“The Brothers have a real sink or fly daring when it comes to aesthetic choice,” says David Tattersall, BSC, in describing the unique look directors Andy and Larry Wachowski were after for their live action resurrection of the classic animated series Speed Racer.

Originally broadcast in the late 1960s and early ‘70s, Speed Racer was the English language adaptation of the Japanese anime series Mach GoGoGo. Created by anime pioneer Tatsuo Yoshida, the series was one of the first anime programs to find success in the West. The easy translation of the series for American consumption was in part due to Yoshida’s affinity for American popular culture. Inspired by Viva Las Vegas and Goldfinger, Yoshida gave his characters Western physical attributes, including Speed’s Elvis-style black hairdo and racing neckerchief as well as sports cars loaded with gadgets straight out of the James Bond franchise.

Unlike domestically produced animation of the era, Speed Racer was an action program with fast cars, fist-fights and plot lines that often involved corporate conspiracies and dark family secrets. US audiences welcomed the show eagerly and it enjoyed a long run in syndication, but in other countries, such as Germany, the program was considered too violent and pulled from broadcast.
“I wasn’t aware of the source material having grown up in England,” says Tattersall, but added that he was thrilled when producer Grant Hill called to offer him the project. “I was flattered, just to be thought of and the chance to work with the Wachowski brothers is a great opportunity. That was my main interest. I almost didn’t care what the show was about because it was a chance to work with Larry and Andy. I have a great respect for them and their work.”

Cinematographer David Tattersall, BSC and crew shot Speed Racer with Sony F23 digital camera systems.

From his initial conversations with the directors, Tattersall knew the project would break new ground visually. The Wachowskis are

best known for their dark and gritty Matrix trilogy, which forever altered the look of action movies. Since Speed Racer is a family movie, it marks something of a departure from their previous work and Tattersall said they embraced the opportunity to create something new.
“It’s quite different to what the Brothers are known for because it is aimed at a younger audience,” explains Tattersall. “It’s lighter and brighter. It’s got these outrageous racing sequences and monkey kung-fu and a lot of humor—it’s quite different from the Matrix.”

For this project, the Wachowskis wanted to explore an aesthetic that ran contrary to expectation. According to Tattersall, they wanted a look that was “super deep-focus, super color-saturated and very smooth, clean and sharp—quite different to the gritty, grainy, realistic thing that is happening in cinema at the moment.
“The Brothers are very techno-savvy fashionistas—they follow all the fashions and the trends, but they’re also sort of pioneers themselves,” Tattersall explains. “They like to sort of mix and match and there’s a lot of anime iconography dotted throughout this project.”

Early planning discussions for the look of the movie referenced the work of anime directors Hayao Miyazaki (Spirited Away, Princess Mononoke), Katsuhiro Otomo (Akira) and Masamune Shirow (Ghost in the Shell) as well as the “hyper-real” work of photographers David LaChappelle and Jill Greenberg.

From the outset it was clear that Tattersall would have to work closely with visual effects supervisors John Gaeta and Dan Glass, as the Wachowski’s plan for the movie would involve extensive image compositing and manipulation.
“John (Gaeta) really was instrumental in leading the charge in creating this very digital, super sharp look, which he called faux-lensing,” says Tattersall, referring to the way in which many of the shots in Speed Racer were created by capturing multiple photographic elements, manipulating them and combining them to create an image that could not be created using traditional photographic techniques.

Gaeta described some of the reasoning behind their approach by discussing the influence of anime on the project. “Anime tends to be…very purposefully expressive,” he explains. “It has a lot more to do with the way the filmmaker wants you to feel than what’s real. It’s often not a literal attempt at drawing a simulation of reality but more about enabling the emotional part of the scene to be visualized in some stylistic way. And they do it all the time and in so many different ways; with light, color, the arrangement of layers, the spatial motion of cameras, often bending parallax in ways that seem like you would never get away with normally with a camera out in the real world.”

“We did this with the Matrix as well, but we were prevented from going to that next level of style and fakeness… As far out as the Matrixuniverse was, it still, to some degree, needed to feel real and then with Speed Racer, we were like, ‘Let’s just loosen it up even further.’”

