HIGH DEFINITION TELEVISION
Copyright 1982 by Steven Bradford
Electronic Cinema. High Definition Television.
These two phrases have been bandied about considerably in the past year. What do they mean? What do they portend for the filmmaking business? Will they really revolutionize filmmaking? What are the costs and benefits? Are they really inevitable or just hype? What is the future of film?
To answer these questions it is best to examine our present television system first and discover how and why it came to exist.
THE PATH TO THE NTSC
The earliest television systems were clever combinations of mechanics and electronics. Using a spinning disc punched with holes that scanned the image, these systems produced low resolution but recognizable images. They had line rates from around 30 to as great as 250. But it was recognized that electronic methods of scanning the image held the greatest promise for good quality pictures. In the early thirties the first pickup tubes were invented by Philo T Farnsworth and Vladimir Zyworkin. By the early forties progress had been made to the point that regular commercial broadcasts could be started.
But first a standard was needed. The first National Television Standards Committee set them in 1941 for monochrome transmissions. At the time it was determined that it would not be possible to come up with a compatible color television system.1 Yet only twelve years later they managed to do so and the current standards were established.
THE NTSC AND OTHER STANDARDS
It is important to understand why the standards are what they are. They are standards set by people for various technical, economic and political reasons.
They are not immutable or engraved in granite. The standards could easily have 523 lines and 28 frames per second established as what we have now. In fact, NTSC equipment can be converted to the PAL system (Phase Alternating Line) without a great deal of trouble.
The established NTSC standards are:
1. 525 scanning lines
2. 30 frames per second with 2 to 1 interlace.
Each frame is divided into two fields. Each field consists of 2621/2 lines, one field consisting of the even numbered lines, the other of the odd numbered lines.
3. 4.25 MHz of bandwidth
4. A color sub-carrier at 3.58 MHz.
5. An aspect ration of 3 X 4.
Since none of these standards is necessarily inherent in good television broadcasting, why are they chosen? The 30 frame/60 field standard came about because the electrical system is standardized at 60 Hz. Therefore, one field for every cycle of the alternating current. It was difficult to make a system that could scan 525 lines 60 times a second in the late thirties. But a rate of 60 per second was required to eliminate flickering. To get out of this quandary the engineers devised the interlace method of alternating half resolution frames or fields. This high field rate is also the reason why television pictures do not strobe when the camera is panned quickly. It is also why freeze frames on videotape look so horrible. The recorder is only playing back one of the half-resolution fields. If it plays back the entire full resolution frame, the picture will jump wherever movement occurred between the two fields.
In countries using the 50 Hz alternating current the standard is 25 frames/50 fields per second. These are the PAL and SECAM systems. To make the flicker less objectionable these systems use the display tube at a lower voltage, making it dimmer.
The 4.25 MHz bandwidth was chosen so that channels could be separated by 6 MHz and provide for the number of television stations in a transmission area that the Federal Communications Commission wanted. A wider bandwidth would have reduced the possible number of channels while a narrower one would have allowed for more channels. This was primarily a political decision.
525 scanning lines is simply the number that results when 4.25 MHz and 30 frames per second is used. If these standards are changed then the number of scanning lines changes.2
The 1.33 aspect ration was chosen because that was the motion picture standard during the thirties.
In the fifties it was realized that color broadcasts could be made compatible with monochrome sets if the color information was encoded into the luminance signal. This color subcarrier is located at the 3.58 MHz point of the television signal. Because it is encoded into the luminance carrier, detail information is lost.
Television standards are like rubber; they can be pushed and pulled according to what trade-offs and compromises one wishes to make.
WHY HIGH DEFINITION TELEVISION?
The people, such as Francis Coppola, that have been calling for a HDTV system see it as a way of combining the advantages of both film and video production.
Videotapes advantages generally deal with lower cost and ease of manipulation of the image. These advantages include:
Instant playback. A take can be played back to determine if it is correct technically or if the performance is right. The advantage of this over video-tap playback is that the original recording is actually being checked, not a secondary recording off the viewfinder.
Ease and speed of editing. This is a complex subject. Using just about any modern TV editing system, an editor can make a first cut of a show much faster than he could with scissors, tape and film. But the first cut of a high-quality program is never used.
