Here's the Definitive Story of the Mammoth Restoration Project, Which Encompassed Frame-Rate Retiming, Digital Image Restoration, Colorization, and Stereo Conversion to Startling Effect

Scratchy, sped-up, black and white footage of soldiers in varying degrees of gray uniforms, zipping around the screen like dull, Chaplinesque warriors, blithely posing for a fixed camera. That’s how most of us see what soldiers experienced in the First World War when viewing historic wartime footage — a flat, gray, silent war, as seen in the first portion of Peter Jackson’s recent documentary, They Shall Not Grow Old. Until the war starts. Then it’s as if Dorothy has opened the door to Oz — suddenly, those flat characters become human beings, looking like we do, moving as we do, feeling what we would feel in their shoes.

Jackson’s remarkable documentary, produced by his Wingnut Films Productions and released by Warner Bros., draws on 100 hours of archival footage, transformed both by Burbank-based Stereo D and his own team at Park Road Post in New Zealand. The artists turned age-ravaged silent footage into colorized, dimensionalized and speed-corrected imagery with natural movement and appearance that finally places viewers into the shoes of those who fought the war. It’s described, via archival interviews in voiceover, by those who fought it.

“The humanity of the people on the film jumps out at you, especially the faces,” Jackson said in the film’s production notes. “They’re no longer buried in a fog of film grain and scratches and stuttering and sped-up footage. This became the First World War film I was waiting all these years to make.”

The Great War, as it was known, was in Jackson’s universe from childhood. His grandfather, a professional British soldier since even before the war began, fought throughout its entirety. The family bookcases were filled with World War I books. “In a way, while I was making this film, I felt that this was my chance to learn what he would have experienced,” the director said.

In 2015, to commemorate the 100th anniversary of the Battle of Gallipoli, Jackson created a photo exhibit, The Great War Exhibition, which ran from April 2015 to December 2018 at the Dominion Museum Building at the Pukeahu National War Memorial in Wellington, New Zealand. For that exhibit, Jackson drew on still photographic images from Britain’s Imperial War Museum (IWM), which he then colorized, overseen by Weta Digital Visual Effects Supervisor Wayne Stables, who also had family members who served in the war.

“Peter is very much of the view that it’s very easy to stay detached when things are black-and-white,” explains They Shall Not Grow Old producer Clare Olssen. “They look old, and you’ve got that detachment, whereas when we colored some stills for this, all of a sudden it evoked a whole other level of emotion.”

That process initiated a relationship between the director and the IWM. Not long after, the organization contacted him and asked if he might be interested in making a documentary marking the centenary of the WWI Armistice using film footage from their archive.

“Lots of people over the years have asked me about making a First World War film,” he said, “but I’ve never really had a desire to make a Hollywood movie about the war.” The Museum holds one of the greatest archives of original footage that was shot at the time for the First World War — at least 2,200 hours’ worth, Jackson noted. “Their only brief was that, while the documentary could be about any aspect of the War that I wanted to do, I had to use their archive footage.”

He returned to New Zealand and began thinking about ways to restore the 100-year-old footage using computer technology available today. After requesting three or four minutes of footage, he and the team at Wingnut Films’ post operation, Park Road Post (PRP), spent several months developing a restoration pipeline to apply to the footage. “They kind of thought outside the box a little bit,” Olssen notes.

First were re-speed tests. The footage was shot hand-cranked at anywhere from 13 fps to 17 fps, so computer software was used to convert the footage to 24 fps. “We tried out different options of [frame] interpolation to try and get the feel of movement back to a more realistic, correct, non-Charlie Chaplin feeling,” explains PRP senior colorist Matthew Wear. “And we worked on correcting spotting and blemishes, to see what could be automated and what would need to be done by hand.” The clips were also given a base black-and-white color grade to even out varying dynamics of shading.

Jackson was astonished by how good the footage looked, and pursued the final step — the addition of color, done in testing by Park Road Post. “We had discovered with the photo exhibition that, by colorizing images, it removed that detachment,” Olssen states. “Peter thought, ‘Well, if I can restore this footage, and we could colorize it, then that would be really cool.’ And the IWM wanted something that could go out to schools — youth engagement was important to them. So the color was really the icing on the cake.”

