecisely or incorrectly. DTV encompasses both SD and HD television but, strictly speaking, the terms HDTV and SDTV relate only to the resolution of the video image and, historically, both HDTV and SDTV have been produced and broadcast in analog form as well as digital. Today the trend is strongly toward all-digital systems for both SD and HD, although analog SD television transmissions are still ubiquitous worldwide and analog HD transmission by satellite is still operational in Japan (although closing soon). The enhancements of DTV are enabled by the Advanced Television Systems Committee (ATSC), Digital Video Broadcasting (DVB), and Integrated Services Digital Broadcasting (ISDB) standards. These are covered in the respective papers elsewhere in this special issue of the PROCEEDINGS OF THE IEEE. The standards relate largely to the format and content of the transmitted signals. Nonetheless, DTV has a major effect on many other parts of the broadcast system. As an example, Fig. 1 shows a much-simplified program chain for the legacy analog National Television System Committee (NTSC) [1] and ATSC DTV terrestrial broadcasting from a local U.S. television station, taking programming from a network, and transmitting over the air. The ATSC standard applies directly to the DTV signal emitted by the station, but the introduction of DTV affects the whole of the program chain including acquisition, production, postproduction, distribution, and transmission, in several ways, as follows. A. Video Format Where HD DTV is allowed or required by the regulatory authority, broadcasters typically need to implement new JONES et al.: DIGITAL TELEVISION STATION AND NETWORK IMPLEMENTATION

equipment for the appropriate HD video format throughout the program chain, from image acquisition through to encoding for transmission. Depending on the DTV standard and regulations for the country, there is a choice of multiple formats based on 1080 or 720 active lines—each with various advantages and disadvantages, but outside the scope of this paper to discuss. The HD formats always use a widescreen picture aspect ratio of 16 : 9. For SD DTV, the choice of digital video format is usually based on the number of lines of the legacy analog system used in that country. Countries using the 525-line NTSC standard normally use a 480 active line DTV format, while countries using the 625-line analog standards known as Phase Alternating Line (PAL) [1] or Sequential Couleur Avec Memoire (SECAM) [1] normally use a 576 active line format. The SD DTV formats may be implemented in either 4 : 3 or 16 : 9 aspect ratios, with a choice (in some countries) of interlaced or progressive scan; this requires changes to production, and monitoring equipment throughout the program chain. The frame rate for both HD and SD DTV programming in a given country is usually based on that for the legacy analog system, 25 frames per second (fps) or (approximately) 30 fps (related to the 50- or 60-Hz power frequency for the country). Progressive scan formats may use twice the frame rate of interlaced formats. In 60-Hz countries, broadcasters may also choose to produce and/or transmit some programming using the 24 fps rate with progressive scan, usually associated with film. This latter choice requires changes to equipment for production and monitoring at various points in the program chain. In 50-Hz countries, film-based material for television is usually produced and run at 25 fps. B. Audio The main effect of DTV on audio systems is the possibility for 5.1 channel surround sound. DTV also allows for multiple languages, audio for the sight-impaired, and other services. Legacy analog systems are usually designed for two-channel stereo audio (sometimes with provision for a second, foreign language service). Therefore, production and distribution for the additional channels for surround sound, 23

plus possibly more for other services, requires significant changes to equipment throughout the program chain. There are also issues related to audio processing and dynamic range that require audio for DTV to be handled differently from that for analog transmission.

provide enhanced production and creative capabilities and potentially enable more efficient operations. For these reasons, the transition to digital video and audio systems had been underway for many years—long before the change to DTV transmission.

C. Multiple Program Services

III. SHORTCOMINGS OF ANALOG

Where a broadcaster chooses to transmit multiple programs on one DTV multiplex, expansion of the playout and master control facilities for the originating station or network (where the program is finally assembled and distributed—see Section X) are required to accommodate the additional services. D. Electronic Program Guide and System Information The inclusion of an electronic program guide (EPG) in the DTV transmitted bit stream, plus the other service information associated with the program, requires provision of the sources needed to produce and update the information needed. It also requires methods for communicating it from those sources to where the DTV bit stream is generated. This information enables a DTV receiver to rapidly tune to a selected channel and to display program guide information for the viewer. For more information, please see the paper in this issue on the ATSC transport layer program and Program and System Information Protocol (PSIP). The DVB and ISDB systems have similar features for electronic program guides, although implementation arrangements are different. E. Datacasting Where a broadcaster chooses to include data broadcasting service in the DTV multiplex, a completely new production unit may be required to produce such services, and a new infrastructure may be needed to distribute and integrate them into the DTV service. More information on data broadcasting services is provided in related papers in this special issue. F. Digital Equipment and Infrastructure As explained in Sections III and IV, in order to maintain optimum quality and exploit the full capabilities of the DTV system, both transmission and studio systems should be based on digital signal processing and, in particular, component digital video (see [2] and [3] for a general introduction to digital video and audio). Most current systems for HDTV have been designed from the start using digital equipment throughout the program air chain. However, much SDTV programming still originates from legacy systems originally designed to supply signals to NTSC, PAL, or SECAM analog transmitters. Such systems may still include some analog equipment for production or other parts of the system and archived program material may only be available in analog form. Therefore, depending on their age, current SDTV program chains may not be digital throughout. Irrespective of the transmission format, there are many advantages to using digital studio equipment. Not only can digital component video and digital audio production, postproduction, archive, and distribution systems greatly improve image and sound quality. They can also often 24

A. Composite Video The traditional analog television standards define not only the transmission arrangements but also the composite video formats used to produce, record, and distribute analog programming. Composite video has certain fundamental picture quality constraints and artifacts caused by the method of carrying the color information on a subcarrier within the bandwidth of the luma (brightness) signal. B. Noise and Distortion Analog video and audio signals suffer degradations due to the introduction of noise and linear or nonlinear distortions during processing and distribution. This applies particularly in analog recording systems, where the build-up of noise and distortions severely limits the number of generations of recording that may occur while maintaining acceptable quality. C. Restricted Capabilities Some desired operational or engineering capabilities or functions are impossible or impractical to achieve using analog signal processing. IV. MIGRATION FROM ANALOG TO DIGITAL A. Digital Studio Islands The quality and capability restrictions of analog video and audio drove the industry to develop digital solutions for several types of studio equipment. Such equipment performed functions not possible with analog processing, but they generally stood-alone as digital “islands” in an otherwise analog system, with conversion back to analog video and audio signals for all interconnections. Examples of early digital video equipment islands include: • digital timebase correctors 1973; • frame synchronizers 1974; • digital video effects 1978; • digital videotape recorders (VTRs) 1986. Similar developments for audio during the 1970s and 1980s included: • digital delays, reverberation, and effects units; • digital audio mixers; • digital audio tape recorders; • digital hard disk recording and editing. Digital video and audio equipment such as that listed above was widely adopted by broadcasters to meet particular functional needs. In a few cases, such islands were interconnected with proprietary methods for digital coding but the absence of practical standardized interfaces hindered development of fully integrated digital systems. PROCEEDINGS OF THE IEEE, VOL. 94, NO. 1, JANUARY 2006

B. Bandwidth Current specifications for uncompressed signals require the following data rates for baseband serial digital bitstreams on the interfaces: Stereo (2–channel) audio

3 Mb/s

SD component video (10 bit, 4 : 2 : 2)

270 or 360 MB/s

HD component video (10 bit, 4 : 2 : 2)

