Why Make Digital Films?

SHOOTING VIDEO FOR FILM RELEASE
by Alan Stewart

The subject is shooting video for film release, so I will begin by saying a little about why one would consider such an option. Only in the most rare situation would one shoot video and convert it to film for aesthetic reasons, so we'll start with the principal reason an independent filmmaker would shoot video... money.

Your goal is to produce entertainment in a medium that traditionally requires a lot of cash. In movie making, maybe more than in other industry, money talks. Money buys time, and creative options, or creative freedom. Obviously, money buys crew, locations, materials, actors, etc. If you have little money, you will have to make good use of the resources you do have, or that you can afford on a skimpy budget. So what do you REALLY need money for?

If you can write a good script, that won't cost you much. Maybe you know some talented actors who will perform in your film for no money, or very little. There seems to be no shortage of hairdressers, lighting technicians, camera operators and so on, who will donate some of their time to a worthy project. What really matters when making a film is what ends up in front of the camera, so let's assume you have directorial talent, a good script and can scrape together an adequate cast and crew. The rest is technical, or 'behind the camera'.

Areas that have always been difficult to negotiate around are camera equipment, film stock and lab costs. I once made a film for 'nothing' and it still cost me $35,000. I owned my own equipment and didn't pay the cast or crew, but the costs of film stock, laboratory, negative conforming, sound mixing and delivery items (film and video elements required by the distributor) added up to quite a bit. I did spend some money on transportation and food, but even if I'd shot it all in one location and let the crew starve (not recommended), I'd still have spent over $30,000. So we come to one means of circumventing some of those 'fixed' costs... shooting video.

Shooting video for a film release can be inexpensive on the front end, but is not cheap on the back end. As a native Texan it comes naturally to me to compare filmmaking to drilling for oil: first you get the money to drill the hole to see if there's anything there. If you find oil, you won't have any trouble getting the money to pump it out. So how does this apply to filmmaking? If you don't get it right off, you might want to consider some other line of work, but I'll explain briefly in case you still plan to hang in there. The 'oil' you are looking to find is the entertainment value in your film. You have to get your film in a form that will demonstrate it's entertainment potential, or in the distributors eyes, the market value. Since you can shoot hours of video at a fraction of the cost of shooting film, and as it is becoming easier/cheaper to edit video and mix sound with computers, you can put together a fairly sophisticated 90-minute film for a few hundred dollars.

Once you have the show in a presentable form on video, you can shop it around and soon find out if you are actually talented, or if your friends and relatives have been lying all along. Seriously, if your movie entertains and has market value, then you are on your way. The costs of transferring video to film and delivering the show to the market place will be covered, if you can make someone a million dollars, or at least make them think you can. "In The Company of Men" is a great example of guerrilla filmmaking... good script, good cast, few locations, no frills, technically weak - shot on 16mm, doesn't look great or sound great. But who cares? It cost something like $25,000 to put together and I saw it twice! "Rhinestone" cost millions to make and I couldn't sit through it once. Recent video-to-film successes are "The Celebration" and "Vietnam: Long Time Coming". These films entertain an audience and achieve their goals nicely, and, had the filmmakers insisted on shooting film rather than video, the films might never have been made at all.

Before I get into some of the technical considerations, I'll say a last word about aesthetics. Look at works that have been shot and/or finished on video and then transferred to film to get a sense of the 'look and feel'. If you are planning to make the next "Out of Africa" or "2001: A Space Odyssey", you may want to start raising the money to shoot film. Be sure the medium suits your material, or the other way around. Things may change in the future, but today video does not look like film. Whether you spend $40,000 or $300,000 on your video to film conversion, you will still not produce a visually beautiful film. But if you can tell a story and entertain people without the glorious images, video-to-film gives you an opportunity to do that at a reduced cost and potentially put yourself in a position to make another film anyway you want.