It was then that, after seeing the restored, colorized footage, Jackson realized he could finally make the World War I film he had always wanted to make.

A restored and color-graded image from They Shall Not Grow Old.
Courtesy Wingnut Films

Park Road Post and Building a Story

Before editorial could begin, the footage had to be put through several preparatory steps by Park Road Post. The first step was re-speeding (or re-timing) the footage – the conversion of the frame rate from its native, in-camera rate to 24 fps, to get rid of that “silent movie” look. “When we see that kind of zipping around normally associated with silent movies, that’s not how we see people move,” explains Olssen. “It’s doesn’t make you think of a human being’s movement, it’s like something else altogether.”

The challenge was that not only was the native frame rate not consistent from shot to shot, it sometimes varied even within shots. “I had been led to believe that the original footage was filmed at roughly 16 fps, but we found quickly that was completely wrong,” Jackson explained. “Most of it was 13 or 14 fps, but we also found everything from 10 frames to an occasional 15 or 16, and, very rarely, 17 or 18. You can’t bring all those different speeds to 24 fps without knowing at what speed the original was filmed. It’s purely guesswork.”

Variations within shots, Olssen notes, were down to simple humanity. “They were hand-cranking. And if it was a calm, nice day, you might see a steady handcrank. But if the adrenaline kicked in — if they were around a firefight, for instance — then, all of a sudden, things might get a bit faster.”

Jackson would sit in the Park Road Post DI suite with senior colorist Matthew Wear and suggest a frame rate, say, 15 fps, and Wear would enter that rate into the system. The two would then view the corrected shot at 24 fps, and make slight adjustments, if needed, until the movement appeared natural. Similarly, after a corrected shot was viewed in the timeline in the edit bay, adjacent to other speed-corrected shots, what appeared correct when viewed on its own might now reveal itself to need a slight 1 fps tick, back or forward, to give a consistent sense of motion across a scene.

Identifying the native frame rate wasn’t always straightforward. It was often dependent on the activity that was filmed. “It was actually easier when people are walking, which has a natural rhythm,” Olssen explains. “You can feel the natural pace of human movement. So if there were troops marching or people walking within a shot, that was easier than other footage that didn’t include human movement,” such as scenes showing soldiers milling around, chatting, or of artillery firing. “That actually proved more difficult.”

Once re-speeding was accomplished, the footage was given a restoration pass by a team of up to eight artists at PRP over several years’ time, again, at this point, strictly for editorial use. The quality of the base image varied. “In many cases, what the Imperial War Museum has,” Jackson said, “is a duplicate, or a duplicate of a duplicate, or a duplicate of a duplicate of a duplicate. So the quality isn’t even as good as the original was.”

One thing stuck out immediately upon viewing the restored footage, even at this level. “The response, whenever we showed people that restored footage, was all about the faces,” Olssen states. “People couldn’t believe that all of a sudden, they were seeing the men, not just a trench. They were seeing the human side of it.”

Lastly, the footage was given a base black-and-white color grade. “We were provided with flat scans, which weren’t particularly dynamic,” says PRP colorist Jon Newell. “Some were too dark, others too bright. And before you can start making an editorial call on a shot, we did a base grade, just to see what was there, so people had an idea of how well we could rebalance each piece of material.”

He and Wear would apply a few layers of grading, Wear notes, “in order to pull detail out of the blacks. And then we had to stretch in different areas, through the mids [and] the low mids, over a series of color grade nodes.”

Jon Newell

Colorist Jon Newell in the DI suite.
Courtesy Wingnut Films

The budget provided by the film’s co-sponsors — the Imperial War Museum and 14-18 (the UK arts commission programming the WW1 Centenary) — was for a 30-minute restored and colorized film. Jackson had spent months reviewing the 100 hours of footage provided by the IWM and had selected the shots he knew he wanted to include. “There were shots that just stood out to him as, ‘Hey, the world needs to see this — and those kids in school in England need to see this,’” Olssen says.

There was another key source of material that would form the other half of Jackson’s storytelling. “As we restored the footage, and the faces of the men became so sharp and clear, I knew that the audio soundtrack should be just the voices of the men who were there,” the director recalled. “No historians, no hosts walking through the trenches, telling us about the First World War. These men should be the ones describing to us what the war was like. This should be an average man’s experience of what it was like to be an infantry soldier in WWI.”