1.485 Gb/s.

These bit rates require system bandwidths greatly exceeding those for carriage of equivalent analog signals and, for many years, the lack of suitable high-speed electronic components impeded development of serial digital interconnections. C. Digital Video and Audio Standards Interoperable digital equipment and integrated systems depend not only on suitable high-speed interfaces but also on the relevant open technical standards being established. Some key milestones in this process during the 1980s and 1990 were as follows. 1) Digital Video Sampling and Format: A critical stage was the development in 1982 of an international standard for the sampling and format of 525- and 625-line digital video systems—CCIR Recommendation 601, now published as ITU-R BT.601 [4]. This specification was the basis of many subsequent digital video developments and led to the first standardized digital VTR (type D1) and other digital video equipment. 2) Serial Digital Audio Interface: In 1985, a digital audio interface standard was agreed jointly between the Audio Engineering Society (AES) and the European Broadcasting Union (EBU). The AES/EBU specification [5] allows distribution of uncompressed stereo audiosignals as a serial digital bitstream. This standard enabled single-cable interconnections for digital audio and laid the foundation for all subsequent integrated digital audio systems. 3) HDTV Production Standard: Following several years of debate, in 1990 the International Telecommunications Union (ITU) published a recommendation of parameter values for HDTV standards for production and international program exchange, now published as ITU-R BT.709 [6]. This established a common interchange format for HDTV production, which, after further revision, was based on a scanning standard of 1920 pixels by 1080 lines. Later a further standard of 1280 pixel by 720 lines in progressive scan at a frame rate of 60 Hz was also adopted by the ITU [7]. 4) Serial Digital Video Interface: Early digital video interconnections were made using bit-parallel interfaces but this arrangement was found to be impractical for long cable runs, as needed for many video systems. A serial digital interface (SDI) for video was conceived and developed during the late 1980s by the EBU and the Society of Motion Picture and Television Engineers (SMPTE), and, in 1993, SMPTE published its standard for a video serial digital interface (SDI), JONES et al.: DIGITAL TELEVISION STATION AND NETWORK IMPLEMENTATION

SMPTE 259M [8]. This specification relates to video conforming to ITU-R BT.601 sampling and allows distribution of uncompressed SD digital video signals over a single cable. ITU adopted a similar standard called ITU-R BT.656 [9]. These standards enabled the implementation of large-scale integrated digital video systems. 5) HD Serial Digital Video Interface (HD-SDI): The equivalent standard for an HD-SDI, SMPTE 292M [10], was published in 1996. The specification defines the interface for 1080-line and 720-line video and allows distribution of uncompressed HD digital video over a single cable. This standard was key to the practical implementation of HD digital systems. There are numerous other important standards and recommended practices relating to all aspects of digital video and audio formats, signals, measurements, and equipment characteristics—far too many to describe here. The majority has been developed by SMPTE or the ITU, with others from the AES, Consumer Electronics Association (CEA), IEEE, and the Society of Cable Telecommunications Engineers (SCTE), and such standards are listed on those organizations’ Web sites [11]–[16]. V. STANDARD DEFINITION SYSTEMS Following establishment of the various digital video and audio standards in the 1980s and 1990s, broadcast equipment manufacturers developed an increasing range of digital equipment with serial digital inputs and outputs. This allowed the implementation of SD digital video and audio production, postproduction, network, and station systems, with superior performance and functionality that eliminated the shortcomings of analog mentioned earlier. During the past 15 years, the trend has therefore been to implement fully digital studio systems for new SD facilities—whether intended to supply signals to analog transmitters, DTV transmitters, cable, or satellite. There are, however, still many legacy broadcast stations and facilities in existence based at least partially on analog equipment, particularly for signal distribution and routing within the facility. A. SD Video and Audio Production Equipment The types of equipment typically included in SD digital video production and postproduction systems include cameras, camcorders, videotape recorders, video servers, editing systems, archives, video switchers, routers and other distribution equipment, electronic graphics and character generator systems, digital video effects, frame synchronizers, standards converters, and reference timing equipment. For audio systems, digital equipment may include mixing consoles, CD players, digital audio tape and disk recorders, hard disk recorders, audio delay and reverberation, audio effects units, routers, and other distribution equipment. It is outside the scope of this paper to discuss these types of equipment individually, but such equipment replicates and exceeds the capabilities and performance of previous analog equipment in virtually every respect. It should be noted that some organizations continue to use audio mixing consoles based on analog processing in what 25

are otherwise largely digital video and audio systems, or even maintain analog audio distribution in parts of the plant. Analog audio mixers are capable of excellent performance and, in some cases, the extra cost of replacing analog audio equipment with a fully digital system may not be justified. B. SD Origination for DTV Before and during the transition to DTV, networks and stations in most countries have employed studio and distribution systems built to produce 525-line or 625-line programming for legacy NTSC, PAL, and SECAM transmitters. As DTV transmission has been introduced, many broadcasters have initially used the same studio systems to produce programming for SD transmissions, with few changes in processing apart from conversion to digital signals (if previously in analog at the master control output) and encoding for the DTV standard. Such programming, if produced and distributed entirely in the digital domain, using the ITU-R BT.601 standard, will benefit from many of the advantages of the DTV transmission standard. However, for analog-originated programming, the ultimate quality will usually be limited by the fundamental restrictions of the analog systems and equipment. One change that may be needed for DTV is conversion of production and monitoring equipment to handle the 16 : 9 picture aspect ratio (see Section VII). While 16 : 9 transmission is not mandatory for SD DTV, some broadcasters may choose to transmit all or most DTV programming in 16 : 9, particularly as the number of widescreen consumer television displays increases. Broadcasters in Australia and the United Kingdom have already made this SD aspect ratio change, and others in Europe are following soon. However, most U.S. broadcasters still produce and broadcast SD programming in 4 : 3 aspect ratio, both for analog and DTV transmission. VI. HIGH DEFINITION SYSTEMS HDTV programming requires installation of completely new production, postproduction, and distribution equipment and systems—typically a major capital investment for a broadcaster. A. HD Video Production Equipment Early HD production systems, developed in Japan in the 1970s and 1980s, were largely based on analog technologies, requiring wide-band analog video component interconnections over multiple cables. Equipment tended to be cumbersome and very expensive and overall production capabilities, e.g., video effects, electronic graphics, and editing, were limited compared to SD systems. In 1988 the first HD digital component VTR, the HDD-1000 was introduced, but systems were still implemented using digital islands with mainly analog interconnections. During the 1990s, manufacturers developed an increasing range of HD video equipment including new cameras and HD VTRs, including most notably the D5 HD from Panasonic (1995) and HDCAM from Sony (1997). The establishment of the HD-SDI interface in 1996 allowed 26

practical all-digital HD systems to be implemented for the first time. However, prices for such HD production systems were still much higher than equivalent SD systems, and deployment, for the most part, was limited to a small number of specialist production companies, major production centers, and networks in countries where HD broadcasting was being introduced. Over the past several years, the range of HD equipment has expanded greatly, with new recording and storage formats and advances in all types of acquisition, production, postproduction, and distribution systems. At the same time, prices have fallen considerably. This has made it possible for program producers, networks and local broadcast stations in various parts of the world to implement cost-effective HD portable and studio systems with a range of capabilities similar to or exceeding those for SD facilities. B. HD Origination for DTV 1) United States: Since the advent of DTV broadcasting in the United States in late 1996, the major networks and many independent producers have installed new HD production equipment and systems. The majority of U.S. primetime programs and major sporting events are now produced in HD, using systems that are substantially digital throughout. Many scripted drama and comedy shows are produced in the 1080-line progressive format at 24 fps, with some acquisition electronically and some on film that is transferred to video for editing and distribution. For live programming and sports, some broadcasters use 1080-line interlaced at 29.97 fps, and others use 720-line progressive at 59.94 fps. The major U.S. networks all distribute an HD program service to their owned and affiliated stations. Most stations owned by the networks, and very many affiliates, have equipment to allow such network HD programming to pass through and be fed to the DTV transmitter. Arrangements for this are discussed later in Section X. Until recently, with some notable exceptions, U.S. local broadcast stations had not installed significant amounts of HD studio production equipment. However, an increasing number of local broadcast stations are now equipping HD studios and systems for local news and other productions. Other stations rely on upconverting their local SD programs to the HD format to enable local programming to be intercut or mixed with the network programs. Upconversion is a process for increasing the number of pixels and lines in an image by interpolation from existing picture information. 2) Canada: The first HDTV broadcast was in 2003 and HDTV is now provided on both cable and satellite channels. Terrestrial DTV was launched officially by CBC early in 2005 with some HDTV programming, and other broadcasters have announced availability of HD signals. Few shows are, however, currently produced in HDTV. 3) Mexico: The ATSC standard was adopted in 2004. As yet, HDTV use is limited but terrestrial DTV deployment is underway and some cable operators already carry HDTV programming. 4) Korea: Test transmissions of terrestrial HDTV in South Korea started in 1998. Since then, an increasing amount of PROCEEDINGS OF THE IEEE, VOL. 94, NO. 1, JANUARY 2006