With film, recutting is a matter of trimming, taking out shots, splicing in new shots etc. Because this is a physical process it is impossible to do with videotape. When videotape is edited it is not cut. Instead each cut or shot is re-recorded on a new piece of tape, one right after another. If just one frame needs to be cut from a shows first minute, then everything after that new edit must be re-recorded.
On most TV editing machines this re-editing process is manual and therefore takes much longer then film re-editing. On computer editors the re-recording of already edited material is done by a computer. But these machines are very expensive to rent. ($300-500 an hour)
Ease of Manipulation of the Image. I believe this to be the major advantage of shooting on tape. Images can be composited, matted, color corrected, etc. in real time. There is no waiting for processing because the laboratory step is cut out all together. This lack of lag time between execution and result is particularly advantageous for special effects work. Live action and effects can be composited simultaneously.
Cost of Recording Medium. This is an important advantage; yet over-rated for major productions. If it were the only advantage to shooting on videotape it wouldnt be worth it. Film and processing are a small percentage of the budget of a feature production.
A one-hour reel of one-inch videotape runs about $853. This is far less than an equivalent amount of film and processing, especially if negative stock is used. So it may seem that small productions will benefit from shooting on tape. However, as I will detail later, the economy of tape can be erased by equipment costs.
Films advantages are perceived as relating to its image quality:
High Resolution. Film resolution is much greater than that of any of the television broadcast standards.
Wide Contrast/luminance Range. To achieve the appearance of a luminance range comparable to film, present television systems require much more careful and precise lighting. Essentially, the gray sc
ale for television is greatly compressed. What may look like flat lighting on film can look low key on television.
Portability. This is an area that film cameras have been far ahead of video cameras. Though the video cameras are now much smaller than in the past they still must be connected by a cable to a recorder. But this is changing rapidly. Sony, RCA and Panasonic have all demonstrated combination camera/recorders that give resolution almost as good as one-inch studio recorders with a package that weighs between 17 and 22 lbs. Yet, a 16mm camera still delivers a better image.
ESTABLISHING GOALS FOR HDTV STANDARDS
The goal then of HDTV researchers is to develop a television system that produces the high image quality of film. In order to do so, it must be determined what is meant by HDTV and what will necessary to achieve it. It would be just as foolish and wasteful to invent and adopt a system that goes beyond our ability to perceive its quality as it would be to come up with one that falls short of satisfying our needs.
Another point to keep in mind is the problem of compatibility. Compatibility of high definition broadcast signals with current NTSC receivers is a question that must be addressed. However, this concern really only applies to the actual broadcasting of the signal. For both feature film and television production, there is no need to lock ourselves into a standard tailored to broadcasting limitations. The HDTV signal can be downcoverted electronically or transferred to film.
For production use many different systems can compete to fulfill different purposes. As the technology evolves new systems will be able to easily replace outdated ones.
There seem to be two national efforts in progress to determine what the basics of a HDTV system should be.
Japanese Design Goals.4
The Japanese realized early on in their research that purely subjective comparisons between film and videotape generated images would be fraught with peril. How much of films vaunted superiority is actually a result of the psychological impact of the size of the screen?
On the other hand, purely objective measurements are superfluous if they go beyond the eyes perceptions. So what the researchers at NHK (Japan Broadcasting Co.) did was to set up subjective evaluations that minimized extraneous factors such as image size and viewing angle.
One of the interesting things they discovered was that people viewing a monitor and a projection television at the same viewing angle preferred the look of the projector. Yet the video projector had poorer resolution than the monitor. A large image viewed from a distance is more impressive than a small one at a close distance.
A corollary to the area and distance of the image is its aspect ratio. The NHK researchers found a definite preference for a greater image ratio of 5:3 or 2:1. (See fig. 1)
FIGURE 1 HERE.
Comparing film and television resolution is another tricky matter. Individual frames from film and videotape cannot be compared side by side. Since most television systems use a 30 fps rate they can squeeze in six more frames of picture information than a movie film at the same standard sound speed can.
