by Steve Wiedemann

In March of 1998, I had created a document called "Film Formats and HDTV: A case for the Future-Proof Standard". It outlined the differences and challenges posed by film production as compared to the promise (or threat) of electronic HDTV origination. In only a few years, it seems the industry has moved from the equivalent of stone knives and bear skins of chemical film production toward a fully electronic origination, post production and distribution system.

At the time the previous article was written, there were few choices for HDTV equipment. Large HDTV studio cameras and wildly expensive HDTV recorders were the norm, if a norm existed. The only practical and affordable HDTV system coming to market was the Sony HDCam, which was a year away. Now, HDTV origination and finishing equipment is becoming commonplace. Very few segments of the production industry are questioning the future viability of HDTV and most are planning for migration of services away from chemical film systems.

A very brief HDTV History:

The original HDTV system as developed by Sony and NHK simply extended the number of scanning lines found in standard television. The line count was increased from 525 to 1125 lines and used the same frame rate and interlaced scanning method. While they were in there, they redefined the aspect ratio assumed by the equipment, adopting a 16:9 widescreen ratio as compared to a 4:3 ratio (which could be called 12:9 when you think about it).

During the late 1980's and early 1990's, the U.S. set out to develop a standardized Advanced Television system (ATV) for national deployment. During the testing process, it became apparent that analog television systems were far less promising than the emerging digital systems. From those torrential times came our current 1080i HDTV format.

Interlace in HDTV:

The term 1080i is shorthand for 1080 scan lines, 1920 horizontal pixels (picture elements) and 60 interlaced fields (30 frames) per second. This format has the highest pixel count available of the standard HDTV formats, which translates to the finest resolving power in the commercial video market. Unfortunately, the 1080i video system is a digital version of the 1125 line standard which uses interlace, a holdover of a bygone era in television. That's what the "i" stands for in "1080i". Where did the other 45 video lines go? Lets call it "blanking". The 1125 line analog standard actually had 1035 displayable lines. That number was elevated to 1080 because, with 1920 horizontal pixels, the 16:9 image structure now had square pixels. Even though virtually all material existing in video today is interlaced, there is no reason to carry this mechanism into a fully digital future.

Interlace was invented during the pioneer days of our NTSC television system. It breaks the television frame into two "fields". This reduces the impression of flicker on the television screen by doubling the speed of the vertical scan but repeating the scan twice per video frame. The screen is refreshed 60 times per second instead of only 30. The flicker would be so bad at 30 refreshes per second, the picture would be unwatchable. Clever circuit design makes the second scan mesh with the first, like teeth on a gear. The apparent number of television lines is the same even though the picture is made of two images meshed together, each with only half of the vertical resolution. High resolution detail was not one of the available realities of early television and interlace solved more problems than it created. With today's equipment, however, the artifacts of interlace can be painfully obvious under some conditions, even in HDTV.

For the viewer, an interlaced picture will create its own artifacts, especially when anything is moving vertically in the frame. There will be a herringbone pattern on near horizontal edges moving up or down the frame. As with interlace in standard resolution television, you must consider the vertical resolution to be only half of the 1080 lines in HDTV. Each video field is really a standalone 540 line image. Combining two interlaced fields requires a softening of vertical detail to keep them from flickering in the presence of highly detailed objects. The full resolution of the video frame is only valid as long as everything is stationary. As soon as something moves, the presence of two individual reduced resolution images, each shifted 1/60th of a second behind the other, will make themselves known. What happens is the two independent fields, which actually refresh at 30hz each, start displaying the flicker of a slow refresh rate that interlaced video tries to mask. These compromises are a strike against the 1080i HDTV format I was speaking of, enough to consider 35mm film the the better choice for theatrical projection over 1080i HDTV.

Some Other Advantages of Film:

Motion Picture Film has none of these artifacts as each film frame is exposed all at once. There is no pixel structure in film like that found on a digitally originated video image. A video image is created with a grid of microscopic photosensors, which, under some conditions, may call attention to itself as a fixed pattern of noise overlaying the video image. Instead of pixels arranged in rows and columns, the individual, randomly placed grains of the film emulsion serve as the smallest area of exposure. With film, a new grain structure appears with each new frame, which randomizes the effective image structure and diffuses the effects of the underlying medium.

Film is scalable and transportable. A production can be released in any other film or video format because its frame rate meshes very nicely (24 frames/sec) with American and European television standards. The Europeans simply run the film 4 percent faster (25 fps) to print it frame for frame into their PAL television standard. A 30 fps American video production (Standard or HDTV) scales less gracefully into a 25 fps European release, so the Europeans always preferred 24 fps film coming from this country. Because of this, film still held an edge over the 1080i video format. With the uncertainty of video formats in the past decade, film was a logical choice to carry high quality images into whatever future video system came along.

Forget Everything You Know:

All of that changed in January 2000 when Sony and others released equipment built to the 1080p/24 video standard. This uses the same pixel count as 1080i (1920 H by 1080 V) but the image is not interlaced. It uses progressive scanning (the "p" in "1080p") which captures a full frame image much like film, plus this video format runs at 24 fps like film. The resultant image is stunning and can easily outperform 35mm film in a variety of areas.