HD programming is being produced and broadcast using the 1080-line interlaced format at 29.97 fps. C. HD Origination for DTV—Japan The Japan Broadcasting Corporation (NHK) and Japanese manufacturers were pioneers in HDTV production in the 1980s, with analog systems using the 1125-line Hi-Vision system and later with digital systems and equipment. In 1997, the HDTV format using 1080 line interlaced at 29.97 fps was adopted for digital broadcasting in Japan. NHK has long experience in HD production, both in the studio and at remote sites, including the Olympic Games, going back to 1984. A wide variety of programming has been produced. At this time (September 2005), all networks (NHK and five commercial broadcasters) have completed their HD infrastructure in the station. Basically, the station is implemented throughout with HDTV including facilities for studio production (including the news studio), editing, and monitoring. Archive SD material and incoming programs from outside that is in SD are upconverted to the HD format in the station. Facilities for remote news production (ENG camera, etc) in HD will be completed in 2006. An implementation program for remote production will be complete before 2008, including interconnection facilities (FPU, Telco line, satellite links, specified fiber, etc.) It is expected that Japanese local broadcasters will follow the networks and will complete their transition to HD before 2008. Currently 80% of all local broadcast programs are provided by the networks. D. HD Origination for DTV—Europe Interest in HDTV began in Europe in the early 1990s. When efforts to persuade North America to move to a worldwide common HDTV 80-Hz standard failed, a 1250/50 format was proposed for the 50-Hz world. An analog HDTV transmission format, HD-MAC, was developed for satellite and cable broadcasting of 1250/50 in the late 1990s. However, market research showed that HDTV receivers at that time would be too bulky and expensive for the European public, and it also seemed better to delay HDTV for a future with digital broadcasting and, hopefully, cheaper and more consumer-acceptable HDTV receivers. HD-MAC was not used, and instead energies were focused on developing digital broadcasting, initially for SDTV, with multi channel delivery via digital satellite, digital cable, and digital terrestrial, through the DVB project. The age of consumer-acceptable HDTV receivers, with flat panel displays, has clearly arrived, and consequently HDTV is now being planned and implemented. Europe now exhibits a limited availability of HDTV services by Euro1080 (HD-1), and some test broadcasts (Pro-7 Germany). There are announcements for HDTV from TF1 and Canal+ in France, and confirmation for HDTV services in 2005/2006 by Pay-TV operators such as Premiere in Germany and BSkyB in the United Kingdom. These services JONES et al.: DIGITAL TELEVISION STATION AND NETWORK IMPLEMENTATION

will be based on generation one and generation two HDTV systems [18]. Generation one is defined as 1080i/25 and 720p/50 with MPEG-2 compression in emission; generation two as 1080i/25 or, preferably, 720p/50 with advanced compression such as MPEG-4 Part 10/H.264 in emission. In accord with the increasing interest, a number of national HD-Forums have been established in Italy, the United Kingdom, Germany, France, Spain, Portugal, and the Nordic countries. The EBU established a European HDTV Forum umbrella group for these forums, with a focus on coordination with respect to interoperability questions. The EBU has recently specified in EBU Tech. 3299 [19] four baseband systems considered relevant for HDTV production in Europe. These are: • System 1 (S1) with 1280 horizontal samples and 720 active lines in progressive scan with a frame rate of 50 Hz, 16 : 9 aspect ratio; • System 2 (S2) with 1920 horizontal samples and 1080 active lines in interlaced scan with a frame rate of 25 Hz, 16 : 9 aspect ratio; • System 3 (S3) with 1920 horizontal samples and 1080 active lines in progressive scan and a frame rate of 25 Hz, 16 : 9 aspect ratio; • an evolutionary System 4 (S4) with 1920 horizontal samples and 1080 active lines in progressive scan at a frame rate of 50 Hz, 16 : 9 aspect ratio. Based on intensive research [20]–[23], the EBU agreed and published a recommendation for HDTV [24]. This states that a progressive HDTV video format is the preferred HDTV system for emission. Currently this should be accomplished by the 720p/50 system but it is envisaged that in the future a 1080p/50 system might be an appropriate option. In accord with this conclusion, new research and work is under way to investigate a progressive 1080p/50 HDTV production and emission chain. On the consumer equipment side, the European Information and Communications Technology Industry Association (EICTA) [25] has agreed on an industry specification for HDTV-capable displays, called “HD-Ready,” and a specification for the minimum requirements for HDTV receivers, including the interface to such displays. They include logos to be used for devices meeting the requirements. In summary, it is assumed that HDTV in Europe will be a natural evolution of television. Pay-TV operators will first introduce HDTV services to the market and large public broadcasters will follow during the years to the end of the decade. Important factors include the penetration and availability of HDTV displays, receiver devices and sufficient HDTV broadcasts so that the consumer feels attracted to the new experience. E. HD Origination for DTV—Australia The Australian DTV system requires SDTV transmission with some percentage of simulcast HDTV. However, the planned quotas for HD have not fully materialized and much of the HD output is in fact currently upconverted SD material produced to the ITU-R BT.601 standard. 27

VII. ASPECT RATIO AND IMAGE DISPLAY ISSUES During the DTV transition period, some programming may be produced with 4 : 3 aspect ratio and some with 16 : 9 aspect ratio, depending on its original format and source. In most countries, there is also an installed based of television receivers and monitors of both aspect ratios. Therefore, it is necessary for broadcasters and receiver manufacturers to ensure that images shot in one aspect ratio will be displayed satisfactorily when transmitted and/or displayed in a frame of a different aspect ratio for DTV and/or analog services. Techniques for aspect ratio conversion include “letterboxing” and “pillar boxing” where bars are added to the top and bottom or sides of the picture. They also include cropping the picture to change its shape, or the process of pan-and-scan where the desired portion of the frame is identified and extracted dynamically. New productions can enforce “shoot-and-protect” where critical picture elements are restricted to portions of the frame that will always be displayed. Implementing these arrangements in a production and distribution environment can be a challenge, The ATSC and DVB standards define parameters that may be included in the transmitted bitstream to indicate the active picture area, for the case when this does not fill the whole encoded frame. This active format description (AFD) and bar data (ATSC only) can be used by a receiving device to help optimize the displayed picture. VIII. SURROUND SOUND 5.1 channel surround sound transmission for both SD and HD programming is enabled by DTV. However, it is not mandatory in any of the transmission standards and stereo or mono audio meets regulatory requirements. The implementation of surround sound production has therefore varied somewhat with different program producers and broadcasters in different parts of the world. Surround sound programming requires production, network, and station facilities to have audio systems that can handle six discrete channels of audio (eight if a stereo mix is also distributed). In particular, audio mixers, monitoring, switching, and distribution systems, previously intended for stereo audio only, all have to be upgraded with provision for at least six or eight-channel operations. It should be noted that, for many years, surround sound with four channels has been distributed using a matrix system for encoding the center and surround channels onto a twochannel stereo signal. 5.1 audio for DTV, with six discrete audio channels, provides a much-improved surround audio experience for television viewers. Major networks and some independent producers in North America have made the investment in new equipment for 5.1 surround sound and a substantial amount of U.S. primetime HD programming is now produced and distributed with surround sound. Many, but not all, U.S. local network stations have installed equipment that allow this programming to be passed through and fed to the DTV transmitter with surround sound encoding for the DTV standard. Other stations still broadcast in stereo only. 28