The Imperial War Museum had made him aware of over 600 hours of oral histories and interviews, some with video, which had been collected from 250 to 300 World War veterans by the BBC during the 1960s, ’70s and ’80s. Jackson requested all 600 hours, which he personally reviewed and studied — along with transcripts made by Olssen and her assistant — multiple times over a period of one and a half years. “Peter just thoroughly immersed himself,” Olssen says, “listening to all 600 hours, over and over and over again.”

Jabez Olssen

Longtime Jackson editor and collaborator Jabez Olssen at work.
Courtesy Wingnut Films

Though he had heard and familiarized himself with the audio content, Jackson had not yet formulated the story itself. So he began the long process of crafting the then-30-minute story with editor Jabez Olssen, who had cut Jackson’s Hobbit and Lord of the Rings trilogies (and is Clare Olssen’s husband), in their editing bay at Park Road Post. Working with the full 100 hours of footage restored by PRP, “They sat together and crafted the imagery without getting into the thick of the audio,” explains Clare Olssen. “There were just shots that he knew he needed in this film. And he knew he needed to get selects to Stereo D if we were going to make delivery of the film. So that 30 minutes needed to be selected.”

The story takes the audience through the soldiers’ lives prior to their entry into the British Army, through their training experience, and on into battle. “From a practical perspective, there’s very little combat footage,” Olssen reveals. “There’s not a lot of footage of the guys going over the top, because cameramen didn’t get quite that close to the front lines. So some of the footage that you see is the guys practicing – they did drills so they were prepared for the real thing.” Those clips were supplemented by soldiers’ own drawings, which appeared at the time in War Illustrated. “Those are as close as we could come to footage.”

Few images, if any, display gory combat injuries or deaths; instead, Jackson and Olssen opt for footage of infantrymen warmly carousing with their comrades, intercut with still images of battlefield deaths. “It, again, puts a human face on that side of things,” Olssen states. “You just look at these faces as people, and realize where so many ended up.”

Olssen herself was personally struck with that sense of humanity upon the arrival of one restored and colorized image. “I remember when the first colorized image came in to my inbox, and I brought it up,” she recalls. “It was an image I had seen many, many times before. And all of a sudden, I looked at the soldier’s face, and he looked just like my brother, and I involuntarily had tears streaming down my face. I made that connection. So by putting that human face on this experience, you’re not just seeing a number or ‘some guy,’ you’re actually seeing the human being. And that was really important to me.”

War Comes to Life

The 30 minutes’ worth of colorized, restored film covered by the budget was cleverly extended to 34 minutes by Jackson and Jabez Olssen, who reused some shots, creating still frames via holds and employing zooms and closeups on specific shots. “If more money had been able to be found, we would have done an 80-minute color film,” says Clare Olssen. “But the reality was [that] those funds couldn’t be raised.”

Stereo D was sent about 360 shots of 2K scanned plates and began the one-year transformation process in October 2017. Jackson had originally approached the company in early 2016 to do a proof-of-concept test on five or six shots to find out what was possible. “We ran those shots through the process, restoring them as best as we could at the time, and added color,” explains Stereo D Color Supervisor Milton Adamou. “We came up with entire processes for how to do that, and then we turned them into stereo. We gave Peter those shots and he really liked them, and then he came back to us last fall and asked us to work with them on the project.”

Working alongside Adamou was the company’s stereographer for the film, Russell McCoy, who also supervised color compositing. In addition, Wingnut asked Weta Digital’s Wayne Stables, who had helped create the color still images for The Great War Exhibition for Jackson, to work with Stereo D as colorization supervisor after giving McCoy and Adamou a tour and explanation of that exhibit to help give them some perspective. “Peter was tied up with a number of other things at the same time and couldn’t always review early work. So my job was to really work with Stereo D,” he explains.

Stereo D’s team was made up of hundreds of artists, including 15 compositors and several top matte painters in Burbank, 40 to 50 skilled matte painters in Toronto, and around 200 artists in India, including around 100 working strictly on rotoscoping.

