NII White Papers--Table of Contents NII White Papers--Chapter 19 NII White Papers--Chapter 21

The Evolution of the Analog Set-Top Terminal to a Digital Interactive Home Communications Terminal



H. Allen Ecker and J. Graham Mobley
Scientific-Atlanta Inc.

ABSTRACT

This paper addresses the evolution of the cable home terminal from past to present and presents a most likely scenario for its future evolution. Initially, a simple converter allowed subscribers to tune more channels than the initial VHF Channel 2 through Channel 13. Next, conditional access and addressability allowed programmers and network operators to offer subscription-paid premium programming. Today, advanced analog home communications terminals (HCTs) allow downloading of software to provide electronic program guides, virtual channels, and menu-driven navigators. Evolution will continue from advanced analog HCTs to a fully interactive digital HCT.

The enabling factors that allowed the cable industry to grow to become the primary entertainment delivery system to the home are the following:

  1. The availability of satellite delivery of content to headends;
  2. The availability of a broadband plant allowing many channels; and
  3. The availability of set-top terminals that have grown in functionality to truly become HCTs with reverse path capability.

Currently, cable systems provide broadband delivery to the home that allows broadcasting of many channels. In the broadcast mode, all programs pass each HCT so that each and all HCTs can receive and display the programs. In most instances, premium programming is scrambled before transmission so that only authorized customers can descramble premium channels. The authorization can be handled through individual addressing of each HCT.

To evolve to an interactive digital terminal that allows client/server type applications, four key ingredients are required.

The first is the requirement for a migration strategy to digital that uses broadband hybrid fiber coaxial systems. This strategy allows the digital signals to coexist with analog broadcast channels from which most revenue is derived today.

The second ingredient is the development of high-density integrated circuits to reduce the cost of the complex digital and analog signal processing inherent in any interactive terminal. Circuit density will increase from 0.8 micron to 0.5 micron to 0.35 micron geometries over time.

Third is the development of a multimedia operating system (OS) and user interface (UI) that allows the user to navigate and interact with the wide variety of content applications made available. This OS/UI must also be programmer friendly so that many different applications providers can develop applications that will run on home terminals.

Finally, the system must be designed for full interoperability at critical interfaces and must operate efficiently within the regulatory environment. Specifically, the interactive home terminal must contain a network interface module (NIM), which would be provided by the network operator but could have a consumer digital entertainment terminal (DET) that could be moved from network to network and that has an open architecture operating system and an open application program interface (API).

HISTORICAL BACKGROUND

Community antenna television, or CATV as it is now called, has its roots extending as far back as the late 1940s. It is widely accepted that the first CATV system was constructed in Mahanoy City, Pennsylvania, by an appliance store owner who wanted to enhance the sale of television sets by making programming available to customers that could not get reception of off-air signals since the community was located in a valley in the Appalachian Mountains. A large antenna was erected on top of a utility pole atop New Boston Mountain and was used to feed wires and repeater amplifiers. The system was "able to provide sufficient reception of television programs so as to not only sell television sets but to obtain subscribers to his cable connection" [1]. From this early attempt to provide cable service, additional systems were constructed with the intent to provide clear reception of locally generated signals to customers unable to receive off-air signals. By the mid-1960s there were over 1,300 cable systems feeding approximately 1.5 million customers.

Architecturally, these early cable systems were merely distribution systems that received and amplified off-air signals without processing and reassigning redistribution frequency. However, the 1970s brought new technological developments. The launching of geostationary satellites, coupled with the allowance by the FCC to make these satellites available for domestic communications, brought about the growth of satellite delivery of programming directly to cable headends. For the first time, television programmers and programming distributors could use satellites to broadcast their programming with complete U.S. coverage. CATV systems could receive and distribute these signals via cable to customers with little incremental investment. Television broadcast stations such as WTBS in Atlanta and WGN in Chicago made use of this technology to create superstations with full continental U.S. coverage of their signals. In addition, FM satellite receivers, large satellite receiving antennas, and video processing equipment provided cable headends with the necessary electronic equipment to receive and distribute these programs over selected channels in the cable system.