Similarly in Europe, Japan, and elsewhere, there is an increasing amount of HD or SD programming produced with 5.1 channel surround sound audio. However, broadcast DTV services still generally include some programs with matrix surround sound and some with stereo audio. A. Multichannel Audio Recording and Distribution The need to distribute and record surround sound signals with six channels of audio initially created a problem because some VTRs have only two or four audio channels, and legacy network distribution systems typically carry one stereo audio signal (often with a second for an alternative language). One early solution for recording was to use a separate multitrack audio recorder synchronized to the video recorder using timecode [26]; this is known as double-system sound. This arrangement enables all needed sound channels to be recorded, with the disadvantage of a considerable increase in system complexity. Some recent HDTV VTRs now have multiple audio channels on board for surround sound recording. One early distribution solution was simply to increase the number of discrete audio channels in the studio and network distribution, with consequent increase in complexity and cost. Another solution sometimes used for network distribution is to use emission compression encoding equipment, such as Dolby AC-3. However, emission compression modes are rather fragile and begin to degrade audio quality after two or three encode/re-encode cycles. Some U.S. networks (e.g., ABC) use AC-3 running at a higher bit rate (e.g., 640 kb/s) than is used for emission. The increased data rate reduces the quality degradation that otherwise can occur when such signals are decoded back to base band at the local station and reencoded for emission B. Embedded Audio It is possible to embed digital audio signals in the ancillary data space available in both SD and HD serial video bitstreams using the SMPTE 272M standard [27]. Up to 16 audio channels (eight 2-channel pairs) can be carried in this way. The technique has many advantages for distribution within a plant, but creates additional complexity whenever the video and audio signals need to be processed, mixed, or switched separately, and is not applicable when the video has to be compressed for distribution over network links. C. Dolby E An alternative solution for both recording and distribution uses a coding system called Dolby E [28], developed by Dolby Laboratories. This can be used to lightly compress up to eight audio channels into the bandwidth occupied by a single two-channel AES/EBU digital audio signal (3 Mb/s). The signal can be recorded onto a VTR or video server and can be distributed over an AES/EBU audio channel from a production facility to the network, and on to the broadcast station. There it can be converted back to individual audio channels for mixing and switching in master control, and finally reencoded as AC-3 audio for transmission. The Dolby E signal can withstand up to 20-plus encode/decode cycles. PROCEEDINGS OF THE IEEE, VOL. 94, NO. 1, JANUARY 2006

Although an efficient method for recording and distribution, it does, however, introduce some extra system complexity for monitoring and postproduction. IX. IT-BASED EQUIPMENT AND SYSTEMS The architecture for the first digital broadcast studio systems was very similar to traditional analog systems. Audio and video equipment was largely based on custom hardware; with point-to-point dedicated interconnect cabling using serial digital bitstreams and crosspoint matrix routing. Recent developments have enabled most video and audio recording and processing tasks to be carried out using computer-based equipment. Related to this is the ability to store video and audio material as computer files, and to transfer them over standard computer networks. The introduction of such systems has taken place in parallel with, but separate from, the change to digital transmission. Some of the most significant new equipment types are hard disk based video and audio servers and nonlinear editing and postproduction equipment. These, and associated systems based on information technology (IT), cover very many parts of the broadcast chain from acquisition to on-air playout and have triggered a revolution in system architecture, workflow and operational practices. This major change in broadcast technologies was analyzed and, to quite a large extent, enabled by a joint EBU/SMPTE Task Force for Harmonized Standards for the Exchange of Program Material as Bitstreams. The report [29] of that group was the foundation for many of the advanced digital television standards and techniques in use today. A. File Transfer In traditional analog and digital systems, audio and video program material exists as linear signals that are produced and distributed in real time. Such signals are usually recorded to, and played back from, videotape machines in real time. Now that digital audio and video material is increasingly stored on hard drives in the form of computer files, such material can also be accessed at random and can be transferred as files over the network to another local or remote location. This may allow faster than real time distribution over high-speed networks (depending on the video format and the network capability) or, when necessary, transfer at lower speeds over restricted bandwidth links. These techniques are now widely used within and between production and postproduction facilities. They will be increasingly used for program distribution to network release centers or to on-air transmission points in advance of transmission time. B. MXF, AAF, GXF, and Metadata The transition to digital video and audio systems has resulted in numerous different video and audio compression and storage formats, some derived from videotape systems, and now used for server-based file storage and distribution. A recent suite of SMPTE standards known as Material eXchange Format (MXF) [30] enables disparate systems from different vendors to have a common standard file wrapper JONES et al.: DIGITAL TELEVISION STATION AND NETWORK IMPLEMENTATION

that allows seamless interchange of material. This applies particularly for distribution of finished programs but also for sequences of clips or program segments. A related set of standards, published by the Advanced Authoring Forum (AAF) is intended for interchange of material in the postproduction process where complex edits and effects are required. The General Exchange Format (GFX) is a third interchange format designed primarily for on-air applications. Television programming has always had metadata (description of the program essence) associated with it, if only a written label describing the title, video format, and length. The MXF, AAF, and GXF standards make extensive use of metadata for the production and technical characteristics of the material, as a fundamental part of the electronic file system and interface and this makes these technologies very powerful tools for media production and management. These interchange formats all have a foundation in the metadata standards published by SMPTE [31]. X. MASTER CONTROL AND EMISSION ENCODING A. Master Control for DTV Master control areas contain the equipment for control and monitoring of the on-air program, whether from a network release center, a station group centralcasting location, or a local station. In a DTV facility, the master control output is usually fed to compression equipment for distribution or emission. The system architecture for these functions varies considerably for different networks and stations. It also depends on the number of program outputs and whether services are in SD or HD, so what follows should be considered as an example only. Fig. 2 illustrates in very simple terms how a master control and emission encoding system may be configured for a local U.S. DTV broadcast station. It shows various local, remote, and network sources feeding (a term often used in broadcasting meaning “sending signals to”) a baseband master control switcher, which may be either SD or HD. The switcher is used to select the program source to be transmitted. The video and audio outputs from this switcher then feed the ATSC video and audio encoders and multiplexer (mux), and the mux output goes to the transmitter. Audio processing before coding for emission is not necessarily required, but is implemented by some broadcasters. Arrangements for a network master control may be similar, but the master control switcher output in that case will feed the encoders for the network or group distribution, and then go to a satellite uplink rather than a terrestrial transmitter. Where a station operates two or more separate program services, e.g., for multiple SD programs or one SD and one HD, there may be a separate master control area for each service. Alternatively, one master control position may be configured to control two or more switchers from the same control panel, and this arrangement will usually apply when a station operates several DTV program services for multicasting. Control of the master control switcher by an automation system is a critical part of station operations, especially for multiprogram services. 29

Fig. 2.

Master control and encoding.

B. MPEG Encoding and Multiplex Systems As shown in Fig. 2, the video compression is carried out in a video encoder, and the audio compression is carried out in an audio encoder, as described in the related papers on those topics. The vast majority of terrestrial DTV services worldwide currently use MPEG-2 video encoding, but in some circumstances advanced codecs such as MPEG-4 Part 10/H.264 and Windows Media 9 (or the SMPTE standardized version, VC-1) are beginning to be introduced. H.264 will probably be used exclusively for HDTV broadcasting in Europe. Each encoder produces an elementary data stream, which is combined into a single program stream in a multiplexer. The output of this multiplexer is then combined, in a second multiplexer, with other data packets to produce the final transport stream that is fed to the transmitter. Depending on the equipment design, the multiplexers may be separate devices or may be in a single box integrated with the video encoder. The audio encoder may be combined in the integrated unit (common for two-channel encoders) or may be a stand-alone device (usual for 5.1 channel encoders). It is most common for a station to locate the encoding and multiplexing equipment at the studio center and send the single transport stream over the studio-transmitter-link (STL) to the transmitter (see Section XII), but other arrangements are possible.

RF channel, forward error correction (FEC), and modulation scheme in use, as defined in the various terrestrial DTV standards.

