Colorist Dave Cole Opens Up The Box

If you’ve followed director Richard Kelly’s career from Donnie Darko through Southland Tales, you know that his films have an otherworldly look that goes beyond their sci-fi story settings. The unique atmosphere of his films is created in large part through collaboration with cinematographer Steven B. Poster, who puts the pictures on just the right, dreamlike wavelength to match Kelly’s fantastic yarns. Their latest, The Box, is one of those incongruities in contemporary film: a period piece shot in HD, using the Panavision Genesis. F&V spoke to the film’s DI colorist at LaserPacific, Dave Cole, about the work that went on behind the scenes to make sure the filmmakers had exactly the right perspective on their footage. The result in the final film-out is a soft, ’70s look with a modern twist.

FILM & VIDEO: How do you start preparing to make a Genesis shoot look the best it can?

Dave Cole: We started with pre-production meetings at Panavision, discussing the technology. Some tests were shot in Boston [where location filming was taking place], and then the DP [Steven Poster] and DIT [Alan Gitlin] came and sat with us. We did a whole lot of looks to test the cameras. From those looks, we built four primary look-up tables [LUTs]: for daylight, tungsten, fluorescent, and sodium vapor.

When production started, they sent us the tapes. Steve [Poster] asked me to look at certain shots and mess around with look development. I would load it up in the Lustre and create looks in the film that were sent back to Boston, where they’d look at it in projection on a seven-foot screen in the trailer. They cut a little promo video as a proof of concept, and then I did a full-on grade in Lustre and took that with me to Boston.

I was there for three days, and I watched that promo to make sure their projection system and on-set monitoring were calibrated correctly. Alan created separate tapes for the production, so they’d shoot daylight scenes on one tape, fluorescent on a different tape, and so on. They’d stick the tape in and play the whole thing through the same look-up table. So these four LUTs allowed them to watch dailies that looked really good. And then we were all on the same page.

I came back to L.A. and did bits and pieces of look-development with Steven, who was continually sending me photos from the set [digital stills worked up with sample looks in Photoshop] with ideas.

What about previews?

We had three days [to grade] for the first preview and a couple of days for the next preview. The advantage of working with the Genesis, and digital acquisition in general, is you’re not color-correcting the dailies and then starting over again with the film scan. You do have to set the look fast in previews, so you can’t get the nuance you get in the final room, but we were obviously heading in the right direction. And when it came to working on the final film, we’d use the mattes and other things we got back from Gradient Effects. We can draw a shape in Lustre and do all that fancy stuff, but Gradient would create mattes for certain effects and we could reuse some of their work. That really is a great thing the Lustre allows us to do. We continued to grade, shot it out to film, and printed at Technicolor.

Had you worked with Genesis-acquired projects before?

I’ve set up other projects where I didn’t personally do the DI, but I’ve been involved with a handful of Genesis projects and this was my third actual grading of one.

It seems like the use of different cameras adds another degree of complexity to your job.

A lot of it is pre-visualizing on set, and emulating that print. There’s no point in the filmmakers seeing something that’s never going to be a reality later on. It’s emulating in the Rec. 709 world, which the monitors are set up for. Every camera has its own character, and in that way it’s the same as film, where each film stock has its own character.

The Genesis is an HD camera, meaning the final result is a 1920×1080 image, so there are things you need to be aware of [in the DI suite]. You don’t want to zoom in too much. You might not have a lot of room to tilt the image after the fact. This one was a 2.40 aspect-ratio show, so we had some extra space where booms came in, and we could reposition for stabilization and things like that. It was really interesting, because other Genesis projects I’ve worked on were 1.85, and at 2.40 we were slicing through the middle of the image. So how would that resolution look on film? As it turned out, we really loved the look of it.

So your final image isn’t necessarily compromised just because of a lack of overall resolution.