Terminology and Pratices:

When you enter into this arena you will need an understanding of the terms and practices used film production, video production and in the computer industry. I cannot bring the novice up to speed here, so you will want to do research and anything else possible to get yourself in a position to make the best decisions.

Resolution:

Lets look briefly at the differences between single video and film frames, or images. Video images are comprised of pixels - little dots that represent the color and luminance values of the image. The more pixels per image, or frame, the greater the detail. The visible area of a frame of NTSC video has some 307,200 pixels, and a PAL frame yields about 368,640 pixels. Film images are etched right onto the film emulsion by light and, while there is film grain, there are not pixels as such. While we see a granular pattern when viewing film, we do not necessarily see the individual silver particles that hold the image. These particles range from about 0.003mm down to about a tenth that size.

An image of 5000 x 3760 pixels is required to truly mirror the discernible quality of a 35mm film frame. When film frames are stored as a computer image file there are three common pixel dimensions used - 2K, 3K or 4K. A 2K image has 2000 pixels side-to-side, a 3K image has 3000 pixels side-to-side, and so on. The pixel dimensions in height vary depending on the aspect ratio. Working with 3K or 4K is desirable because the threshold of human vision is about 2500 x 2500 pixels when viewing an average movie theater screen (if the picture were square). 2K resolution is becoming more common because of the reduced file size and throughput requirements (data flow) - commerce winning out over art again.

Shoot:

Obviously, you would want to shoot the highest quality video possible. If you can afford to rent a High Definition package you could probably also afford to shoot film. Luckily, the quality of the mini DV cameras is excellent and at $2,500 to $5000 for a good one, they are affordable. The DV format is 5:1 compressed, but still looks pretty good. When shooting your video work with lighting ratios around 3:1. Be sure to watch focus and exposure carefully and do not use any autofocus or autoexposure features. In those regards, shoot much like you would shoot film. Don't let you blacks get crushed and try not to clip the video levels.

Edit:

It is not necessary to edit in the native DV format, but if you choose to do so, some editing systems use a Firewire I/O to 'dump' the data from the tape directly to disk. When working with your raw footage (dailies) at 5:1 compression you'll need about 13 gigabytes per hour of source footage. It might be a good idea to work with a system that allows you to further compress the video so you don't need a room full of hard drives. There are many possible workflows to choose from. Avid's products allow you to work at a high level of compression and then recapture the material at little or no compression for final output. Avid's AVR77 compression (2:1) yields a good picture direct off disk. Avid also has systems that work at D1 uncompressed video resolution (Symphony and DS). Another possibility would be to 'off-line' on an inexpensive system that would generate and EDL for a video on-line from the original tapes, or on one of Avid's high end systems.

Conversion:

When NTSC video is converted to film, about 307,200 pixels must be copied to a frame with a potential of holding something like 19,000,000 pixels, but a 2K frame would be only 3,000,000 pixels (at 1.33:1). Obviously taking any picture and trying to increase its resolution is a tricky proposition, but in the case of video-to-film one has to deal with the fact that a video frame is actually two interlaced fields (the odd lines are drawn first and then the even lines), and that there are 30 frames per second (30 frames equals 60 fields), and furthermore NTSC video does not actually run at 30 frames per second, frames are displayed at a rate of 29.97 frames per second. While film runs at 24 frames per second with no 'interlacing'. In video terms, film frames are best compared to progressive scan (like your computer monitor). I am not going to go into much detail here explaining the interlacing of video frames and the frame rate issue. Refer to my website at http://www.zerocut.com for more on this issues and other principles of digital nonlinear post production.

PAL video offers some advantages over NTSC in that there is more video resolution (368,640 pixels), it runs at 25 frames per second, and the cameras are often cheaper in the USA due to lower demand. Because the frame rate is so close to that of film, PAL video can actually be transferred to film frame-for-frame (it simply plays back 1 FPS slower in the theater). This eliminates the need for frame 'averaging' that goes on when converting NTSC 30 to 24, thus reducing motion artifacts.