Emulating the anime style of exaggerated colors, camera movement and perspective would be impossible with traditional real world cinematography—recreating this style with real people would only be possible by combining multiple photographic elements into a form of virtual cinematography. “One of the things that we spent a lot of time discussing with the Wachowski Brothers and David is the idea of cinematography not bordered by the physical media,” explains Gaeta. “If cinematography is really like the acquisition of a complete image then really you have an incredible amount of custom tools at your disposal.”

Glass discussed another aspect of the anime influence. “In traditional cel animation, being hand drawn, they aim to become more efficient in certain ways—like they’ll reuse backgrounds or rather than constructing elaborate perspective in 3-D environments, they’ll hand draw them and slide them across the frame and maybe break them into layers and so on to simulate broader camera

moves. We looked at that and thought that was actually quite interesting—the idea of applying that to photographic-based material, including our foregrounds. So really, we conceived of constructing a movie that was built out of quite visibly flat layers…”
This notion of sliding backgrounds led Gaeta and Glass to the idea of creating 360° 3-D “bubbles” out of panoramic still photographs taken of exotic locations to use as background layers. Similar to the process used to create QuickTime VR files, high-resolution photographs were shot in 360° from a selected nodal point or sometime multiple points in the location—these stills were also shot at varying exposures and at different times of day to capture changing lighting conditions, creating HDR (High Dynamic Range) files that allowed for a large degree of manipulation.

These files were then seamed together and then retouched to create 3-D bubbles that were virtual environments that other photographic elements, including actors can be placed within. Sometimes these bubbles were combined in multiple layers to create even more fantastic environments. Simple camera movement through the virtual space could be simulated by panning or rotating the bubble. For more elaborate perspective shifts, the stills were sometimes mapped onto a 3-D model of the simulated space, much like wallpapering a room.

The Wachowskis were initially leery of shooting HD but because the anime-influenced aesthetic of the project called for super sharp, noise/grain-free images and very deep depth of field, tests were performed to determine whether film or one of the many digital formats would be best suited to the unusual demands of the project.

“There was an early on think-tank meeting and various options were discussed,” reveals Tattersall, “and then we had a week of comparison testing various digital formats and film. We tried pulling film processing to get the sharpness, and then single chip HD systems versus the three chip 2/3” systems. We tried the Arriflex D-20 and the Viper but it just so happened that that month Sony were about to launch the F23.”

Glass explained that the tests were conducted in two phases, one phase was to determine whether the digital formats could be calibrated in such a way as to match film-originated material when output to film. The second phase pushed the formats to see if any held particular advantages for the look and style of movie they were creating. As Glass put it, “Why use the HD camera if all you’re doing is making it look like film?”

Ultimately, the team decided that the F23 was the best match for the project—its smaller 2/3” sensor size meant it would have greater depth of field than 35mm film or the D-20 with it’s 35mm sized sensor. In addition, the digital signal the F-23 produced had less noise/grain than film. Glass added that the 3-chip system employed by the F23 offered better color rendition than the single sensor digital systems as well.

Getting the cameras out of the factory added a bit of drama to pre-production. “We were able to get the first five off of the production line,” notes Tattersall. “Our four main production cameras only arrived two weeks before we started shooting…”
To maintain the sharpest possible image, Tattersall chose Zeiss DigiPrimes and DigiZooms for the project and set all of the

cameras at -3dB to minimize noise in the signal.

Tattersall said Vison Research’s Phantom HD camera was also occasionally used to capture over-cranked footage at 1000 fps, as well as the NAC HiMotion high speed HD camera.

The production utilized the hard disk-based Codex Recording System in tandem with a tape-based Sony HDCAM SRW-1 deck to record the uncompressed signal from the F23.

“We’d record uncompressed 4:4:4 on the Codex and also on the SRW-1,” says Dale Hunter, DIT on the parallel B-unit, which shot most of the cockpit gimbal footage. “So we’d have the disk packs at 4:4:4…it holds about 2 tapes on the disk pack.”
The Codex System offered a number of advantages for the complex workflow that the project would entail. “Codex’s real forte is that it runs what is called the Virtual File System,” explains Glass. “So that when you log on to the drive, you don’t see the images or files as they’re stored, you see the images as you’ve configured it to display them—and that’s totally customizable. This greatly simplified file management and streamlined post as circled takes can be automatically sorted and placed into select folders rather than requiring assistants to cull them from all of the other footage on the disk.

