COFDM VS 8-VSB

A meeting was held in Washington, DC January 11th at the Ronald Reagan Convention Center. It was hosted by MSTV, a Washington based technical support association for broadcasters, and the National Association of Broadcasters (NAB), the Washington based association for a large portion of the broadcasters in the nation. Of the 170 in attendance, 140 were broadcasters. The rest were staff and lawyers from the NAB and MSTV. They had gathered to hear the first official presentation of the MSTV comparison test between 8-VSB and COFDM. As most know, a highly contentious question arose 19 months ago from tests conducted by Sinclair Broadcast Group. In side-by-side demonstrations the 8-VSB receivers were not able to receive indoor signals equal to their rival COFDM receivers. That set off a flurry of activity and no inconsiderable angst throughout the industry, which finally led to more scientific tests being run the last half of 2000 under MSTV supervision.

The NAB will hold their next board meeting in Los Angeles on the 15th of January to discuss the results of the tests and search out the means for drawing its members to a consensus. The NAB has not been able to speak clearly with one voice on this modulation question since its members broke into two camps.

One might come away from a first reading of this report thinking that the differences between the two are now only modest and, indeed, often favoring the 8-VSB. The results certainly do not give COFDM a sweeping victory, which its proponents had long-predicted. The 8-VSB did better than most had thought it would. The problem this editor sees is that neither worked to meet the original criteria set by broadcasters in the early 90s and thought answered by the 8--VSB selection. No greater than 50% of any of the 6 ft mast test sites in difficult areas (the tests were to determine how difficult multipath receiving areas would handle things. The easy receive areas were not tested since they would work under most all conditions) could receive either 8-VSB or COFDM, bringing to question the viability of digital broadcasting as being a suitable replacement for analog.

Manufacturers of HDTV 8-VSB receivers had reacted to the uncertainly raised by the testy modulation argument by placing more emphasis on satellite receiver markets as their means for driving the HDTV business. But no traditional networks yet broadcast their HDTV services from the two commercial satellite services--Echostar and DirecTV. At this point only HBO and Showtime are delivering their HDTV services from the commercial satellite companies, along with some pay-per-view. Both movie channels have looked to traditional broadcasting as the chief driver for the penetration of HDTV receivers. They think of themselves as little more than riders of a trend rather than drivers of it.

The cable DTV "must carry" hearing was also scheduled by the FCC on the 11th and was postponed. The cable industry has time now to look over the MSTV/NAB report and determine if they are being asked to carry DTV stations with coverage equal in area to that of NTSC even though that coverage is only partially achieved with digital signals. Is there a legitimate case, then, for compensation should a must carry ruling occur? It is not the intent of cable "must carry" rules to increase the service area of a broadcaster, but rather to carry signals into the broadcaster's market already covered by that broadcaster's signal.

It will take a few more days to sort out the information presented below, plus the more detailed report being studied by the sponsors of the tests. It would look from here that the FCC can find no grounds--political or technical--to endanger the roll out of H/DTV by opening the standard to a process for including the COFDM option, which Sinclair Broadcast Group has asked for. There appear to be no compelling reason shown from these tests to make that move unless new data is introduced. Already from Europe proponents of the COFDM DVB system are excusing the relative weak performance of DVB saying that the receivers used were not tuned to the US environment, and so were not up to their potential. That would be true also for the origianal receivers used in the Sinclair tests. The only plausible reason for reconsideration by the FCC would be to accomodate the still-untested mobile reception, which COFDM proponents say can only be accomplished using COFDM.

The hope of the industry has been that the MSTV test would settle once-and-for-all time the terrestrial modulation issue. If there was enough margin to warrant the expense of retooling the standard, that would have to be the choice and then done. But that margin would have to be substantial enough, and it would have to be a somewhat fixed margin. By that I mean that improvements in the 8-VSb could not be predicted to further narrow the margin. Few will say that the margin has not already been narrowed by the use of the latest available 8-VSB receiver improvements. So, the issue appears to be well on the way to being settled, though COFDM advocates are not likely to give in to this view easily.

In summary it would appear that the poltical legs are weakened for COFDM in the United States, and one can add to this speculation that other regions of the world seeking affiliations with a large manufacturing base to lower the cost of their own DTV decisions must now reconsider their modulation choices. In other words, 8-VSB is back in play, if it was ever out-of-play.

