EBU Likes Progressive Scan HDTV


Europeans hope to avoid format debate

Although European terrestrial broadcasters, cable operaters, and satellite broadcasters have been broadcasting DTV for several years, they have, until recently, largely avoided HDTV. Probably in part because they are still smarting from the highly expensive and stunningly unsuccessful HD-MAC (High-Definition Multi-plexed Analog Component) experiment, they instead elected to concentrate on widescreen 16:9 SD as they entered the digital television era.

HD-MAC, as its name implies, was an analog component HDTV delivery system that, like the Japanese MUSE system, occupied so much RF bandwidth that it had to be broadcast via satellite.

There is a single European satellite broadcaster, Euro1080, that is delivering (1080i) HDTV to its customers. Others, including the BBC and BSkyB in the U.K., France's TF-1 and German pay-TV provider Premiere are making plans to offer HDTV in the future. This, as we know, involves a lot of choices--production and delivery scanning formats and coding schemes, to name two.

The EBU (European Broadcasting Union), is an organization of European state broadcasters. When the EBU was formed, state broadcasters were virtually the only broadcasters in Europe, but there now many private terrestrial broadcasters, cablecasters, and satellite broadcasters in Europe. Despite this television proliferation, the EBU is a large and influential body in European television, particularly in regards to technical matters. Their technical choices will have a significant impact on all HDTV broadcasting in Europe and elsewhere.

THE EBU & HDTV

The EBU has recently been addressing the questions involved with HDTV broadcasting, and there are some documents available on the EBU Technical Review Web site ( www.ebu.ch ) that describe its deliberations and conclusions. Of particular interest is "High Definition for Europe--a progressive ap-proach," by David Wood, secretary of the EBU Project Group B/TQE (Televi-sion Quality Evo-lution). This document describes the deliberations and conclusions of the TQE group regarding emission or delivery scanning formats. We note in passing that the European broadcasters have had the benefit of weighing the experiences of those areas that have a history of HDTV broadcasting; notably, the United States, Australia, and others.

Although the HD-MAC experience caused European broadcasters to question whether HDTV would ever be viable in Europe, the situation is rather different now from 12 years ago. There are three reasons to indicate the time for HDTV has come in Europe. First, large advanced flat-screen displays are becoming readily available at prices that the public is willing to pay. Second, HD-DVD hardware and software are imminent, which will stimulate the public's interest in HDTV. Third, HD-quality PC displays are becoming widely used, which will acclimate the public to HDTV.

The paper states three core technical issues that require analyses and decisions in order to implement HDTV in Europe: 1) The HDTV delivery platform(s)--satellite, cable, and/or terrestrial; 2) The scanning format to use for the delivery channel; and 3) The compression system to be used. The TQE paper focuses on the delivery scanning format.

In April 2004, the EBU Technical Committee recommended the use of progressive scanning, such as 720p/50 or 1080p/50, for HDTV emission standards. They further noted that production and emission standards do not have to be identical, as we in the United States are well aware.

The TQE explored whether a 60 Hz HD scanning format, rather than a 50 Hz format, should be adopted. They noted that 60 Hz offers advantages in motion portrayal and large-area flicker, and that it would certainly be advantageous to have a common worldwide standard. The price to be paid would be a higher bit-rate requirement.

However, because a worldwide 60 Hz HD scanning format is unlikely to materialize, and because of the serious frequency congestion in Europe argues for the lowest bit-rates possible, the TQE does not recommend pursuing a 60 Hz scanning format.

The TQE paper argues that interlaced scanning may have been the correct decision when made in an earlier, CRT-based world, and that it might still be correct under other circumstances. But in today's world of advanced displays, contemplating a new service beginning with a "blank sheet of paper," the committee can only conclude in favor of progressive scanning.