George Lucas used a prototype Sony-Panavision 1080p/24 camera to shoot several scenes in the "Phantom Menace" Star Wars feature. Nobody is saying which scenes they are, but people in the know will tell you not to look for video artifacts, look for the really nice scenes. The video scenes were printed to film and intercut with the rest of the film material. The success of that test drove the decision to shoot the next two installments of the Star Wars series completely in the 1080p/24 HDTV format, not in film.

During the production of Star Wars Episode II, principle photography was well ahead of schedule due to the easy setup and flexibility of the HDTV 24p camera system. Tape stock is also much cheaper than film, which gives rise to a dramatic change in capturing scenes. Instead of wasting time on the set starting and stopping the film cameras, the inclination is to just let the tape run and let the actors act. The economics of shooting on tape come into play and multiply when removing the inefficiencies of resetting film equipment for each shot. For the total production cycle, the time to screen is shorter, less expensive and it looks better when compared to 35mm film. Star Wars Episode II was also the first feature film with a complete digital delivery chain to the home through a DVD release. From the glass of the camera to the home viewer's screen, everything was digital. Now, if we can just get the HD/DVD standard locked down...

Special Effects:

The special effects seen in feature films today are, for the most part, created electronically. Film originated material is scanned into digital workstations and manipulated using computer animation and artificial image techniques. Fully rendered and animated 3D objects can be seamlessly integrated with live action material. This is being done so successfully as to potentially upend the established domain of live screen actors and replace them with convincing computer generated characters. The potential is certainly there and is shaking a few foundations, but that's another discussion.

The advantage of using HDTV originated footage is the elimination of the scanning step and the subsequent printing step moving to and from film. The HDTV signal is already digital and may be imported and exported directly to and from the digital effects environment.

There is still a place for capturing some special effects using physical models, such as the simplicity of blowing up a model space ship rather than artificially creating the complex organic interactions of an explosion. The decision against using an HDTV camera for model photography is the ability to overcrank the film camera to produce slow motion effects. However, future HDTV equipment will be able to do this natively and may currently be simulated in off the shelf software. Motion prediction and image tracking software to create inbetween frames produces astonishing results smoothly simulating hundreds of frames per second.

Projection:

On the projection side, several theaters across the US were outfitted to show Star Wars Episode I with a digital video projector, not from a film print. This was a noble experiment, called "E-Cinema", watched very closely by all of the major Hollywood feature film houses. People (engineers) I've talked to who would run from theater to theater watching successive showings of Star Wars alternately in projected film and video preferred the video overall. The film was relatively fuzzy, had plenty of gate weave, dirt jumping around, scratches and other detractions compared to the video projection. Of course nobody noticed those things on the film until they compared it to the 1080p/24 video. Video projector technology has a few more steps to take for increased brightness and improved resolution (the video projector was only capable of 1280 horizontal pixels), but it at least equaled if not exceeded most peoples wildest expectations.

At the same time Star Wars Episode I was released, another feature, "An Ideal Husband" from Miramax, was also running in a few places with the same technology but nobody noticed. :) The equipment was a mix of Hughes-JVC and Texas Instruments projectors fed by hard drive based Pluto HD video recorders. As of NAB 2002, there were approximately 20 permanent digital theater projector installations in the US with plans to outfit several thousand more over the next five years.

The Demise of Film?:

This galvanizes the position of the 1080p/24 HDTV video format as a direct replacement for 35mm film production. There seems to be enough detail to please the eye without calling attention to the fixed pixel pattern. There were some very loud comments on how bad the HDTV projection of Star Wars looked. For one thing, most of the DLP projectors used at the time were only capable of 1280 horizontal pixels, far short of the 1920 available. To be really correct, the HDCam format only captures 1440 horizontal pixels and interpolates up to 1920, so you cold argue the creation of scaling artifacts in the process destroyed some of the resolution. Also, many of the film proponents were sitting close to the front of the theater in an effort to see how bad HD looks. As with any theater setting, you should sit no closer than three time screen height. There's a reason for that. Fear not... it is only a matter of time before the full 1920 native pixels are captured and displayed.

So far, the Europeans are embracing 1080p/24 (shock of shocks). I have not heard them express the need to modify the system to run at 25 fps. They are (so far) happy to work with the 4 percent speed increase for video, as they do now with film. Like here in the U.S., they can cram in a little more commercial time by picking up a few minutes every hour. In the theater they just run it at 24 fps anyway. This may be the first universal video standard which transcends continental boundaries, an area previously reserved for film.

With the low 24 fps frame rate, 1080p/24 would not be suited for fast action sports photography. For features, you will still have to follow the panning speed tables of 24 fps film so as not to induce large amounts of judder on the big screen. But the 1080p/24 system has no new visible artifacts when compared to 35mm theatrical film and eliminates many of the problems associated with film. At the 24 fps frame rate, the 1080p/24 HDTV system takes on the "veil of separation" that film has. This is appealing to feature film cinematographers who rely on that veil to suspend belief. At higher frame rates, the viewer experiences a telepresence that would trash the mood 24 fps film provides, a primary blow against 30 frame/60 field television acquisition.