The first step, as was the case at PRP, was to retime the footage, based on information PRP provided. “A lot of the exploratory work was already done by Park Road Post, which was excellent,” Adamou says. “They would send us the shot and tell us, ‘We believe this shot is 11 fps and needs to be retimed to 24.’ So we would take the source frame rate and punch in the target frame rate, and we would have our tool calculate the in-between frames.”

The process was complicated by lateral movement and parallax — side-to-side movement in different planes. “Those would cause problems for the retiming software. Those we would send to our artists, who would painstakingly go through and make sure that the horses had legs and any smears that were introduced during that process would be painted out. So it was largely a two-step process — the retime tool got us 60 to 70% there, and then an additional 30 to 40% to clean up each of the frames by hand.”

Next came an extensive restoration pass. While PRP had done a great amount of restoration work on a large volume of material, Stereo D did its own pass, for the most part, on the specific shots that they would be colorizing. “They did a level of restoration needed to allow them to cut picture with,” Adamou explains. “But then the material would come to us to do a high-level restoration pass, if you like,” to create the sharp edges Stereo D’s artists would need to do their color work. “It’s analogous to color timing. You have dailies coloring, and you have final color. Dailies color is what you would use to cut your feature, final color is what you do in the DI.”

The quality level of the scanned material the team received varied. “As you would expect with material that’s over 100 years old, everything that could be wrong with a film, we would have on various shots,” Adamou explains, including shrinkage, warping, blemishes, scratches and extreme grain. Heavy grain was removed, allowing the images to be smoothed out, and then added back in, with additional sharpening then applied.

A restored and color-graded image from They Shall Not Grow Old.
Courtesy Wingnut Films

One of the biggest challenges laid in footage where either the film was so degraded that it had “big patches of almost nothingness inside the image,” Stables says, or simply frames that were completely missing, requiring reconstruction of the missing material by hand. “We had about 20 paint artists in India who were incredibly skilled at recreating legs that were missing, missing frames, horses running and turning little smears into real imagery,” McCoy states. “Our India artists worked their butts off to get the horse to look like a real horse again.”

In one particularly challenging shot, seen in the training portion of the film, a group of soldiers is seen chatting and cavorting together, with a difficult move as one soldier turns his head and smiles towards the camera. “In that shot, we were missing about 10 frames. It was one of our most challenging recreations of frames. We would get a single frame, and then he’d be in a different position and we’d try and smooth it out.”

The colorization process was as far from what is normally thought of by viewers as one could imagine. “We’ve all seen lots of colorization done over the years, some well, and some not so well,” Stables states, mostly examples with a single tone representing the color of any item in the image. “One of the biggest things people tend to start off with is they just put color washes across things. ‘The grass is green, we’ll make it all green.’” That quickly takes an audience out of the experience.

His early direction to Stereo D was to think about color much differently, resulting in an incredibly complex, time-consuming process that delivered spectacular results. “The images we were seeing, particularly when I showed them the photos from the Great War Exhibition, was much akin to the selfies people take today, which is what people expect to see now. I said, ‘Think about going outside and taking a picture with your iPhone. Have a look at the richness and the vibrancy of color that you see. That’s what we need to see.’ You need to think about it as if this was actually color film this was shot on.”

Milton Adamou, Russell McCoy and Mark Simone

Stereo D Color Supervisor Milton Adamou, Stereographer Russell McCoy and Stereo D producer Mark Simone at work at Stereo D’s Burbank studios
Courtesy Stereo D

The team decided to forego waiting for finished footage from the lengthy restoration process and began by crafting key frames — single frames from each shot that offered representations of each item that were painted by Stereo D’s artists and assigned colors.

The frames were then sent to New Zealand, where they were presented for review by Jackson and another key player, Wingnut’s in-house World War I historian, Pete Connor. “He was involved in all our feedback sessions with Stereo D,” Matthew Wear relates. “He’s someone who can see a small number of pixels or a problem in the background and be able to identify exactly where things might be off, even slightly.”

“It was important for us to know exactly what was inside there, even down to the unit patches on the sleeves,” Stables comments. “Pete did a huge amount of research into every shot we did.”