C. Bit Rates Bit rates typically used for SD video for emission encoding with MPEG-2 compression may range from about 2 to 8 Mb/s, depending on the program content and desired quality level. Bit rates for HD video may range from about 12 to 18 Mb/s depending on the video format, program content, and desired quality level. These bit rates can be greatly reduced by using one of the new advanced codecs. Bit rates used for the audio content are typically in the range 192–448 kb/s per service, depending on whether the program is stereo or 5.1 surround. A small amount of data bandwidth is required for system information and metadata. In addition, some bandwidth may be allocated for data broadcasting services. In all cases, the total bit rate allocated cannot exceed the capacity of the broadcast transmission channel—between about 14 and 23 Mb/s, depending on the

Some networks distribute a multiplex of programs and local stations or regional centers may need to modify the mix with local programming. In this case, the incoming bitstream is fed to the multiplexer where it can be reconfigured by adding or dropping program elements.

30

D. Multicasting Operations Fig. 2 illustrates one possible basic station output arrangement for a single DTV program service. However, when stations multicast two or more program services in a single output bitstream on their DTV channel, there will be multiple video and audio feeds (the broadcasting term for the signal being sent) from master control, to multiple video and audio encoders. These other program services are all combined in the final mux, as indicated in the figure. In general, increasing the number of programs in one multiplex results in lower picture quality and more visible artifacts, but this depends very much on the format of the video carried and the program content. When multiple program are carried, statistical multiplexing may be used, which can improve the quality of each service considerably. Statistical multiplexing allocates available bandwidth to each program on an as-required basis, rather than using a fixed bit rate for each one. The random (statistical) nature of video means that peak bandwidth requirements for multiple programs rarely coincide. E. Remultiplexing

F. Closed Captioning Equipment It is outside the scope of this paper to cover how closed captions (subtitles intended to allow deaf people to follow the program audio) are generated and carried, but most programs will be received from the network or other supplier with caption information carried along with the video signal but not visible. Live local programs, such as news, are typically captioned at the station, although the person producing the captions can, in fact, be located off-site and provide the service remotely. PROCEEDINGS OF THE IEEE, VOL. 94, NO. 1, JANUARY 2006

The caption information may be carried in the studio center embedded as data in the digital version of the video vertical blanking interval or vertical ancillary data space (VANC) [32]. Those elements are not encoded for transmission. Therefore, the caption information has to be extracted and placed in the appropriate place in the DTV encoded video bitstream for transmission. This process may take place entirely inside the encoder, or may require an external device, called a caption server, as shown in Fig. 2. The caption server is also needed if DTV captions are being generated locally. G. PSIP or EPG Generator For an ATSC station, the PSIP tables are produced in the PSIP generator computer. The output of this is a data stream feeding the ATSC multiplexer, as shown in Fig. 2. The PSIP generator may be connected through a network with other computers, such as the traffic system, and automation system, which produce and manage the program schedule and associated information needed to generate the PSIP tables. H. Data Broadcasting Equipment If present, a datacasting service may be combined with the television programs in the multiplexer at the output of the studios. This requires a data connection to the multiplexer from the data server as shown in Fig. 2. The data server stores and communicates the content of the data service. It is usually connected through a network with other computers, either at the station or elsewhere, which produce and manage the data information. I. Alternative Master Control Arrangements The previous sections on master control and encoding assume that the DTV program feed from the network is received and converted to baseband video and audio before being distributed at the station, switched to air through a master control switcher, and then encoded for transmission. Not all networks and stations follow this model, however. Another possible arrangement, used by two U.S. networks (Fox and PBS) is to carry out ATSC encoding at the network release center and distribute this emission-level (19.39 Mb/s) compressed signal to the affiliate stations. Local programming is encoded for transmission at the affiliate station and ground terminal equipment is provided there that is capable of MPEG splicing to enable the transitions from local programming and interstitials to network programming as required. Such compressed bitstream splicing allows for the seamless integration of MPEG program streams, such that the decoder in the viewer’s home (as well as the viewer) is unaware that the network and local material was not encoded by one single encoder. To be useful in today’s television environment, other functionality needs to be implemented that allows local station identification and branding, including graphic overlays (both static and dynamic), video sizing, as well as audio voiceover. These must be implemented without the complexity and quality degradation of decoding the whole signal back to baseband and reencoding for transJONES et al.: DIGITAL TELEVISION STATION AND NETWORK IMPLEMENTATION

mission. Some of these features have been implemented, while others have been demonstrated and will be deployed in the near future. Other arrangements are possible, and the decoding, switching, distribution, and emission encoding arrangements at different stations vary considerably. XI. NETWORK DISTRIBUTION ARCHITECTURE This section is written largely from a U.S. perspective, where several different strategies have been used for distribution of DTV programming from networks to local broadcast stations, largely driven by the need for HD program distribution. The basic concepts and choices are similar for different countries, but there are many possible configurations and variations, particularly where distribution of several bitstreams with many programs is required, e.g., as for the U.K. Freeview service. Factors to be considered include the number and location of the network release facilities, local stations, and transmission facilities; the multiplex configuration; and sources of programming. It should be noted that architectures and implementation arrangements for network distribution are still being developed and improved. A. Background HD service from national networks has evolved from the launch of DTV service in the United States until today. Each network has had unique challenges and applied unique solutions. In general, the solutions have been digital and have covered the country using geosynchronous communications satellites. In the early years, when HDTV programming was sparse, most networks used ad hoc distribution that used spare transmission capacities with limited capabilities. HD playout facilities were islands within the normal SD infrastructure (which in some cases was still analog). As time progressed and the amount of HD programming increased, these ad hoc distribution systems were replaced by homogeneous digital satellite systems that transport both HD and SD programs. B. Link Platforms Aside from local links from the network satellite receive point to an affiliate station, which may use fiber or microwave transmissions, all national network HD programming in the United States is delivered by satellite, with its easy implementation of point to multipoint distribution. Digital satellite systems are in use that deliver data rates from 30 Mb/s (using QPSK) to greater than 70 Mb/s (using 8 PSK) depending on the ground terminal G/T as well as group delay and LNB phase noise performance. Satellite systems are also used for many contribution circuits but fiber-optic cables are a viable alternative, since such links are often point to point or with a limited number of destination drops. C. Compression Levels There are many different compression and coding methods in use by the networks to deliver digital HD programs. For 31

live contribution feeds to network centers, MPEG-2 encoding is typically used with 4 : 2 : 2 sampling and data rates typically in the range 35–45 Mb/s. For distribution from network centers to local affiliate stations, most networks use MPEG-2, long GOP, 4 : 2 : 0 profile, with data rates typically in the range 25–40 Mb/s, although one network (CBS) is moving to 4 : 2 : 2 sampling at 43 Mb/s for network distribution. At the affiliate station’s satellite receive site, the receive equipment decodes the compressed satellite HD program and supplies an uncompressed HD video signal. This signal is either routed to the stations HD master control (see Fig. 2) directly, or, if the receive site is not collocated, compressed again and sent via terrestrial microwave or via fiber (uncompressed) to the station master control. Compressed data rates that fall between the fully compressed rate required for emission (about 20 Mb/s) and uncompressed HD video (about 1.5 Gb/s) are referred to as mezzanine levels. Using these for contribution and distribution links allows compression to be less aggressive. This enables multiple cycles of encode/decode and video processing between encode passes without degrading the final compressed program for emission. Rates around 45 Mb/s, as mentioned above, may be considered low-level mezzanine. The use of higher rates (100–400 Mb/s) would provide higher quality margins, but there are no available national satellite distribution systems that can deliver such high bit rates. D. Distribution Compressed for Emission Two national U.S. networks, Fox and PBS, have implemented their national HD program distribution with a model using emission level encoding at the network release center. In order to avoid unacceptable video (and audio) quality degradation by decoding such signals to baseband at the station and reencoding for emission, the use of MPEG splicing is required, as discussed earlier in Section X-I. This approach saves distribution bandwidth (both over satellite and locally) as well as improving video quality by avoiding a decoding/encoding step. The network emission coding can be either constant bit rate (CBR) or variable bit rate (VBR) with the advantage of minimizing bandwidth used both over the satellite as well as providing more bandwidth for secondary programs (such as weather or news channels) that the local affiliate station may broadcast on its multiplex. E. Audio Distribution Some distribution systems use MPEG-2 Layer 2 audio coding for two-channel audio distribution. As mentioned in Section VIII, distribution for surround sound audio may use high-rate AC-3 encoders or the more robust Dolby E. Those networks that use emission level coding and MPEG splicing can utilize the bit efficiency of emission-rate AC-3 without the degradation of reencoding, thus saving more bandwidth. XII. TRANSMISSION SYSTEMS This section is written largely from a U.S. perspective, with reference to the ATSC 8-VSB transmission standard. However, most of the basic concepts and choices are 32