Not as long as you’re aware of what you’re trying to do. You can’t just run and gun. If you’ve got someone like Steven as your cinematographer, you’re going to be in a good place. Steven tested the camera in terms of overexposure and underexposure, learning where he could push it and where he couldn’t. We worked out the effective shooting range of the camera, and he rated where he wanted to shoot it. Like any show, the more tests you do, the better results you get. And Steven really did that on this film.

What about the film’s period look?

The movie is set in 1976, so even though we’re using the newest camera technology, Steven used the equivalent filters that he would have used in that era. We went through tests of using Pro-Mists and Glimmerglass and low-cons and double fogs. We came to the conclusion that he could use Pro-Mists and enhance them a little bit more in the DI. The result, hopefully, is this interesting look that feels 1970s but has a modern aesthetic about it, a quite unique vibe.

So you were combining in-camera looks with DI image manipulation.

It was both. We wanted to make sure the image going into the camera was in the ballpark, and then we enhanced it after. We didn’t want to go purely in camera, and we didn’t want to go purely digital.

I understand that you’ve cooked up some custom plug-ins that streamline your workflow for certain tasks in the Lustre.

I’ve been a colorist for 14 years, but before that my university degree was computer science, so I can program. That’s just a happy luxury for me. It means we can enhance Lustre through its plug-in API to create look filters as needed on a job-by-job basis. If the Lustre didn’t support this, every Lustre in the world would be the same. But, by having the capability to add to the pipeline and the toolset, we can have something unique that’s not available everywhere else. On the previous couple of films I did with Steven, I had written two or three specific plug-ins, and of course I’m reusing them.

What kind of tasks do they handle?

Something Steven detests is light switches or power outlets in the middle of a nice wall. I wrote a plug-in to remove elements like that. We didn’t have many shots in The Box where something was out of place in terms of period – there weren’t any satellite dishes in this. But, as an example, I just did some special promo music videos for The Box, which are online at youaretheexperiment.com, and a wide shot of Langley Air Force Base had some modern cars in it. I removed those cars and replaced them with bushes. We do elementary visual effects work in the DI suite. There will be times when it’s more efficient to farm it out to VFX, but very now and then I can do something in two minutes and it’s done.

Does the capability of the Lustre continue to expand with each new release?

It’s really starting to push forward. We’ve had a couple of versions [of Lustre] with a stereo toolset, but now we’ve got advanced grading functionality for moving around pre-set grades between projects and reels. There’s a really great tracker in there. And, obviously, as the hardware improves it benefits the software application. The new Lustres we’re on are four times faster than what I was using on The Box. It’s continually benefiting from Moore’s Law, and the application itself is becoming more powerful. Nothing can touch it in terms of shape manipulation. That’s one of the things I’m really good at, and Steven very much uses it as a tool in his arsenal.

What about advances in sheer resolution? Are we getting any closer to moving from 2K to 4K as a de facto standard for mastering, and does it have clear implications for image quality?

We are obviously moving there. The biggest things holding us back are time, speed and money. Say we were working in 4K from a film scan. That 4K scan is four times larger than a 2K. In terms of space on hard drives and a storage area network, it needs four times the capacity and you have to be able to pipe through networks to render it and play it, all at four times the size. It’s not that it can’t be done. It just uses more resources: time and money. That’s why jobs have been done in 4K, but they’re not as prolific as they could be. There is a tax to doing 4K. However, as we move forward in technology everything gets faster and space gets cheaper. We’ll be moving more and more toward doing that.

Having said that, the resolution of the Genesis is amazing. It’s not a native 1920 [resolution] because it shoots a striped array. But it is a really sharp camera. It looks sharper than a supposed 4K digital camera. So numbers aren’t everything. When you say “4K,” 4K doesn’t necessarily mean 4K. You can have that many picture elements, but if you have 4K, is that 4K of RGB or is that 4K of just one channel and not of the others because it’s a Bayer pattern and then you have to interpolate pixels? There are lots of different ways of seeing numbers, and bigger is not necessarily better.