The principal reasons follow. First is the coding gain offered by progressive scanning. The group canvassed academic opinion among those who had no financial or personal stake in interlaced or progressive scanning, and found that a consistent answer was forthcoming. There is a "coding gain" associated with progressive scanning and adaptive compression, compared to interlaced scanning and adaptive compression, although, as it is scene- and content-dependent, quantifying the gain is difficult. It was noted that tests with H.264 (MPEG-4 AVC, which according to all indications will be used for HDTV compression in Europe), and Windows Media 9 have established that both compress progressive images "better" than they compress interlaced images.

The second major reason to use a progressive distribution format is to avoid interlaced-to-progressive conversion at the display. All the advanced displays are fixed-pixel displays, and all are progressively scanned themselves. While it is a relatively straightforward process to convert a progressive image to interlace, it is not at all easy to convert an interlaced image to progressive, particularly an interlaced image that portrays motion.

It was further noted that the future broadcast chain will begin and end with progressive scanning. Advanced displays are progressive as stated above, and current capture devices are as well. The electrical signals from all rows of both CCD and CMOS sensors are transferred to storage at the same instant, and a progressive or an interlaced image is read out of storage as required.

In fact, the interlaced image, consisting of every other line of the stored image, is generated by discarding information. Essentially the same thing is true of a telecine or film-to-video transfer device. The argument is that if both ends of the television system are progressive, there is no reason to introduce an unnecessary limitation on quality, in the form of interlace, into the middle of the chain.

1080p OR 720p

Having come down squarely in favor of 50 Hz progressive distribution, the question of which progressive format, 1080p/50 or 720p/50, remained. A series of tests conducted by the BBC revealed that the median viewing distance for a large flat screen in the home would be about 2.7 meters (8.9 feet), a number remarkably similar to the one the BBC measured 15 years ago, and to the RCA Labs Lechner Distance. Further tests on perceived detail led to the conclusion that 720p/50 would saturate the eye with detail for screens up to 50 inches at the representative viewing distance. 1080p/50 would, then, provide more detail than necessary for the screen sizes and viewing distances that would probably exist in European homes.

A survey of display manufacturers indicated to the committee that the "core" or "design target" flat-panel display for television should be the "Wide XGA" panel, a 768-line progressively scanned display. This is a good fit with the above finding that 720p would saturate the eye with detail in the screen sizes and viewing distances expected to exist in European homes.

Another question considered was what the implications would be if, as one manufacturer and several Japanese broadcasters stated, the core flat-panel display becomes a 1080p display rather than a Wide XGA one. If people were watching 1080p displays and 1080p DVDs were prevalent, would 720p broadcasts look inadequate? It was pointed out that if watched at the 2.7-meter viewing distance, the difference between 1080p and 720p would not be noticeable. It would, however, be noticeable if viewed at closer distances.

The committee concluded that it was best to begin HDTV broadcasting with the adequate, lower bit-rate 720p system. If receivers and set-top boxes were able to decode formats up to and including 1080p, this would permit broadcasters to move to 1080p broadcasting if the alternative scenario develops.

It should also be pointed out that the anticipated compression efficiencies that may be achieved with MPEG-4 AVC would likely permit 1080p/50 broadcasting, although on the other hand, 720p/50 would ensure fewer coding artifacts as well as lower emission bit-rates.

The conclusion of this work is that the EBU Technical Committee recommends progressive scanning for HD delivery for a number of reasons, the principal ones being that it obviates the requirement for interlace-to-progressive conversion at the display, and that it makes the delivery system more future-proof. The EBU Project Group B/TQE has concluded that a single standard, 720p/50, could be chosen to deliver HDTV-quality images to large flat-panel displays, providing the lowest delivery bit-rate requirement for images that satisfy the demands of the viewer, assuming the Wide XGA display.

Earlier HDTV implementations, such as those in the United States, were largely based on earlier presumptions, including the presumption of CRT displays. The European HD implementation will take place in a much different world of display choices, mainly based on progressively scanned fixed-pixel technologies, and in a world where more advanced compression technologies are available. The European broadcasters are therefore in the enviable position of starting with a "blank sheet of paper." In the world they contemplate, they see little reason to use interlaced scanning for delivery to the viewer.

Randy Hoffner