similar for different countries, although various parameters—particularly frequency bands and power levels for transmitters—will vary significantly. Factors to be considered include frequency allocations and availability of spectrum, number and location of stations and transmission facilities, and population areas to be covered A. Background The output of the television studio, whether analog or digital, feeds the STL, which in turn feeds the broadcast transmitter. Over the air signals are used in several ways: • direct reception by tens of millions of viewers provided by tall television towers; • reception at tens of thousands of cable headends for redistribution to subscribers; • extended reception by thousands of television translators serving small communities that are isolated from the coverage of the large broadcast transmitters. This combination provides television signal coverage to over 99% of the U.S. population. A complex and expensive broadcast infrastructure has grown over the past five decades that in the United States utilizes much of the VHF and UHF (plus some SHF) spectrum and requires extensive reserve capacity and backup transmission and power supply facilities. The change to digital transmission in effect requires a duplication of these facilities during the transition and a hardening of those new facilities to continue to maintain a reliable broadcast service. The DTV signal is a complex work of engineering that incorporates: • MPEG-2 compressed digital video and audio; • a transport stream containing multiple video, audio, data services, system information, and PSIP signaling; • 8-VSB modulation techniques, combined with powerful error reduction coding that provides a signal that needs only about 15-dB signal-to-noise ratio for perfect reception (even less if the new standard for Enhanced VSB transmission is used), which is also shaped to sharply attenuate the signal to fit within a 6-MHz channel. The new digital facility needed to place this signal on air requires a new STL, digital transmitter, bandpass filter to meet Federal Communications Commission (FCC) mask requirements, transmission line, antenna, and test and monitoring equipment. The addition of the components to typically already crowded transmitter buildings often requires new building construction, and the DTV antenna may require a new tower or strengthening of an existing structure. B. Coverage, Frequency, and Power Based on assessments of various planning factors and the characteristics of the 8-VSB DTV signal, the FCC assigned a new DTV channel and power for each analog licensee that, for most stations, was designed to replicate their current analog coverage area. Many VHF analog stations were assigned new channels in the UHF band. The shift from VHF to UHF for digital PROCEEDINGS OF THE IEEE, VOL. 94, NO. 1, JANUARY 2006

resulted in much higher power levels to achieve the same coverage. For example, VHF stations operating at 100 kW (maximum for channels 2–6) or 316 kW (maximum for channels 7–13) peak effective radiated power (ERP), the move to UHF generally required nearly 1 MW average ERP to achieve similar coverage for the DTV signal. While higher gain antennas for UHF help reduce transmitter power, the difference was so great that most stations with a new UHF DTV assignment decided to begin operation with lower than authorized power to reduce both purchase and operating costs. On the other hand, UHF analog stations with a new UHF digital assignment found that they needed lower radiated power to achieve comparable coverage. For example, a maximum power analog UHF station at 5 MW ERP required less than 500 kW ERP for comparable DTV coverage on UHF. This fact is due to the unique characteristics of the digital and analog signals, propagation differences between UHF and VHF, and revised planning factors for DTV reception. Note that analog power is measured to accommodate the peak synchronizing signal while digital signals are measured as average power because of their consistent and wide-band spectral characteristic. Transmitted 8-VSB signals have a peak-to-average ratio of about 7 dB. Field tests backed up by laboratory analysis demonstrated that digital signals can be received perfectly (albeit not always reliably) at 12 dB below the point at which an analog signal becomes less than passable quality (Grade B). C. DTV Transmitters The change to digital forms of modulation from the traditional analog technology required a reassessment of transmission system parameters including power levels, linearity, frequency response, out-of-band emissions, and reliability. Meeting the higher standards required of digital facilities meant new research and development efforts on the part of equipment manufacturers who successfully developed linear amplifiers and high-level bandpass filters meeting the stringent out-of-channel emission limits. These, in turn, allow adjacent channel operation either with analog or digital television broadcast signals. D. Antenna Systems and Towers A new DTV antenna, often on a different band, is needed in addition to the existing antenna. Most stations have found that their existing tower will not support the new antenna without modification and strengthening, often at a cost in some cases nearly equaling that of a new tower. In other cases, there was no alternative but to plan for a new and expensive tower. However, a new tower frequently takes years for approval and construction. Some stations are still waiting for local approval for a new tower. E. Shared Facilities In some markets, broadcasters have formed collaborative organizations to establish a common transmitter and antenna facility. In dozens of these cases, towers, buildings, and antennas are shared by two or more stations resulting in cost savings for equipment purchase and maintenance. Another JONES et al.: DIGITAL TELEVISION STATION AND NETWORK IMPLEMENTATION

important benefit is that it provides a single common direction for viewers to aim their receive antennas. The classic example is DTV Utah in Salt Lake City, where eight stations collaborated to build a single building and two towers instead of multiple individual towers and buildings at different locations. F. Studio–Transmitter Links for DTV Most stations have their transmitter located well away from the studio, which may be in a downtown area convenient for businesses and personnel connected with the broadcast operations. To interconnect the two locations, an STL is employed which may consist of a one or two hop microwave relay in the 2-, 7-, and 13-GHz UHF and SHF bands or a fiber-optic cable provided by a local common carrier. The microwave system is owned by the station and requires only a line-of-sight path and coordination with other users of the bands to insure interference free operation. The fiber-optic circuit is leased from a common carrier but because the path is on or in the ground, it is susceptible to occasional outages due to rerouting and inadvertent damage from construction crews digging up conduits and trucks toppling telephone poles. Converting a microwave STL for DTV operation means replacement of the transmitters and receivers (usually main and backup). In some cases, dual-use microwave systems are employed to carry the program signals for both the analog and digital transmitters. Depending upon the lease terms, a fiber-optic STL can be converted or modified to carry both analog and digital broadcast signals to the transmitter site. G. DTV Translators Just as broadcasters need a new transmitter to accommodate digital transmissions from their main station, the comparatively low power translators serving isolated and smaller communities also need a new facility to provide the signal for the new digital service. Translators receive their input signal over the air from full power transmitters that are often over 100 mi distant. With so many new DTV stations using previously unused channels, interference to translator input as well as output can occur, making the implementation of DTV translators more difficult than for analog facilities. Because translators operate at comparatively low power levels, the FCC may, under certain conditions, permit a relaxed out-of-channel emission mask filter to be employed, which will substantially reduce installation cost and complexity. Translators traditionally convert an incoming analog channel to a different output channel through a heterodyne process, rather than demodulating the signal to video and audio baseband and remodulating to obtain the output channel. While this heterodyne process eliminates degradation due to demodulation and remodulation of the signal, the signal can still be degraded by noise and interference in the transmission path. When using a string of translators (daisy chain) the reduction in quality can be serious. DTV translators overcome this deficiency by demodulating the signal to the transport data stream, which can be recovered 33

with virtually no errors and remodulating the transport data stream for transmission on the new frequency. As a result, the quality remains the same as the original transmission. Multiple hops of translators using this technique have no adverse effect on the end users video and audio (either cable TV headends or direct to the home), including HD video and 5.1 surround sound. This fact alone provides an incentive to implement DTV translators earlier in the transition process. The cost of a DTV translator can be lower in some respects but higher in others than an equivalent analog translator. Lower DTV output power provides the same coverage as analog but the DTV remodulation process is more expensive than simple heterodyne conversion or analog remodulation. It is expected that DTV remodulation costs will drop substantially in time and with volume purchases. Today there are 4400 analog translators serving about 7% of the total television audience.