“Second huge advantage of the Codex is that in addition to organizing things, you can specify the format that the images come out. So we were able to build formats that actually matched exactly, for example, the Avid’s MXF (Material Exchange Format) format. So the images could be ingested straight into the Avid—there was no conversion process needed, including the metadata that came with them was imported in. It meant for visual effects we were able to set up a DPX format, which is our standard.”
Digital effects supervisor Jake Morrison underlined the importance of getting footage in the DPX (Digital Picture Exchange) format, which he described as kind of a “lingua franca” in the effects world. DPX is derived from the Kodak Cineon file format, which is used to express the logarithmic values of scanned film negative.

“With DPX files, everybody can agree on exactly what happens in there,” explains Morrison. “Everybody’s kind of agreed on what a stop is if you need to go up or down a stop in the image. These are constants that are very handy…(given) all of the experience of the people you’re going to hire.”

Furthermore, one difficulty with doing effects work on HD-originated material is that the video color space is very different than that of scanned film and can vary depending on the camera system used and how it is recorded. “If you change the type of files and the color that everyone’s working with simply because the camera captures it a little bit better in a particular range, you’re really doing a lot more damage than you need to because the act of actually converting it as if you had scanned film just gives you so many more benefits on the backend,” explains Morrison.

Morrison did point out a side-benefit of shooting digital, from an effects perspective. “Because there’s no moving film in the

camera, a lock-off is truly a lock-off. That seems like a minor point, and you do lose some of that organic interest that you get with film. But from a technical point of view and if you’re doing all of the interesting tricks like we’re doing on this one, it’s very important.”

When combining two or more layers of material shot on film, gate weave, both in camera and in the film scanner is a given. “There’s only so much pin registration that you can do with 35mm film,” continues Morrison. “Typically what you have to do is get in there and track one and then stabilize that and pin it to the other one, so it looks like it was shot as one neg.

“The byproduct of that stabilization and pinning is that the image gets soft…This is why visual effects work can sometimes stand out when cut next to unretouched negative. Digitally acquired imagery does not require stabilization since the sensor does not move, so its sharpness does not suffer this additional level of processing. One of the challenges of shooting a movie where so much is shot against blue or green screen is the ability to conceptualize the final composited image while shooting the individual elements.”

As Tattersall remarked, “The success of an image often has to do with how a foreground sits on a background. When you have an actor on a real set, you choose how to do that and where to pitch the light levels across the frame.
“So when you’re dealing with a green screen it is more tricky. If there isn’t a very clear idea of what’s going to be in the background, it’s more difficult dealing with the light/dark tonal separations…I try to have detailed conversations with the compositing department, in this case with John and Dan, about what exactly we are going to put right behind that actor’s head, in the close-up. So that I know that it will sit well in the medium and the wide shots, always considering interactive light and what the light sources will be.”

To insure the 3-D background bubbles would be available for principal photography, a still unit was dispatched far in advance of the main shoot to gather backgrounds at exotic locations all over Europe and North Africa.
“While we were shooting this film it wasn’t just looking into an empty void of green,” says Gaeta. “We shot all of these locations slightly in advance, so that we came to stage always, not just with a jpeg sample that one could look at on a printed sheet, but a high definition navigatable background to composite on stage in HD.”

Glass said that this was accomplished using an on-set composite station. “We built a system through Digital Domain that we called Sparky,” he explains. “It was really a custom live composite station that would take a feed from the set and we could put our spherical background behind it as a layer and rotate it and scale it and pan and tilt it around so that we could create the line up. This was mostly done in the set up stage so that the directors could sign off on it and so David then could plan his interactive lighting accordingly.”

“On every occasion on Speed Racer when we went onto the total green or blue sets, the effects brothers in conjunction with the art department would produce these beautiful composite montage photo-paintings of the possible backgrounds,” says Tattersall, who was very pleased with the process. “So we would all have a pretty good idea of what was going to be put into the background.”
Ultimately Speed Racer was a project that required a close collaboration between the cinematographer and the visual effects department, a relationship that grew out of the mutual respect that they shared for one another.

“Dave was very forward looking and open-minded about the possibilities of making a movie that had so many components that would be sort of mixed and re-framed and colorized beyond the live action, actor moments,” explained Gaeta. “It’s not as simple as saying, ‘Dave can you give us exquisite green-screen exposures?’… Working with Dave was a fairly essential and positive relationship.”