HOW THE ANALOG CABLE SET TOP EVOLVED

For the first time, there were more channels available over the CATV system than the normal Channel 2-13 VHF television could tune. This availability of greater channel capacity led to the development of the cable tuner or "set-top converter." These early units, developed in the late 1970s, were capable of tuning up to 36 channels. They did little more than convert the tuned cable channel to a standard Channel 3 or Channel 4 output frequency so that a standard VHF television set could receive the cable signals. In effect, these units were the cable analog of the UHF converter prevalent in the late 1950s. Recognizing the vast market made available by the satellite and cable distribution system in place, content providers such as Home Box Office, Cinemax, Showtime, The Movie Channel, and The Disney Channel as well as ESPN, USA Network, and others sought to use the existing infrastructure to provide "cable only" delivery of premium and "niche" programming content to a growing audience hungry for content variety.

This programming was meant to service only those who purchased it and therefore required means for selectively enabling only those viewers who paid for these special services. To fulfill the requirement, frequency traps were installed in cable drops to nonsubscribing customers. This technique had several drawbacks:

  1. Traps were used to prevent reception, proving to be a costly proposition since the expense was distributed over nonsubscribing customers.
  2. Customers who changed their viewing tastes required a service call by the cable operator to either remove or add traps.
  3. If the cable operator added a new premium service, new traps were needed for each subscriber not desiring the new service.

Scrambling and addressability removed these road blocks and made premium programming a viable business. In the 1980s, scramblers began replacing frequency trap technology. Premium programming content was scrambled before being redistributed over the cable system. Newer set-top converters were provided that contained addressable descramblers. These units had unique preloaded serial numbers so that individual converters could be addressed remotely from the cable headend and either enabled or disabled for descrambling of particular premium programming content or tiers of programs.

More Channels and Better Signal Quality Using Analog Fiber

In an attempt to increase channel capacity, improve penetration, and provide better quality of service, cable systems in the 1990s substituted analog fiber technology for the coaxial RF trunk and distribution systems. Analog fiber provided a number of advantages:

  1. It reduced the number of amplifiers needed to span large distances, thereby increasing reliability and signal quality.
  2. It allowed cable systems to penetrate deeper into outlying areas without a sacrifice in signal quality.
  3. It reduced the powering requirements.
  4. Analog fiber was selected instead of digital fiber because it made conversion from fiber back to analog coaxial distribution inexpensive and simple.

The addition of analog fiber also changed the architecture of a typical cable distribution plant. In the process of providing better signal quality deeper into the distribution system, the topology slowly converted from a "tree and branch" layout, in which feeder lines split from a main trunk, to a "star" topology where clusters of subscribers were connected to a fiber node that was fed directly from the cable headend. This architecture, known as fiber to the serving area (FSA) or hybrid fiber coax (HFC), reduced the number of subscribers being fed from a single node and increased the number of nodes fed directly from the cable headend. This evolution has enhanced a cable system's ability to convert from a broadcast topology, where bandwidth is assigned to a program, to an interactive topology, where bandwidth is assigned specifically to connect a content provider and a single customer.

Hybrid fiber coax also provides more bandwidth for reverse path signaling. In the Time Warner Full Service Network in Orlando, reverse path amplifiers are used to connect HCTs to the fiber node in the spectrum between 750 MHz and 1 GHz. Reverse path time division multiple access (TDMA) signals are converted from RF to analog fiber at the nodes and piped back to the headend over fiber. Other systems operate in the same way except that the band from 5 to 30 MHz is used between the fiber nodes and the home HCT. The "star" topology of HFC provides more reverse path bandwidth per home because of the lower number of homes connected to a node.

Cable Becomes a System

Cable in the 1990s has evolved into a complete communications system that provides broadband delivery to the home of many channels, both basic unscrambled as well as premium scrambled channels. Through the use of computer-controlled system managers coupled with a conditional access system, individual set-top terminals can be quickly and efficiently electronically addressed to allow for the descrambling of particular premium channels.