H. Costs DTV transmitter prices range from $50 000 for a low-power DTV to nearly $1 000 000 for high-power UHF facilities and about $500 000 for VHF DTV facilities. In addition to the cost of the transmitters, a range of ancillary equipment costing from $50 000 to $250 000 is necessary to provide long-range support and reliability. This would be in addition to the cost of the antenna, tower changes, and installation. Recognizing that there are more than 1700 full power stations and over 2500 low-power stations (including stations licensed by the FCC as Class A, which allows higher power levels in some circumstances) in the United States, it is clear that the cost to broadcasters in carrying out the transition to digital transmission is tremendous. It represents a substantial investment that many broadcasters believe will not result in payback for many years to come.

XIII. TEST EQUIPMENT Since digital television signals are so fundamentally different from analog, DTV facilities require various items of specialized test and monitoring equipment that are fundamentally different from those for analog systems. Without the right equipment, the performance verification of equipment and systems and, in particular, the diagnosis and rectification of errors in digital video, audio, compressed bitstreams, and digitally modulated RF signals will be difficult or impossible. Such equipment types include: • baseband digital video measurement set; • baseband digital audio measurement set; • transport stream analyzer; • 8-VSB and/or COFDM analyzer; • vector signal analyzer. In addition, since so many digital television systems are based on computer and IT systems, nearly all the IT diagnostic and test tools are relevant. 34

XIV. PROGRAM SERVICE AND MULTIPLEX ARRANGEMENTS A. DTV Services in ATSC Countries In the United States, most full-power broadcast television stations are associated with one of the national networks (ABC, CBS, Fox, NBC, PBS, PAX, the WB, and UPN). Typically, each station has one main VHF or UHF analog broadcast channel, although stations in some areas also have translators and repeaters. Many state broadcasters associated with PBS operate multiple transmitters to cover a whole state. During the DTV transition, each station has been allocated an additional VHF or UHF broadcast channel for digital services. As of September 2005, 1508 DTV stations were on air in 211 U.S. markets, covering the vast majority of the population. It is now proposed that analog terrestrial broadcasting in the United States will terminate on 31 December 2008. Most stations broadcast some programming from their network feed, some produced locally (primarily news) and some syndicated programs from other sources. The FCC requirement is for the program content on the analog channel to be largely replicated in digital, but stations may also carry additional digital programs not provided in analog. There is no mandatory requirement for any DTV programs to be in HD. The networks have migrated much of their primetime program production to HD and many stations carry those programs in HD. At this time, most local stations still produce their own local news and programming in standard definition. Except for USDTV,1 all terrestrial DTV broadcasting in the United States is currently unencrypted and free of charge. Multiplex Arrangements: The ATSC broadcast channel can carry a bit rate of 19.39 Mb/s, which may be allocated to virtual channels in the multiplex with a variety of combinations of HD and SD programs. Stations (or station groups, when owned by a group) make their own decisions on services to be carried, with typical arrangements as follows: • one HD program only; • one HD and one SD program; • four SD programs. Other combinations are possible. An additional low bit rate SD virtual channel is often carried for services such as weather radar, and some stations may allocate some proportion of their bandwidth to data broadcast services. It is common for PBS member stations to vary their program multiplex at different times, perhaps carrying multiple SD programs during the day and switching to one HD or one HD and one SD for the evening. Service arrangements in other ATSC countries are generally similar, although the rollout is less advanced. 1The U.S. Digital Television (USDTV) organization has established cooperative ventures with stations in some areas of the country to offer a subscription service. Each station maintains its own free-to-air HD or SD programming, but allocates a portion of its DTV multiplex bandwidth to additional encrypted SD programs. The additional channels use the Windows Media 9 advanced video codec and comprise a total of 12 programs otherwise available only on cable or satellite.

PROCEEDINGS OF THE IEEE, VOL. 94, NO. 1, JANUARY 2006

B. DTV Services in ISDB Areas The basic concept of ISDB for digital services is that it provides interoperable broadcasting system and content for various physical media such as terrestrial and satellite channels Given the band-segmented transmission for the terrestrial channel and slot assignments for the satellite channel, one can choose the most efficient transmission parameter set for one segment/slot, or the most robust one for another segment/slot in the same channel. The common feature is the adoption of HDTV for terrestrial and satellite. ISDB-S digital satellite service started in December 2000, providing seven HDTV, three SDTV, and a variety of sound and multimedia data broadcasting services. In 2007, following the termination of analog HDTV satellite service, digital HDTV satellite broadcasting will be expanded. DTV terrestrial broadcasting using ISDB-T (simulcast with conventional analog channels) was started 1 December 2003 in the Tokyo, Osaka, and Nagoya, Japan, areas. Other local broadcasters have now started digital transmissions, and it is expected that terrestrial broadcasting will cover 60% of households by the end of 2005, and major cities all over Japan by the end of 2006. It is planned that analog terrestrial broadcasting will be terminated by 2011. The on-air terrestrial DTV channels provide a full service for almost 24 hours a day, and government guidelines require that more than 50% of all programs should be pure HD and not upconverted. The ISDB-T specifications allow flexible allocation of bandwidth in a channel for different services. It enables a maximum of three independent segments for targeting home reception, mobile reception, and portable reception. Currently each broadcaster provides one HDTV or multiple SDTV services for home or mobile reception. In 2006, some broadcasters will start to use one segment for a new service for portable reception by handheld terminals and mobile phones. It is expected to create new interactive broadcasting services, cooperating with other communication media. More information on the service arrangements for ISDB is provided in the related ISDB papers in this special issue. C. DTV Services in DVB Areas The DVB standards have been implemented in large areas around the world. Digital satellite and digital cable services using DVB-S and DVB-C have been available with SDTV programming for a number of years. European commercial as well as public service broadcasters are well advanced with digital SDTV services. The compression system used in emission has been MPEG-2 MP@ML with bit-rates between 2.5 Mb/s and (more recently) up to 8 Mb/s for SDTV. Bit rates are chosen depending on economic considerations, the degree to which high-quality flat panel displays are in the public’s hands, acceptable quality, and the availability of statistical multiplexing. Recently, DVB-S2 has been developed for future digital satellite transport. DVB-S2 will provide an efficiency improvement of 30% compared to DVB-S [33]. JONES et al.: DIGITAL TELEVISION STATION AND NETWORK IMPLEMENTATION

Digital terrestrial DVB-T is under continuous rollout in Europe. However there are large differences in the time schedule from country to country, with the most developed digital broadcasting being in the United Kingdom, where DVB-T, DVB-C, and DVB-S are now well established. For DVB-T, there are about 30 television and 20 radio channels in six Freeview multiplexes. Analog television broadcasting in the United Kingdom is slated to end by 2010. Several hundred channels are available via DVB-S. An ambitious DVB-T rollout is in progress in Germany, region by region, with the analog service being switched off within months of the start of DVB-T. Digital DVB-T broadcasting began in France in 2005. Other countries have also begun digital terrestrial and satellite broadcasting. Regular HDTV services in Europe will first be broadcast via satellite utilizing DVB-S2 and H.264 compression. First announcements from cable operators have also been made to introduce HDTV channels. HDTV via DVB-T is under discussion but further investigations and trials are required. Countries that have not yet begun DVB-T SDTV may decide to go straight to HDTV. Countries that have already started with SDTV may decide to use the “digital dividend,” following analog switchoff, for HDTV. However, a competitor for spectrum—broadcasting to handheld receivers via a ruggedized broadcasting system, DVB-H—is also being considered. More information on the service arrangements for DVB is provided in the related ISDB papers. ACKNOWLEDGMENT Fig. 2 and portions of the text of Section X are extracted and adapted from A Broadcast Engineering Tutorial for NonEngineers by Graham Jones, published 2005 by Focal Press. The authors would like to thank H. Katoh, NHK, and N. Katsura, previously with NTV, for providing information in Sections VI-C and XIV-B about HD origination for DTV in Japan and DTV services in ISDB areas. REFERENCES [1] D. H. Pritchard and J. J. Gibson, “Television transmission standards,” in Standard Handbook of Broadcast Engineering, J. C. Whitaker, Ed. New York: McGraw-Hill, 2005, pp. 3.9–3.33. [2] M. Robin and M. Poulin, Digital Television Fundamentals. New York: McGraw Hill, 1998, pp. 131–211, 235–282. [3] J. Watkinson, The Art of Digital Video. Oxford, U.K.: Focal Press, 2000. [4] “Studio encoding parameters of digital television for standard 4 : 3 and wide-screen 16 : 9 aspect ratios,” Recommendation ITU-R BT.601-5, 1995. [5] AES recommended practice for digital audio engineering—Serial transmission format for two-channel linearly represented digital audio data, AES Standard AES3-2003, Audio Engineering Society. [6] International Telecommunications Union, “Parameter values for the HDTV standards for production and international program exchange,” ITU Recommendation BT.709-5, 2002. [7] ——“1280 720, 16 9 progressively-captured image format for production and international programme exchange in the 60 Hz environment,” ITU Recommendation BT.1543, 2001. [8] 10-bit 4 : 2 : 2 component and 4fsc composite digital signals—Serial digital interface, SMPTE Standard 259M-1997, Society of Motion Picture and Television Engineers.