TV scanning lines are not the same as lines of resolution. The number of lines of vertical resolution can never be as great as the number of scan lines. Forty-one of the 525 lines are used in the vertical hold bar, the black bar you see when the vertical hold is out of adjustment. And, because of various interference effects the full number of picture lines will not resolve detail to their theoretical limit, although excellent equipment adjusted well can get close.5
Using a high definition monochrome television system, with variable line rate and a 27 monitor, the Japanese made comparative evaluations at a distance of three times the picture height from the monitor. They found that after 1600 scanning lines, no improvement could be discerned.
FIGURE 2 HERE
To compare the sharpness of film and an 1125 line television the NHK scientists printed simulation transparencies on a facsimile printer that gave a line structure equivalent to that of an 1125 line television system. The result was that the TV image appeared equal to both the facsimile slide and a 35mm cine print, although not as good as a 35mm original slide.
In tests of total picture quality using slides simulating various line rates, the Japanese again found that a plateau was reached at 1600 scanning lines. (This was on a .8 meter by 2-meter screen, viewed at a distance of 2 meters.)
In all they concluded that the sharpness of a 1500 line system would be comparable to a 35mm slide and a 1000 line system would be superior to 35mm movie films.
Notice that they are talking here about sharpness and image quality and not resolution. There is no question that film has greater resolution that any presently practical television system that can record of live action. (Although I think film chauvinists often cheat by using original camera material instead of release prints for their comparisons.)
The reason for this is probably that television resolution is bandwidth limited while film resolution is grain limited. The resolution of film drops off gradually as it reaches its limit but television resolution doesnt drop off until just before it reaches the limits of its bandwidth.6
American Design Goals.7 In the late seventies, the Society of Motion Picture and Television Engineers formed a study group on HDTV. The purpose of this group was to examine the progress of HDTV and make recommendations for further research. No actual research was done by the group although some of its members have been involved in HDTV work.
Basically they set up parameters for what a HDTV system should be, both for film production and broadcast.
The premises of their study were:
FIGURE 3 HERE
AN INTERNATIONAL STANDARD
Today there are four incompatible television systems: NTSC, PAL, PAL-M and SECAM. There is a definite need for a worldwide standard. HDTV presents an opportunity to obtain such a standard that also incorporates the stereo audio and digital TV proposals now being made. If this does come about though, it is doubtful that the final system will be suitable for film production. There will simply be too many economic and political issues involved causing inevitable compromises. Since such a standard is not necessary for film production we should not hesitate to make use of HDTV when it becomes available simply because standards are not yet set. A system that is used on just one film would work fine as long as the end product is release print projectable anywhere.
PROGRESS TOWARDS HDTV
Japan. The Japanese are the only ones to have made any progress towards HDTV. This is understandable as they have devoted many years and millions of dollars to its development. They have in fact come up with a system that meets the SMPTE criteria for an HDTV system.
There is really nothing new about HDTV. Various systems have been around for years. The problem has been to come up with a system that can record moving images in color. Most of the systems have been still picture systems. This is true of both the RCA and CBS 10,000 line systems. Getting a high enough frame rate to record smooth motion is difficult. When it is achieved, something else usually suffers, such as color or resolution. Several monochrome systems exist or have existed. For example RCA had an 1870 line, 20 frame/three-to-one interlace system with a 60 MHz bandwidth. But the tube was four and a half inches wide and no one knew in 1975, or know now, how to record a 60 MHz signal.8
The researchers at NHK, Sony and Panasonic have broken through these barriers. Their first accomplishment was building practically sized color cameras and monitors. The camera utilizes three one-inch diode gun impregnated cathode saticon tubes. These tubes tubes have a limiting resolution of 1600 lines at the center and suffer from a very small lag.9
A wide variety of monitors and projectors for HDTV have also been invented. The monitors are capable of delivering the full effect of the HDTV image but the projectors lag somewhat behind.10 Fig. 4 compares the shadow masks of NTSC monitors and the mask of the NHK monitors.11
FIGURE 4 HERE
The monitors range in size from 22 to 30 and are in the 5:3 aspect ratio. The HNK researchers believe that the future of HDTV large screens lies in flat panel gas discharge displays, which they are experimenting with.
Using these monitors they have achieved a resolution of 800 TV lines vertical by 1000 TV lines horizontal.12
But the amazing achievement is the invention of a videotape recorder for HDTV.13 NHK has developed a recorder that records on one-inch videotape by doubling the rotational speed of the recording head and quadrupling the tape speed. The luminance and chrominance information is recorded on completely separate channels. A ten-inch reel of tape whizzes through the machine in 221/2 minutes.