Dale Cripps

January 12, 2001

In the Spring of 2000, the VSB/COFDM Project, implemented and funded by 30 broadcast organizations, began a testing and research program to provide thorough, impartial and scientific facts to inform the television broadcast industry about the current state of the art implementations of VSB and COFDM technologies for service in the United States and the directions both technologies might take in the future. Under the guidance of a Steering Committee composed of the Project members, and under the ultimate supervision of the MSTV Board, two Technical Groups were established:

1) The VSB Technical Group was tasked to test the newest VSB receiver products manufacturers made available and to investigate potential improvements and modifications to the 8VSB standard to accommodate existing and new service applications; and

2) The COFDM Technical Group was asked to evaluate the relative performance of the best available implementations of COFDM and 8VSB in a comprehensive comparative field test program for outdoor, indoor and portable reception at various distances from a number of different transmitters. The results of this Phase 1 effort were to be used to determine whether to continue with a subsequent Phase 2 effort to further investigate COFDM technology for the U.S. broadcasting environment.

This is the report of the COFDM Technical Group. The VSB Technical Group report is being submitted to the Steering Committee contemporaneously.

The COFDM Technical Group first developed a test plan for the comparative field test program and developed methods for comparing various aspects of the VSB and COFDM technologies as currently implemented. These aspects include portable reception and ease of reception as well as traditional outdoor and indoor receivability. To perform the comparative testing, COFDM and VSB signals were alternately transmitted from the same transmission facilities at the same average power level.

Testing was conducted in the Washington DC/Baltimore area and in the Cleveland area. The Washington/Baltimore tests involved transmissions from four different UHF DTV stations to about 200 outdoor sites and 45 indoor sites. The Cleveland tests involved transmissions from a low VHF DTV station to about 100 outdoor sites and 20 indoor sites.

In both Washington/Baltimore and Cleveland, reception was tested with antennas placed outdoors at 30 feet, outdoors at 6 feet, and indoors. The Washington/Baltimore location offered a wide range of challenges for DTV reception, including severe multipath, interference and low signal areas. The Cleveland location offered the particular challenge of impulse noise and other RF propagation characteristics associated with low VHF operation. Because the Cleveland DTV facility (channel 2) was collocated with an NTSC facility (channel 3), the tests also compared DTV and NTSC performance.

VSB and COFDM equipment was solicited from manufacturers for use in this program. Because there is no COFDM system currently employed in the U.S., equipment currently in use abroad had to be configured for the U.S. 6 MHz channel spacing. The parameters of the COFDM system were chosen to yield the same approximate net data rate as the VSB system (19.76 Mb/s for COFDM compared to the 8VSB net data rate of 19.39 Mb/s). The COFDM equipment complied with the DVB-T standard while VSB equipment complied with the ATSC DTV Standard. Laboratory evaluation of several candidate modulators and receivers was initially performed to select the best equipment available and confirm the expected performance of the units to be used in the field. In order to accelerate the data collection process, four similarly equipped field trucks were used to gather the field data. Field tests began in mid-August and were completed in mid-December.

In addition to conducting the comparative field test program, the COFDM Technical Group assessed the predicted coverage and interference impact of VSB and COFDM modulation schemes using the existing FCC DTV channel assignment plan. The Group used the same average transmitted Effective Radiated Power for both modulation schemes in order to minimize any new interference caused to the existing NTSC service. It then conducted laboratory tests to determine the appropriate interference criteria values for COFDM and VSB to be used in a DTV spectrum planning model. This model, implemented using a DTV spectrum planning computer program, then calculated predicted nationwide coverage and interference impacts for each of the two technologies.

Findings
0000000The COFDM Technical Group made the following Findings.

Washington/Baltimore (UHF)

At a 30 foot receive antenna height, 8VSB was successfully received at a greater percentage of sites than was COFDM for all four stations. This was true at all distances from the transmitter (grids, clusters, arcs and extended radials). 8VSB performed better up to the farthest distances measured from the transmitter (55 miles).

At a 6 foot receive antenna height, using a simple antenna, COFDM was successfully received at a greater percentage of sites than was 8VSB for all four stations. This was true for sites close to the transmitters (grids and clusters). At greater distances from the transmitter (arcs), the performance was very close between the two systems. Successful reception of either system was achieved at less than 50% of sites, which is disappointing.

