By Greg Pak
After years of pushing feature film projects which needed budgets of millions, I wised up and concentrated on writing a project which I could shoot on digital video with the smaller chunk of change I could raise from private investors. The result was “Robot Stories,” an anthology of four stories about love, death, and robots. We shot on digital video, with the intention of blowing up to 35mm film. Here’s the skinny on our tape-to-film strategies and experiences:
Low budget filmmakers shoot on video largely because it’s a tenth to a hundredth of the cost of shooting on film. But film has better latitude and depth and color – and for some ineffable combination of reasons, movies just feel like movies when they’re projected on film. So we can’t afford to shoot on film. But transferring our video to film can give it some of that filmic glory.
The format and equipment
We shot “Robot Stories” on PAL DVCAM with the Sony DSR-500 in 16×9.
What’s that mean?
PAL is the European video standard – we chose it over NTSC, the US standard, because PAL has about 20% more lines of horizontal resolution. The extra pixels in those lines improve picture quality noticeably – particularly in reducing typical elements of video ugliness, like jagged breakup along hard edges.
DVCAM ranks in the middle of digital video formats – better than MiniDV, not as sweet as HD. We couldn’t afford HD, or High Definition video – even if we got the camera for free, the editing costs would have killed us. On the other hand, I’d made a number of very successful short films over the years using MiniDV camcorders like the Sony VX-1000. The image quality was crisp and clear, but also clearly video. These were short comedies – two were actually porn spoofs – so the specific video look wasn’t a problem. But “Robot Stories” was no porn spoof. And in all the MiniDV features I’d seen blown up to 35mm, even when lighting had been carefully controlled to avoid blown-out white spots (for which video is notorious), tell-tale video-style jagged edges remained. DVCAM, specifically DVCAM shot with the Sony DSR-500, proved much better.
The Sony DSR 500 is a great professional video camcorder – larger and equipped with a much better lens than prosumer camcorders like the PD-150. We ran tests and were thrilled with the fineness of the images and the camera’s latitude – the camera can record a fairly wide range of darkness and lightness, meaning that with careful lighting, we could shoot dark scenes and still retain details in the shadows and we could have some bright areas in the frame which wouldn’t entirely blow out.
Finally, shooting 16×9 means we wanted our final image to be widescreen – on the DVD, you’ll see the film letterboxed, with black bands at the top and bottom of your screen. But many camcorders with 16×9 settings merely mask the top and bottom of the frame, meaning you lose pixels and image quality. Since we were transferring to film, we needed all the picture quality we could get. We needed a camera which could shoot true 16×9, meaning it would record images on the entire frame. On a non-16×9 monitor, the image would thus appear stretched – people look tall. But on a 16×9 monitor, the images are squashed to the proper ratio – but retain all the pixels of the stretched image.
Point of reference: Our choice of format and camera – 16×9 PAL, Sony DSR-500 – was the same used by Jennifer Jason Leigh & Co. to shoot their DV feature “The Anniversary Party.”
The Shooting Strategy
From years of making MiniDV short films and documentaries, my cinematographer Peter Olsen and I knew the pitfalls of video. Here are the dangers and how we countered them:
- Bad latitude. Film does a beautiful job of rendering very dark areas and very bright areas without losing information. Video is less adept at this – one of the most common and ugly features of DV films are blown-out windows and skies. The brightness turns pure white – it’s a void which I think can subconsciously shut doors in a viewer’s mind. So Pete assiduously avoided overly bright spots – putting gels on windows to bring down the brightness, working with the costume designer and production designer to avoid white clothing and walls.
- Murkiness and noise with underexposure. DV camcorders have an amazing ability to record images in extremely low light – when you use what’s called the “gain” range of the camera’s exposure settings, you can virtually shoot in the dark, recording things you sometimes can’t see with the naked eye. But shooting under those conditions come at a price – video images shot with high gain tend to moil a bit with video noise. We avoided the problem by not dipping into the gain settings on the camera – we shot dark scenes, but Pete lit them enough to be able to record in the normal, non-gain exposure settings.
- Jaggies and moire. Since digital images are made up of thousands of tiny, square pixels, certain kinds of patterns record poorly with digital video. The jaggies, one of the ugliest features of bad DV, comes when individual pixels seem to pop out, usually when the image includes sharp diagonal lines at high contrast. Film would render those lines cleanly – in digital video, the fineness of the line gets interrupted with tiny squares. Moire is a weird, subtle color shifting in an image – in digital video, moire can appear in fine patterns with many narrow lines or squares close together. To avoid the jaggies, we controlled exposure and contrast – overexposure greatly exaggerates the phenomenon. To avoid moire, our costume designer and production designer tested complex patterns in front of the camera and tossed what didn’t work.