To aid in program selection, advanced analog set-top terminals today feature downloaded fully integrated program guides and a system navigator that informs subscribers of the vast programming available. Also, using the 150 kbps average bandwidth available in the vertical blanking interval of a standard video channel, these units incorporate "virtual channels" that can be used to send sports updates, financial services, news, and other types of digital services. These units also can be equipped with a standard read-only-memory (ROM)-based operating system or with an addressably renewable operating system (AROS) that allows the operator to download new software to the HCT via the system manager to provide new user interfaces, unique electronic program guide on-screen formats, bitmapped graphics, and multilingual on-screen displays.

Also, reverse path signaling has made impulse pay-per-view (IPPV) and near-video-on-demand (NVOD) a practicality. With an integrated IPPV module, subscribers can view events at their leisure. Also, NVOD capability offers subscribers advantages similar to those of VCR tape rental by defining events with staggered start times on multiple channels. Advanced HCTs have the capability to emulate a VCR's pause, fast-forward, and rewind features by making use of the staggered simulcast nature of NVOD channels and the built-in intelligent software resident in the HCT to keep track of the appropriate channel needed to view a desired program segment.

Digital Communications with Analog TV Signals: An Added Dimension

While current cable systems use analog signals for video and audio, advancements in digital technology now allow cable systems to add a digital video layer to increase channel capacity with little or no increased distribution bandwidth.

Two technological breakthroughs in digital processing are clearing the way for digital video and audio program content. The first is the adoption of standards for digitizing, compressing, and decompressing video programming. In 1992, the Moving Picture Experts Group (MPEG) of the ISO set out to develop a set of standards for digitizing and compressing an analog video signal. This international standards group laid the groundwork for standardizing the algorithms, syntax, and transport format necessary to allow interoperability among different suppliers of video compression and decompression systems. The attitude at the outset was that if digital television was to flourish, equipment built by different vendors must be compatible and interchangeable. The international standards adopted by the MPEG committee allow freedom to be creative in the encoding process within the parameters of the defined "encoding language" while maintaining compatibility with standard MPEG decoders. The analogy in the computer programming world is to say that software programmers can approach a programming problem in many different ways; however, if their program is written in C, for example, any C compiler can compile and execute the program.

The second development was a means for delivering the digital signals to the customer. Schemes for modulating and demodulating an RF carrier using either quadrature amplitude modulation (QAM) or vestigial sideband modulation (VSB) have been developed. These approaches are compatible with standard analog cable systems and can deliver data information rates up to 36 Mbps in a single 6-MHz channel. The combination of digital compression of video and audio and digital transmission over cable can increase the number of video services in a single 6-MHz channel by a factor of approximately 5 to 10 depending on programming content and picture resolution.

Not only does digital processing increase capacity, but it also allows greater flexibility in the types of services that can be provided. No longer will there be the restriction that the information provided be a video signal. Digital computer games, digital music, and other multimedia applications will all be likely candidates for the new broadband digital cable.

The cable set-top terminal needed to fully utilize these new services will evolve to a fully interactive home communications terminal (HCT) that allows client/server functionality.

Migration Strategy for Broadband Cable Systems of the Future

So how do we evolve from the cable system of today that broadcasts analog video services to the home to the fully interactive digital system of the future? This evolution will most likely occur in four phases.

The first phase in the migration would employ advanced analog home communications terminals (HCTs) that are compatible with existing hybrid fiber coax distribution plants. These terminals, available today, not only can tune up to 750 MHz but also have both forward- and reverse-path digital communications capability for downloading digital applications, providing on-screen display of program guides and hand-held remote navigators for ordering pay-per-view and NVOD.

The second phase most probably will use digital broadcast HCTs that can not only tune and display the standard analog video but also receive and decode digital video services. These digital services would be overlaid on the HFC cable plant using unused spectrum above the standard analog video channels. This strategy would allow the digital services to coexist with analog broadcast channels from which most of today's revenue is derived. System upgrades and even "new builds" will need to provide today's analog channels to existing set-top terminals and customers who are content in viewing basic and premium scrambled broadcast services. A hybrid fiber coax system transmits the digital information on analog RF carriers designed to be frequency multiplexed into a normal broadband channel. The bandwidth required is compatible with a normal analog video channel so that the digital signals can be assigned a vacant space in the spectrum or, because of enhanced ruggedness, can be placed at the upper end of the spectrum outside the guaranteed performance range of a distribution plant.