For indoor measurements, the percentage of successful reception was similar for 8VSB and COFDM. 8VSB held a slight advantage for signals from three of the four stations and COFDM held a slight advantage for signals from the fourth station. This was true for both the bowtie and the Silver Sensor antennas. Successful indoor reception was achieved at only about 30% of sites, which is disappointing.

In the "ease of reception" testing (antenna pointing sensitivity) for each station individually, COFDM outperformed 8VSB at 6 feet. Both systems performed equally at 30 feet. Because of the directionality and gain of the antenna used at 30 feet, the 6 foot data is more representative of a viewer experience with a simple antenna.

Cleveland (low VHF)

At a 30 foot receive antenna height, 8VSB was successfully received at a greater percentage of sites than was COFDM. 8VSB performed better than COFDM farther away from the transmitter (for radials and 25 and 50 mile arcs). COFDM performed better than 8VSB closer to the transmitter (grids).

At a 6 foot receive antenna height, 8VSB was successfully received at twice as many sites as COFDM for arcs, grids and radials. 8VSB achieved only a 28% success rate, which is disappointing.

For indoor measurements, successful reception for 8VSB was 9 percentage points better than for COFDM. However, even 8VSB achieved only a 26% success rate, which is disappointing.

The Cleveland facility offered the unique opportunity to study a co-located channel 2 DTV and channel 3 NTSC facility. For sites that had a NTSC CCIR 3 rating or greater (i.e., acceptable NTSC picture quality as conventionally defined for analog television broadcast service planning), 8VSB was successful 92% of the time, while COFDM was successful 78% of the time.

Other:

The data confirmed the theoretical carrier-to-noise performance difference of about a 4 dB advantage of 8VSB over COFDM.

For the 30 foot failed sites, 55% of the COFDM sites failed due to low carrier- to-noise ratio, while 40% of the 8VSB sites failed due to low carrier-to-noise ratio.

Based on the current DTV Table of Allotments (that is, assuming that no changes were made and that facilities using COFDM operated according to the parameters in the Table), COFDM would reduce the total DTV viewing population by 5.9% and the service area by 13.9%, when compared to 8VSB. The impact on individual station viewers and areas would vary greatly. The difference between COFDM and 8VSB in interference caused to the existing NTSC service would be minimal.

The 8VSB and COFDM receivers selected for use in the comparative field tests were carefully evaluated for typical performance areas to insure that they were the best receivers available for the test program and represented the current available state-of-the-art in DTV technology. Both technologies had areas where their performance characteristics could be improved. The performance deficiencies in some areas probably account for some of the failures observed in the field. Examples include the range of echo handling for the VSB receiver and a documented but unexplained anomaly (an apparently non-linear mechanism) in the RF section of the COFDM receiver.

Conclusions:

00000The COFDM Technical Grpoup Concluded the Following"

The results of the field testing of 8VSB and COFDM at the specified data rates indicate that 8VSB should be suitable for a broadcast service when a 30 foot outdoor antenna is used for reception.

The results of the field testing of 8VSB and COFDM for outdoor reception using a simple antenna at 6 feet are far less optimistic that the reliability of service for that reception configuration will be adequate for defining a broadcast service with either system tested.

The results of the field testing of 8VSB and COFDM indicate that although some viewers would be able to enjoy indoor reception with either system, neither system exhibited the level of reliability that would be required of a practical broadcast service based solely on service to indoor antennas.

Given the high level of failures at moderate-to weak-signal levels in Cleveland for both systems, the data suggests that the planning factors used by the FCC to predict the low VHF service are inadequate. Additional power may be needed to maintain adequate service levels.

Recommendations

The COFDM Technical Group was formed to investigate the relative performance of COFDM and 8VSB modulation for DTV service in the United States. The group conducted extensive field tests comparing the performance of both systems at the same field locations in a rigorously supervised manner. The results of these tests are summarized in this report. System performance for the two technologies is different. Each system showed varying degrees of performance under different circumstances. However, neither system in its current implementation will meet all broadcasters' or viewers' needs.

The COFDM Technical Group was further charged with the responsibility to make a recommendation at the end of the Phase 1 activities. If the Steering Committee wishes to develop a more complete assessment of the services supported by 8VSB and COFDM, the following expanded activities for a Phase 2 are recommended. However, these activities exceed the original Phase 2 scope and will require additional time and funding beyond that originally proposed.