2

2

35

[9] International Telecommunication Union, “Interfaces for digital component video signals in 525-line and 625-line television systems operating at the 4 : 2 : 2 level of Recommendation ITU-R BT.601 (Part A),” ITU Recommendation BT.656-4, 1998. [10] Bit-serial digital interface for high-definition television systems, SMPTE Standard 292M-1998, Society of Motion Picture and Television Engineers. [11] ——“Standards,” [Online]. Available: http://www.smpte.org/ smpte_store/standards/ [12] International Telecommunications Union, International Telecommunications Union home page [Online]. Available: http://www.itu. int/home/index.html [13] Audio Engineering Society [Online]. Available: http://www.aes. org/publications/standards/list.cfm [14] Consumer Electronics Association [Online]. Available: http://www.ce.org/standards/StandardsCatalog.aspx [15] IEEE [Online]. Available: http://www.ieee.org [16] Society of Cable Telecommunications Engineers [Online]. Available: http://www.scte.org/content/index.cfm?pID=59 [17] B. Fox, “The digital dawn in Europe [HDTV],” IEEE Spectr., vol. 32, no. 4, pp. 50–53, Apr. 1995. [18] H. Hoffmann, D. T. Itagaki, and D. Wood, “HDTV revisited—User requirements and opportunities,” presented at the 146th SMPTE Tech. Conf, Pasadena, CA, 2004. [19] European Broadcasting Union, “High definition image formats for television production,” EBU Tech. 3299-2004. [20] ——“The potential impact of flat panel displays on broadcast delivery of television,” EBU Tech. Inf. I34-2002. [21] ——“Further considerations on the impact of flat panel home displays on the broadcasting chain— First edition,” EBU Tech. Inf. I35-2003. [22] ——“EBU guidelines for the RRC-06,” EBU Tech. Inf. I37-2005. [23] ——“Maximizing the quality of conventional quality broadcasting in the flat panel environment,” EBU Tech. Inf. I39-2004. [24] ——“EBU statement on HDTV standards,” EBU Tech. Rec. R1122004. [25] —— [Online]. Available: http://www.eicta.org/ [26] Audio and film—Time and control code, SMPTE Standard 12M1999, Society of Motion Picture and Television Engineers. [27] Formatting AES/EBU audio and auxiliary data into digital video ancillary data space, SMPTE Standard 272M-2004, Society of Motion Picture and Television Engineers. [28] Dolby Laboratories, “Dolby E multichannel audio coding,” [Online]. Available: http://www.dolby.com/assets/pdf/tech_library/49_Dolby_E_Innovation.pdf [29] Society of Motion Picture and Television Engineers, “Task force for harmonized standards for the exchange of program material as bitstreams, final report: Analysis and results, 1998,” [Online]. Available: http://www.smpte.org/engineering_committees/pdf/tfrpt2w6.pdf [30] B. Devlin, “MXF is ready for you, are you ready for it?,” in Proc. NAB Broadcast Engineering Conf. 2005, pp. 438–443. [31] O. Morgan, “Metadata system architecture,” in Proc. IBC Conf. 2002 [Online]. Available: http://www.broadcastpapers.com/IBC2002/ibc2002.htm [32] Vertical ancillary data mapping for bit-serial interface, SMPTE Standard 3342M-2000, Society of Motion Picture and Television Engineers. [33] D. Breynaert and M. d’Oreye de Lantremange, “Analysis of the bandwidth efficiency of DVB-S2 in a typical data distribution network,” presented at the CCBN 2005, Beijing, China. Graham A. Jones was born in Stoke-on-Trent, U.K., in 1944. He received the B.Sc. degree in physics from Nottingham University, U.K., in 1964. He started his career with the BBC and for 18 years was a partner of International Broadcasting Consultants. Previously with Harris Corporation, he was Engineering Director for the Harris/PBS DTV Express road show. He is currently Director of Communications Engineering with the National Association of Broadcasters, Washington,

36

DC, working on advanced television issues, standards, and education. He is author of A Broadcast Engineering Tutorial for Non-Engineers (Elsevier, 2005), and editor for the forthcoming 10th NAB Engineering Handbook. Mr. Jones is a Fellow and Governor of the Society of Motion Picture and Television Engineers. He is a Chartered Engineer and a member of the Institution of Electrical Engineers, the Society of Broadcast Engineers, and the Royal Television Society. He was chair of the ATSC TSG/S1 specialist group on PSIP Metadata Communications and chairs the SMPTE S22 working groups on lip sync and image formatting. In 2004 he received the ATSC Bernard J. Lechner Outstanding Contributor Award.

James M. DeFilippis (Member, IEEE) was born in Morristown, NJ, on January 27, 1958. He received the B.S. and M.S. degrees in electrical engineering from the School of Engineering, Columbia University, New York, in 1980 and 1990, repsectively. He has worked in radio and television broadcasting for 25 years, including the ABC radio network, the ABC television network, the Advanced Television Test Center, and the Atlanta Olympic Broadcast Organization. Currently he is Senior Vice President of Television Engineering for the Fox Technology Group, Los Angeles, CA. His primary research focus is on new video and audio compression systems. Mr. DeFilippis is a member of SMPTE and is involved in standards development at the International Telecommunications Union.

Hans Hoffmann (Member, IEEE) was born in Munich, Germany. He received the Engineering Diploma from the University of Applied Sciences, Munich, in 1992. He joined the Insitut fuer Rundfunktechnik (IRT) in 1993 as a Member of the Research Staff in new television production technologies. In 2000 he moved as Senior Engineer to the Technical Department, European Broadcasting Union, Geneva, Switzerland. For the last three years, he has been involved in the EBU technical activities on high definition in television production and emission. He has chaired the EBU project groups P/BRRTV and P/PITV, both involved in new technology innovations in television. Mr. Hoffmann is a Fellow of the SMPTE and a member of SID and FKT. He was chairman of the SMPTE technology committee N26 on Networks and File Management and in 2002 became SMPTE Engineering Director, Television and chair of the Television steering committee.

Edmund A. Williams (Senior Member, IEEE) was born in Cleveland, OH, in 1939. He received the B.S. degree in communications engineering from Franklin University, Columbus, OH, in 1961. His professional career has been in public broadcasting engineering positions since 1958, and he retired from the Public Broadcasting Service in 2004. He developed and conducted technical seminars on the 40-city nationwide tour of the Harris/PBS DTV Express. He is author of numerous technical publications and editor-in-chief of the 10th NAB Engineering Handbook. Mr. Williams is a member of the IEEE Broadcast Television Society AdCom and Broadcast Symposium committee; he received the Third Millennium Medal in 2000, and is a member of the Association of Federal Communications Consulting Engineers, the Society of Motion Picture and Television Engineers, Society of Cable Telecommunications Engineers, and the Society of Broadcast Engineers.

PROCEEDINGS OF THE IEEE, VOL. 94, NO. 1, JANUARY 2006