- Ugliness in wide shots. When watching movies which had been shot on video and transferred to film, I noticed that close-ups usually looked fabulous but wideshots often didn’t. Wideshots tend to provide more opportunities for the pitfalls of video, including jaggies and latitude problems. So Pete and I made a conscious decision to avoid lingering on wideshots with “Robot Stories.” This was difficult – often the most cinematic and dramatically powerful choice may be to use a glorious wideshot. But “Robot Stories” is ultimately an intimate picture about personal relationships and struggles – I think our shooting strategy complements the film’s dramatic requirements.
- Video sharpness and coolness. Crisp and clean seem to be positive attributes. But strangely, the hyperreal clarity of digital video can actually be distracting in a feature film. I’m not entirely sure why – maybe that supercrisp video feel is just a bit too real – too much like televised news and sports and reality shows. We used a Black Pro-Mist filter over the camera lens to put a subtle haze into our images. And we used a chocolate filter for three of our four stories to introduce some warmness.
All of our precautions paid off – when watching dailies during the shoot, I was thrilled. We had almost no blown-out areas – all of our windows and walls retained lovely detail. And the filters gave the picture a kind of subtle organic feel – hard to explain, but you feel it when ya see it.
But the true look of the film only came together with the tape-to-film transfer.
The Tape to Film Transfer
We edited the entire film on video using Final Cut Pro 1.2.5 on a 533 MHz Dual Processor G4 Macintosh. After locking picture, we exported to video. Since we’d imported and exported pure digital video in and out of the computer, our video master had the identical picture quality as our production footage.
We then took that video master to a lab for a tape-to-tape color correction. Now the footage looked great, since Pete had done such a beautiful job, but there were still corrections to make from shot to shot and scene to scene. And we had special plans for certain flashback and virtual reality sequences. And finally, we needed to prepare the video for transfer to film.
Swiss Effects, our tape-to-film lab, gave us a sheet of specs for preparing video for the transfer to film. When we reached the tape-to-tape color correction, what we needed to remember was to keep contrast and color saturation levels lower than we might normally prefer. As Joe, our colorist at Magno explained, the idea is to bring the intensity of the image about 75 or 80 percent there – the transfer to film adds the rest. Another way to think about it is that if you pump up the contrast all the way, you lose information in your dark areas, leaving you no latitude later on down the line. This was one of the toughest parts of the whole process – I understood the plan, but when I looked at the images on the screen, they just didn’t look punchy enough to me. I had to trust what the pros were telling me, that the transfer to film would bring that punch back.
Now several advantage of shooting digitally came home to me while we were doing our tape-to-tape. First, if you shoot on film and don’t have money for digital effects, what you have on your negative is largely what you’re stuck with – if you have a corner of a frame that’s a bit too bright, there’s nothing you can do about it except make the whole scene darker when you print. But when you shoot digitally and do a tape to tape, the colorist can use the magic Power Windows feature on his or her expensive machine to lighten or darken or change colors in specific parts of a given frame. Since we were paying by the hour, we chose our Power Windows moments strategically. But it was invaluable, for example, in darkening up the edge of one scene to obscure a boom shadow.
Another advantage of the tape-to-tape color correction was the mind-blowing array of tools and effects the colorist has at his or her disposal. Just one example: at the inspired advice of my editor Stephanie Sterner, we gave the “Machine Love” segment of “Robot Stories” a subtle electric sheen by bringing up the blue in the black elements of the image. In a film negative to film print color correction, we could increase the amount of blue in the picture. But in the video color correction session, we were able to specifically increase the blue in the black undertones of the images, which made a subtle but importance difference. The images, shot in a spare, modernist office, had always felt a bit drab, a touch disappointing. But the blue undersheen worked with the blue elements we’d consciously worked into the costume and production design – suddenly everything came together.
So at the end of our tape to tape color correction, we exported the movie to tape, again retaining the pure digital video quality, and reimported it into the computer. I cut forty of the more problematic images of the film into a two minute clip which we sent to Swiss Effects for them to transfer to 35mm as a test. After ten days of nail biting, the test came back, and it was brilliant. I’d purposefully sent them our trickiest moments – the dark scenes, the high contrast moments, an instance of moire on a couch pattern. And everything looked better than we’d dared to hope. The moire pattern had even disappeared, which we had no right to expect.
So we sent the whole package off to Swiss Effects. They transferred the video to 35mm film and sent the negative to Technicolor in New York. Technicolor made our first answer print, based on notes about the colors and look we were going for.
And it was gorgeous.