Further analyze Phase I data including continued investigation of a COFDM receiver anomaly;

Conduct an initial Phase 2 test program, including the testing of other COFDM modes for both DVB-T (European) and ISDB-T (Japanese), evaluation of RF Signal Capture for channel characterization, and the technical investigation of the legacy issues for consumer and transmission equipment;

Investigate and field test on-channel repeaters, conduct high-band VHF testing, and investigate improved receiver performance.

The COFDM Technical Group concluded the following:
Conclusions

The VSB/COFDM Project is an all-industry broadcast project funded by voluntary contributions from broadcasters. The project was formed to conduct parallel scientific and impartial investigations of VSB improvements and COFDM performance. This report addresses only the VSB investigation.

The goals of the VSB investigation were to: (1) independently evaluate next-generation products and test their performance in the lab and in the field; and (2) investigate improvements to the 8VSB standard and possible modification of the standard to provide more robust reception and accommodate new services. The investigation was carried out by the VSB Technical Group, with participation by eleven broadcast groups in addition to MSTV and NAB.

The group issued a "Request for Proposals on VSB Improvement" to some two-dozen manufacturers and research and development organizations, seeking information on three options, in order of priority:

Receiver-only improvements that would not require any changes to the transmitted signal.
Compatible changes - or "enhancements" - to the transmitted signal, which could be used by new receivers to achieve a higher level of performance but would not affect the performance of existing receivers.
Non-compatible changes, based on VSB modulation, which would affect the performance of existing receivers.

Only seven responses were received, four of which were deemed "non-responsive". However, the group was pleased with the quality and potential of the three identified for follow- up action, submitted by well-respected companies with substantial research and development talent and resources. Members of the VSB Technical Group visited all three companies to obtain more information and to discuss ways in which we could provide support to expedite their efforts.

The principal technologies that have been proposed and are being pursued by the respondents to the RFP are: improved equalization, to handle real-world multipath conditions having complex combinations of delay, amplitude, and rate of change; training signal improvements; diversity reception; alternative modes of VSB modulation, 2VSB and 4VSB; and scalable approaches. The VSB Technical Group found the latter approaches particularly appealing, since they can transmit - within the same signal - a scalable combination of robust data that new receivers could decode, and "backward-compatible" data usable by both existing and new receivers.

The VSB Technical Group has identified five ways in which it can help to accelerate the development of VSB improvements - RF captures; simulation hardware and software; field testing; human resources; and "matchmaking" - and it has begun the deployment of appropriate resources, both in-kind services and funding. Since few next-generation products became available for testing during the time frame of this investigation, but both next-generation devices and simulations or prototypes of enhanced-VSB technologies are expected to become available for testing in early 2001, the VSB Technical Group elected to reserve a major portion of its funds to support this future testing.

Conclusions

The VSB Investigation has reinforced the findings and recommendations of the previous work conducted in 1999 by a broadcaster task force comprised of MSTV, NAB and members of the MSTV Engineering Committee.

Progress continues to be made in improving 8VSB reception performance, particularly with regard to the better equalization needed to handle complex dynamic multipath conditions. The few next-generation receivers that have been available for testing to date have shown advances over their predecessors. However, the pace at which improved consumer products are reaching the marketplace is disappointing.

The VSB Technical Group has seen presentations and simulations of further advances in components which will become available in the first quarter of 2001. We will work with manufacturers to conduct laboratory and field tests as soon as possible. However, due to the lead-time of a year or more between first simulation and integrated-circuit products, some of the improvements already developed and disclosed to the group under NDA will not emerge from that pipeline until 2002. Incorporation into receivers generally takes another year. We understand that very significant investment would be required to accelerate these cycles. However, without such acceleration, consumers may not see these improvements before the holiday season of 2003.

It is encouraging that several major, well-respected organizations with substantial VSB knowledge and experience are actively pursuing 8VSB reception improvements. These include both backward-compatible and non-backward-compatible enhancements, which can provide greater robustness for both existing and new services. The primary focus of the VSB Technical Group is on backward-compatible enhancements and, with our support to expedite their efforts, we will expect these organizations to follow through and provide simulations for testing in early 2001.

It is preferable to enhance performance and support new services through backward- compatible changes within the VSB family. However, it is essential that all testing and evaluation be conducted in a timely manner, since the impact of non-compatible approaches increases with time as the deployment of DTV rolls ahead.