Sony has improved on the machine by slowing the tape speed to the point where a reel of tape can record 48 minutes of material. This is of course quite adequate for production purposes. In addition, this new machine can be used in an editing system to make edited tapes. A few short experimental productions have been done with the system here in the United States.14 Francis Coppolas Zoetrope Studios produced two six minute shows, Glen Larson Productions shot segments of the TV show The Fall Guy side by side with the film cameras and CBS shot some clips of a football game and the Rose Bowl Parade.
Reports on the results were all enthusiastic, although some had reservations. Glen Larson called the look of the productions quite remarkable and said it is a totally new experienceas different as color television is from black-and-white television.
Mike Young, engineer in charge of HDTV at Glen Larson Productions, is more specific. He says, As a replacement for what were doing now (35mm film) it looks pretty hopeful. He adds that the cameras and lenses need to be made smaller and lighter so the camera can get in small rooms, fit on steadicam, etc. He thinks though that the technology is there; it just needs to be put in smaller boxes. With some tinkering he believes that you wont be able to tell the difference between HDTV and film.
Michael Lehman,15 director of Electronic Cinema research at Zoetrope agreed that the images produced by the HDTV system are almost the equal of film, and that the size of the cameras is a problem for now. They are big, monster studio cameras.
Kenneth Holland, president of Image Transform, and the inventor of an American HDTV system, was impressed by the ability of the NHK camera to handle highlights. He said it was able to show details in chrome bumpers, one of the traditional problem areas of television cameras. But, he said that the shadow detail needed improvement. The problem, he says, is that pickup tubes have an inherently linear response, and need special pre-amps to imitate the gamma sensititivity curve of film. He thinks though that the system is capable of a much higher dynamic range if the pickup tubes are improved.16
In the commotion over the Japanese HDTV achievements, the fact that there is an American HDTV system has been overlooked. This is the Imagevision system, a proprietary process developed by Kenneth Holland at Image Transform. Though it is not nearly as good as the NHK system it is far better than the NTSC standard and it can be and is used now.
The system is primarily an upgrade of the NTSC standard and uses the same cameras, switchers, monitors etc. that are used today. A new synch generator is added on to the camera that increases the line rate to 655 and decreases the frame/field rate to 24/48. Then the color encoding system is replaced by one that is much better than the one employed by NTSC or PAL systems. It does not introduce as many artifacts such as dot crawl into the color video image. This is the part of the system that is patented and enables it to work as well as it does even when the standard 4.25 MHz bandwidth is used.
When it is used over a ten MHz bandwidth, as it is designed to do, it can give 600 lines of vertical resolution and 800 lines of horizontal resolution. This compares to the 330 lines of vertical resolution of the NTSC signal at its best.
The only special equipment required beyond the converters are the recorders. These are Type B one-inch recorders with heads modified to spin twice as fast. This is necessary to record the ten MHz signal. Otherwise, all standard equipment can be used through the add-on converters.
Two productions have used the Imagevision systemMonty Python at the Hollywood Bowl and the TV series Greatest American Hero. The Python film will be intermixed with other tape to film transfers and straight film clips, so it should make an interesting comparison test.
The Hero show only uses Imagevision for special effects, specifically Chromakey mattes. These are not done with the 10 MHz bandwidth. They are then transferred to film and cut in with film shot material. I have seen the show and the mattes are excellent, though stylistically sloppy, and I could not see any difference on the TV screen between the tape original material and the film original material.
(The tape-to-film transfers are done on ImageTransforms electron beam film recorder. An interesting feature of this machine is that it automatically makes archival red-blue-green records on black-and-white negative film.)
Holland does not see Imagevision as an economical system for single-camera productions. He claims it increases the costs over normal TV production by 30 to 50%. He sees its greatest use in Multi-Camera productions, (Concerts, live events, etc.) that will be released theatrically and in special effects for television. In these instances he believes considerable swings can be achieved over film camera productions.
FUTURE AND CONCLUSIONS.
The Effect on Production.
With a little more work, the Japanese should be able to get their HDTV system to the point where it is at least as good as film. But I cannot see why Ken Hollands evaluation of Imagevisions economics would not apply ten times over to the NHK system.