The third phase in the migration would provide the analog/digital HCTs with full interactive capability. This phase would require an incremental upgrade in the HCT to provide more memory and enhanced reverse path signaling and full interactivity. However, the major upgrades would evolve over time as the plant architecture migrates from a "tree and branch" to a "star" configuration and more cable headends become interconnected. This change would be coupled with the growth of content-provider digital applications stored on file servers, market demand for specialized applications, and the success of true "video dial tone" as compared with NVOD.

The final phase, illustrated in Figure 1, is a fully integrated broadband network that provides full connectivity to the home for POTS services, broadband digital modems for personal computers (PCs), analog and digital video services, digital applications, and utility monitoring. Either a fully integrated HCT or a split between the functionality of an HCT and a customer interface unit (CIU) could be used. This last phase could happen as slowly or as rapidly as needed depending on market demand for interactive services.

The phases outlined make sense because they allow for the coexistence of analog only, digital broadcast, and full interactive services and HCTs on the same system and do not require flash cuts to implement, nor do they sacrifice analog video revenues. Thus, the revenues track with the cost of upgrades.

COST DRIVERS FOR THE FULLY INTERACTIVE HOME COMMUNICATIONS TERMINAL

This migration will be paced by key cost drivers in the HCT, the distribution plant, and the headend. HCT cost appears to be the controlling factor because each digital user must have a digital HCT. Also, developing an economic model that allows enough incremental revenues to support needed upgrades will be key.

The cost of the fully interactive digital HCT depends on the availability of a cost-effective set of application-specific integrated circuits (ASICs) for digital and analog signal processing and digital memory inherent in the interactive digital terminal. Currently most if not all the necessary ASICs have been designed to perform the critical processing functions. These functions include digital demodulation of 64/256 QAM digital signals, adaptive delay equalization of the digital signals, error correction, program demultiplexing and conditional access, and finally MPEG video and audio decompression. Also, microprocessors and graphics accelerators are available for a platform for the terminal operating system (OS) and user interface (UI). Several trial systems such as the Time Warner Full Service Network in Orlando, Florida, and the US West fully interactive trial in Omaha, Nebraska, are using HCTs that include these ASICs. In 1995 other systems using these chips will be deployed by BellSouth, Ameritech, Southern New England Telephone, and PACTEL.

However, these chips are currently designed using today's 0.8-micron technology, and the number needed to build a fully interactive HCT is about seven to nine chips containing about 975,000 gates and 1,500 pinouts. To reduce cost, the goal is to design the required functionality into one or two chips. Projections are that by mid-1996 the number of ASICs could be reduced to four using 0.5-micron technology, and by 1998 the goal of two

chips could be realized using 0.35-micron technology. Reducing the number of chips would dramatically reduce the number of pinouts, thereby decreasing cost and increasing liability.

The second key cost driver is the cost per subscriber for the headend and distribution plant. In an analog broadcast or digital broadcast scenario, including NVOD, the cost per subscriber for the headend and distribution plant can be normalized over the number of subscribers. Therefore, increasing penetration reduces operating cost per subscriber. However, in a fully interactive network, interactive server/client bandwidth must grow proportionally with the number of subscribers, so that plant and headend cost increases with penetration to meet the anticipated heavier demand for simultaneous interactive sessions between client and server. An acceptable value for subscriber costs for the fully interactive system is heavily dependent on the economic model for incremental revenues that might be expected from the additional services available to a customer. To put it another way: Just how much is a customer willing to pay and for what types of services? To answer this question, a number of fully interactive trials are being conducted for consumer acceptance as well as technical evaluation.

Current projections for 1995 96 revenues that might be expected with various services using advanced analog HCTs, digital broadcast mode HCTs, and digital interactive HCTs are shown in Table 1.

These projections indicate that a fully interactive HCT could provide almost a 100 percent increase in monthly revenues over an advanced analog unit. However, a large portion of the projected increase is in the vague category defined as "other services" that include home shopping and other services whose revenue potential is not proven.