There are several methods of converting video to film ranging from filming the show off a video monitor [ :-) ] to scanning every frame into a high-end computer system which then will blend the frames and up-res the frames to 3K using sophisticated filters. The latter method is the best and costs about $2.00 per frame. Do the math and weep.

There are essentially three different video-to-film transfer devices: Film Recorders, Kinescopes and Electron Beam Recorders. Each has their positive and negative aspects, including quality and (most noticeably) costs. The facility will give you back (depending on the facility) either a film negative or a print (with or without sound), or a combination of these. You need to work closely with the facility you select to fully understand their process, what they need from you and what you will get from them.

The Film Recorder:

Film Recorders that reproduce the highest quality electronic image on film, typically from digital files that were filmed live or created on computers. They boast vertical resolutions of 2000 or 4000 video lines (called 2K and 4K, respectively.) There are two basic technologies: 1) Film Cameras that shoot hi-res monochrome CRT (high quality video screens) through red, green and blue filters, and 2) red, green and blue micro-lasers that scan an image on the film's surface without the use of lenses.

CRT-based systems are made by Management Graphics (Solitaire) and Celco. These devices are slow, 15-40 seconds-per-frame for a 4K output image, and cost several dollars per frame.

Laser systems are made by Eastman Kodak (Cineon Lightning). Digital Cinema Systems (Lux) and Pthalo Systems (Verite). Laser systems are faster and brighter that CRT systems, but at a slightly higher cost starting at several dollars per frame.

To get your video transferred, you will need to send your videotape to the facility. It is best to master to the highest quality video you can afford (i.e., D-1 or DigiBeta). They digitize the tape on their computer, then feed the images a frame at a time into the film-recording machine. The film recorder will then film-capture each frame to a film negative.

The Kinescope:

Another method uses a kinescope, which is a very sharp monochrome video monitor photographed by a specially designed film camera. The kinescopes drop selected fields to produce a 24 frame-per-second film (actually 23.976 frames per second) and eliminating the visible roll bars that occur when filming TV screens.

Electron Beam Recorders:

The electron beam recording (EBR) systems fire electrons directly onto unexposed film, which is held in a vacuum. The film is black & white because electrons don't have color, only intensity. Three exposures must be made for each frame of color, one for the red, green and blue components of the image. These three separate exposures are re-printed through red, green and blue filters on an optical produce a color negative.

There are very few facilities that use EBRs. Most of the machines are now owned and operated by Four Media Company (4MC) in Burbank. 4MC's EBRs transform 16mm only. Both 16mm color negatives or 35mm color negatives are made on an optical printer, the latter via a 16mm-to-35mm blow up.

The other major player that uses EBR technology is the Sony Pictures High Definition Center. The facility opened in Culver City back in 1991 to create new methods of video-to-film transfers. They have a unique system based on non-real time, pin-registered 35mm EBR to transfer 1125-vertical line resolution HDTV images. Sony's EBR outputs three black-and-white separations per second, which equals one color film frame. Their color correction and switchers are all HDTV compliant. NTSC video resolution images are up-converted to 1125-line HDTV, the line-doubled again to approximated 2K output. The costs are high, around $7 per second of video, print only. For a 16mm output, they reduce the 35mm negative on an optical printer.

The LOOK:

The only way to truly know is to test it.

Due to the variables in equipment, lighting conditions and the specific equipment used, it is impossible to predict what will be created from a video-to-film transfer. Tests are crucial to find out what works best for you

Parting Notes:

• Get the framing right for your target release medium. Are you going to 1.66, 1.78, 1.85, or what? You need to know before you shoot.
• Know that pans and other camera moves may look odd when converted to film.
• Visual effects that look smooth on a video monitor at 29.97 FPS may not work to your liking at 24 fps.
• Pay careful attention to the audio specifications. It is probably best to record the audio apart from the video with a DAT recorder and then sync it up in post production.
• Be sure you know how much the services will cost and what you are paying for. Some facilities will not give you the negative for the quoted price, only one print.