Field acquisition using any HDTV format is especially thrilling for the nature photographers. How often have they shot thousands of feet of film waiting for the leopard to drop out of the tree only to be changing magazines when it catches the gazelle? Now, they can shoot with abandon and reshoot on the same stock if nothing is worth keeping. Think of what they have to haul around if it was all film. Think of the costs involved. A 400 foot load of 16mm color film is around $100 and runs about 12 minutes. A 40 minute HDTV tape is $65, and you don't have to process it. At 24p video speeds, the same tape load runs 50 minutes.

Intercutting Film and 1080p/24:

In this transitional time (which will probably last 20 years), it will be necessary to mix visual formats, especially when the use of available historical material must be intercut with newly produced footage. Film, especially 35mm, will intercut nicely with 1080p/24 video. Super 16mm is less desirable than 35mm but can make an excellent image if transferred on a wildly expensive HD Telecine. Relative to 1080p/24, the 1080i/30 format looks a little too "video" like to make seamless intercutting possible.

The Long Run:

The collection of original images has some very new tools available and the producer must balance several factors when choosing the originating format. One factor is the long term asset value of the images you collect. In the future, HDTV interlaced images running at 60 fields per second may be less desirable than Super 16mm film. Also, video tape formats swarm and multiply only to disappear within a decade. Photographic plates of the Civil War still make excellent reproductions but a 20 year old video tape may, or may not, play in a machine - if you can find a machine. I'm certain Matthew Brady doesn't care right now, but we appreciate the quality of his work today. However, film has long term stability issues but is generally usable further into the future than any current video tape format

It's important to separate the concept of a tape format from the actual video data format being recorded. For example, the moniker of "D1" video was applied to almost anything having to do with component digital, even though D1 is a tape recording format and the video associated with it is a standard in its own right. Component digital video may be stored in a variety of ways with D1 only being one of them. The same holds true for HDTV formats. The important part of the image format is the structure of the imaging elements, the scanning method and the frame rate. The type of recorder used to store the video data is only important to the extent that it faithfully reproduces the original image.

If a video project requires a long shelf life, consider taking any video originated images and storing them on a digital media type other than video tape. This includes finished programs as well as the outs - they may have archival value. Some data recorders project a life expectancy of the media exceeding 100 years. A proper data archiving facility will test for media integrity and perform a lossless transfer of the material to new media if the original media has degraded.

Any of the high quality formats, be it film or a flavor of HDTV, will create images for which you won't have to apologize. The selection of your originating format will, however, become a factor later when time has erased the memory of today's conventional wisdom. Please choose wisely.

E-Cinema Network:

The future Sony saw for HDTV and E-Cinema in the early 1980's is coming true. Way back then, Sony had developed second generation HDTV production tools, had bought a Hollywood feature film company, had bought a chain of movie theaters and had purchased a satellite from Hughes. What they were planning was to create feature films using HDTV and distribute them electronically to their own theaters via satellite. This is exactly what everyone in Hollywood is weighing. It would eliminate the cost of 5,000 release prints, eliminate stolen films that seem to end up on the Taiwanese pirate market, eliminate theater operators underreporting how many times a film was run and will give them complete quality control from end to end. The only thing missing 20 years ago was a viable video projector. Now, that final puzzle piece is emerging very nicely.

Currently, we have the equipment needed to shoot and edit 1080p/24 feature films. Already, we have 1080 HDTV field equipment that we rent to our production clients and we operate HDTV editing facilities containing large screen projection. The latest cameras and recorders form Sony are format agile between 1080/30i and 1080/24p, so you can choose your production path. Incidentally, Sony and other manufacturers are developing 1080p/30 and 1080p/60 systems to be released within the next three years.

Conclusions:

Film is not a bad choice for documentary or theatrical releases but it is rapidly becoming a technology eclipsed by the advances of HDTV video. Couple that with my belief that video will emerge as a common, on-demand service in the next five years, the role of film may well be a thing of the past. Digital archiving of digitally acquired images and delivery from servers on demand has no room for film.

If you must shoot today, shooting on 35mm film is still not a bad choice given the systems in place to produce and distribute using film. However, give strong consideration to the 1080p/24 system. Some of the clients we have introduced to this method of production have deleted film from their radar screen almost completely. Even though this flies in the face of some of our existing business units that rely on chemical film production and transfer, as a company, we must adapt our business units to embrace and extend new technology developments.

The writing is on the wall for film in general and we will choose to adapt to the changing environment. Equipment and techniques change very rapidly but our primary asset in this area is our talent pool. There is no better background to approach all electronic feature production than to have a deep understanding of film. For that reason, we don't see HDTV feature production as a threat, we see ourselves as an asset for any producer who approaches us.

Steve Wiedemann
Sr.VP, Director of Technology
Henninger Media Services
703.908.4018
http://www.henninger.com

Also of interest: Film Formats and HDTV: A case for the Future-Proof Standard.