A dozen folks who have seen the film told me they had no idea it was shot on video. Filmmakers and industry types who know what to look for could tell it originated on video – there are a few giveaway wide shots here and there. But folks have uniformly been impressed with the lush, organic feel of the colors and look.
I’d seen the movie projected on video many times while we were editing. And I always felt good about the image quality. But I never felt as emotionally connected with the film as I did when I saw it projected on film. I can’t explain this entirely – Gene Siskel used to talk about alpha waves and beta waves in the brain when he tried to explain why film was more involving than video. But I have other theories. First, contrast. Transferring to film gave the black parts of the images true blackness. So I found that when watching close-ups of characters, I could lose myself in their eyes. I wasn’t looking at the image; I was looking into it. Second, frame rate. We shot the movie on PAL video at 25 frames per second. But each frame of video actually consists of two fields, so we had 50 images making up each second of screen time. The result is that video tends to render movement very smoothly and cleanly. Film, on the other hand, is recorded at 24 frames per second, with a frame consisting of a single image. Fewer images making up each second may sound bad – why give up more reality? Because I think there’s something magic about that 24 frames per second frame rate. I think it stylizes movement in a perfect way, makes it unreal enough to make it feel more real than real. I can’t explain it. But I know that when I saw our movie projected on film at 24 frames per second, the movement of the characters and camera felt right – it felt like a movie.
Finally, we needed our video master of the final film. We went back to Joe at Magno and did a film-to-tape color correction, this time going from our negative from Swiss Effects back to tape. We were able to do a bit more tweaking with color. And by going from the film negative, we were able to capitalize on those subtle organic differences the transfer to film had introduced into the movie.
Drawbacks to Digital
I’m thrilled with how good “Robot Stories” looks. But I don’t pretend that it wouldn’t look and feel even better if we’d been able to shoot on 35mm film. A brief list of what we sacrificed by shooting digitally:
- The ability to exploit wide shots. As mentioned before, we avoided using too many wide shots because we knew they had a better chance of looking crappy in the end. But even though I can justify the look of the film based on the specific dramatic needs of its stories, here and there I miss the wide shots.
- Subtle pacing issues. The lab transfers 25 fps PAL video to 24 fps film frame for frame, meaning a frame of video becomes one frame of film. The result is that the material you shot 25 fps gets projected 24 fps. So your movie becomes 4 percent longer. The labs do a pitch shift on your sound, so the music and people’s voices don’t get all low and distorted. And most viewers would never notice the difference. But I think the movie was a hair funnier at 25 fps. Everyone else thinks I’m hallucinating and the film works brilliantly at 25 fps. But if I were to do it again, I’d try pumping up the pace of our funnier stories just a touch to try to compensate for the slowdown.
- Okay, I admit it: Film is prettier. It just looks better. Better than video. Better than video transferred to film. There’s a depth and glory to film which just feels better. Maybe it’s not a question of better – maybe it’s different. Video can look fantastic – certainly top of the line digital video has better color and sharpness than some crappy old films you can see. But film is organic, and maybe we just respond to it a touch more deeply because (at the moment I write this, anyway) we’re organic.
Advantages of Digital
To end on an up note, here’s the DV cheerleading section:
- It made our project possible. We simply couldn’t afford to make the movie on 35mm with our budget. Without digital video, the movie wouldn’t exist.
- I got a lot of takes. As a first time feature filmmaker shooting emotionally demanding material, I never wanted to be told we didn’t have enough film to shoot another take. The number of takes I could get of a given scene was limited by a million other factors, from traffic to rain to equipment failure to the destruction of the World Trade Center on day two of our shoot. But the fact that we shot on video meant that I was able to shoot as long as we had time.
Playback. I could see exactly what we were recording while we were recording by looking at the video monitor. And when I was acting in “Machine Love,” I could look at playback to see if we’d captured what we needed. It’s nice to know.
- Speed. I don’t make as much of this as some digital filmmakers – we paid great attention to lighting “Robot Stories” properly, which means we took as much time as we would have if it had been a 35mm shoot. But the camera and tripod were lighter than a 35mm rig would have been. And we didn’t have to mess with heavy 35mm film cans and reloading magazines every twelve minutes.
- Special effects. We had a few digital effects in the picture, which were much easier to do when all the material was digital from the beginning. If we’d shot on film, we would have had to transfer the material to video to introduce the effects and then transfer it back to film and then transfer that film to video so we could cut it back into the picture and then –I’m getting a headache just thinking about it. Suffice it to say that after a (admittedly considerable) bit of trial and error, we were able to export clips from Final Cut Pro which our special effects guys could import into their AfterEffects program and then export back out for us to use in the project.