Currently, a reasonable compromise is to provide advanced analog HCTs with NVOD capability for the near term, but add additional capacity using digital broadcast, and offer digital interactive services on a customer demand basis with digital interactive HCTs.

REQUIREMENT FOR INTEROPERABILITY AND INDUSTRY STANDARDS

The promise of emerging digital interactive services cannot reach critical mass without the elements of interoperability, portability, and open standards at critical interface points in the system. Because of the complexity of the systems involved, these elements are best resolved through industry groups representing major providers of equipment as opposed to government standards that could overlook critical issues.

The following are one set of definitions for the above terms:

Interoperability, consumer equipment portability, and open standards are critical to the success of a fully interactive system for at least three reasons.

First, the risk is lowered for all participants. Equipment manufacturers are able to invest in the development of products that are guaranteed to be compatible with all networks and the suppliers of content on those networks.

Second, it assures the lowest cost hardware and content. Competition between equipment manufacturers to produce in large volume and therefore achieve economies of scale is enhanced because these products can serve the entire market base as opposed to many smaller proprietary networks. Also, content providers will be inclined to participate in content and applications development because these applications will be compatible with all consumer terminals.

Third, it will jump start the creation of many new interactive applications by removing the chicken-and-egg phenomenon. Once standards are agreed upon, content providers can begin working on applications software knowing that network architectures and consumer terminals will be available in the near future. Likewise, once attractive applications are available, consumers can acquire home terminals and subscribe to the network.

A major step toward the goals of achieving interoperability, portability, and industry standards has been accomplished by the MPEG committee in the adoption of standards for compressing, transporting, and decompressing video signals. As a result, chip manufacturers have developed a set of decompression chips that are functionally interchangeable, and equipment manufacturers now have a video standard to work from. Because of MPEG's success, a digital audiovisual council called DAVIC was formed in April 1994 to help resolve remaining compatibility issues. This committee currently has over 120 member companies worldwide. The charter for DAVIC is to assist in the success of the emerging digital video/audio applications and services by providing timely agreed-upon specifications for interfaces and protocols. DAVIC is working on a timetable that called for a strawman baseline document in December 1994, experiments and tests to occur during the spring and summer of 1995, and publication of final recommendations by December 1995.

While the goals of DAVIC are ambitious to say the least, much of the intent could be realized by recognizing that two critical interface points exist in the fully interactive network.

The first interface where standards are important is the connection between broadband integrated gateways at a cable headend and the data networks used to achieve full connectivity between cable systems and between cable systems and telephone companies. An interface standard must be adopted that allows SONET, FDDI, Ethernet, ATM, and other protocols to provide inputs to the access part of the network easily. This approach would provide the connectivity necessary to access remote content providers either over satellite, or from remote video file servers on the metropolitan network or connection through a trunk from other areas. Interactive applications providers with the capability to communicate over the Internet or get applications resident outside the local cable network can also be sources of content.

The second critical interface connects the broadband network to the home. If interoperability and portability are to be realized along with network integrity and flexibility, the home terminal must split into a part that must be provided by the network operator and a part that is not necessarily a part of the network and could evolve to becoming consumer electronics sold at retail.

A first part of the terminal that could be called the network interface module (NIM) must be provided by the network operator for several reasons:

The NIM would be provided by the network provider as a standard sized plug-in module with standardized I/O on the customer side and specialized interface to the network. It would contain the tuner, demodulators, conditional access system, and reverse path signaling common to a specific network and would allow the many different network architectures, signaling structures, and connection management systems, many of which are already in place today, to communicate with a part of the HCT that one day might be purchased at retail. To add further flexibility and network integrity, the addressability and conditional access system (AC&E) could be installed on a "smart card" that plugs into the NIM.

This HCT concept can be explained by Figure 2, which shows the major functional blocks contained in an interactive digital HCT. These include the following:

CONCLUDING REMARKS

The following statements are supported by the previous sections:

Reference

[1] Goodale, James C., 1989, "All About Cable," Law Journal Seminars-Press, Chapter 1, p. 6.