At present, broadcast quality cameras run between 30 and 90 thousand dollars. I would guess that a quarter of a million dollars to be the price of a HDTV camera. The same figures apply to recorders. That could mean half a million dollars for one camera and one recorder! I have no idea how much all the other components of the system would add to this.
For standard production work, I cannot see how there can be any cost savings over film, especially since color video-taps and video post-production techniques can provide almost all the same advantages at a far lower cost. HDTV can only increase crew size by adding a highly paid video engineer to the standard union crew.
And there is still the portability problem. The camera must be connected by a cable to a large recorder. Combination camera-recorders are just now appearing for use with the NTSC system. It will be years before similar (expensive) machines are developed for HDTV.
It is in the field of special effects that HDTV will make its first impact. It is not the cost of film and equipment that makes special effects expensive, it is the time and labor required to execute them.
Using television techniques, an HDTV special effect could have all its elements combined simultaneously in one pass. The effect can be viewed exactly as it will appear while it is being shot and played back instantaneously. If it doesnt work it can be done again right away.
Almost any effect that can be done on film can be duplicated on tape. There are many effects that can be done on tape by punching a few buttons that cannot even be duplicated on film (At least not practically). Digital techniques make possible even more effects, in addition to making the whole problem of multiple generations moot.
Of course some effects will remain the exclusive province of film, primarily those dealing with alteration of shutter speeds and frame rates. But television is gaining on that problem too. Eastman Kodak is now marketing a 2000 frame per second monochrome television system.
I believe that special effects houses that catch up fast enough will find HDTV to be a boon to their business while those that attempt to ignore it will suffer from the competition.
The first place we are likely to see HDTV is in the theatres on tape-to-film transfers. There is too great an investment in film projectors on the part of theatres to expect them to replace them all overnight. This is good because video projectors arent up to projecting an HDTV picture yet.
Eventually though, HDTV projectors will appear and I predict that videodiscs will prevail over tape and satellite distribution. For theatrical purposes, the indestructibility and relative operational simplicity of discs will be great advantages. Tape wears out quickly, and satellites are prone to transmission and piracy. (See Fig. 4)18
Future of Film
Film will stick around for a long, long time. As television technologies make more inroads into films turf, it uses will become more and more specialized. When extremely high resolution is required, picture makers will turn to film, as they will when they need to manipulate time or use extra-light equipment.
I say this despite the probable high costs of HDTV equipment. Those clever Japanese are sure to bring the prices down eventually, especially if they succeed in eliminating the tubes, making the cameras entirely solid state.
FIGURE 4 HERE
1Donald G. Fink, The future of High-Definition Television: Conclusion of a Report of the SMPTE Study Group on High-Definition Television, SMPTE Journal. 89, (March 1980):153.
2Mark Schubin, Video Research, Videography, 5, (August 1980):58-9.
3Comprehensive Video Supply Corporation, Catalog, 1980-1981, p. 136.
4Kozo Hayashi, Research and Development on High-Definition Television in Japan, SMPTE Journal, 90 (March 1981).
5Mark Schubin, Video Research, Videography, 5, (August 1980):58-59.
6Otto Schade, SR., Image Quality: A Comparison of Photographic and Television Systems, (Princeton; RCA Laboratories, 1975), p. 1
7Donald G. Fink, The Future of High-Definition Television: First Portion and Conclusion of a Report of the SMPTE Study Group on High Definition Television, SMPTE Journal, 89 (February-March 1980).
8Otto Schade, Sr., Image Quality: A Comparison of Photographic and Television Systems, (Princeton; RCA Laboratories, 1975), p. 18.
9Kozo Hayashi, Research and Development on High-Definition Television in Japan, SMPTE Journal, 90 (March 1981): 181.
11Takashi Fujo, Future High-Definition Television System, Symposium Record, 12th Intl. Television Symposium, (May, June 1981) P. 193.
13High Definition Television, American Cinematographer, 63, (March 1982):244.
14The Brave New World of HDTB, Broadcasting, (Feb. 1, 1982) p.82-5.
15Interview with Michael Lehman.
16Interview with Kenneth Holland.
18American Cinematographer, P. 247