Aiding the entire team was the real thing: pieces from Peter Jackson’s own collection of World War I items. “Being the WWI enthusiast that he is, Peter has developed a large collection, over the years, of artifacts from the war,” says Olssen, describing a cache including everything from uniforms, helmets and boots to weapons and even aircraft. “The Stereo D guys would be in a room at Park Road Post, surrounded by First World War uniforms of every description — helmets, boots. And Peter would encourage them to put them on, go out in the sun, take photos. The real reference was there.”

“We would use all of that input, from both our eyes and our own creativity, to create the key frames,” Adamou explains. “Those key frames were presented to Peter, or members of Peter’s team, who vetted them, and then they went to Peter. He would provide comments on those, and then we would refine the stills to a level that we felt was good enough to proceed and move ahead with the color process.”

That next step was a tedious one, albeit Stereo D’s bread-and-butter work: rotoscoping. Each and every item in every frame had to be isolated — boots, trousers, jackets, tools, guns — before color could be applied.

“We would be dealing with hundreds, if not thousands, of roto shapes, depending on the shot,” McCoy says. That was all accomplished by hand, mostly by the 100 roto artists in Stereo D’s studio in India. Says Stables, “Unfortunately, there’s no magic wand for making that stuff happen. A lot of it just comes down to a huge amount of manual rotoscoping and labor.”

There were sometimes deep variations within each item. “Grass is particularly difficult to do,” Stables explains. “We usually think of grass as green or brown, depending on the time of year. But the amount of tones and shades and color shifts inside grass is really complex.”

Human skin is no less difficult. “You can get a tone in there that looks right, and then just iterate on it for as long as you can until you feel you’ve got it right.” And if five soldiers are in a shot, none of them have the same tone of skin. “If we walk down the street and see five different guys, even if they’re five different young men of roughly the same age, they’re going to be radically different, in terms of their skin tone.

“And black-and-white footage can lie to you. Because some old film stock would react in slightly different ways, and slightly different odd ways, to colors. Some of the blues would come out really dark, whereas if you photographed them today, it was just turn it black and white. It wouldn’t be that tone. Sometimes we would realize, ‘You know what? This guy’s got what seems to be a particular tone to his skin, but I actually don’t think that’s the case. I think our film stock is playing with us here a little bit.’”

To help manage the massive amount of color information and rotoscoped shapes, Stereo D designed a number of tools specifically for the project. Among the most important was one known as their “palette system.” “That tool would allow us to take hundreds of shapes of roto and run it through the system, to allow us to change colors of hundreds of shapes on the fly,” McCoy explains.

“It’s kind of a three-part key. There’s no flat colors on any of the characters in the image with our palette system. The highs, mids and lows are all colored differently, per object. And sometimes there would be some palettes that were more than shadows, midtones and highlights. We would have five distinct colors, depending on the shape. So if they’re holding a gun, the wood tones are different in the shadows than the midtones and the highlights. There’s no single green put on any blade of grass — it’s always a green-yellow to blue cool mix, and it worked really well. It’s something that we’re very proud of that we developed especially for this project.”

Completed shots, and shots in progress, were reviewed twice weekly, typically over T-VIPS (Tietz Video and Image Processing System) video conferencing, allowing the Stereo D team to present their shots to the folks in New Zealand for discussion. Besides Wear and Newell, Stables was on hand to make specific review notes, helping the team get the imagery to a level that all agreed was ready to present to Jackson. “He was a proxy internal resource for them, with some knowledge of how Peter and Wingnut operate,” Olssen explains. “Having Wayne review was as close as having Peter’s eyes as one could get.”

The other half of Stereo D’s colorization work took place in the backgrounds — the creation of matte paintings by their artists in Burbank and Toronto, based on the images in the existing frames’ environments. “We really just painted in as much detail as we possibly could,” McCoy states.

The company’s artists each brought their own talents to the game. “We had certain people who are really, really good at buildings, others were great at environments, others that were fantastic at skin tones,” Adamou says. “We even have a guy in Burbank, Zack Smothers, who was our mud guy. Mud sounds like it would be easy to do, but it’s actually one of the hardest.”

A group of Lancashire Fusiliers wait in the Sunken Lane on the morning of July 1, 1916. At 7:30 a.m. they clamber up the dirt bank to make a rush at the Germans. This shot was taken about 15 minutes before that attack. What these men don’t know is that the vast majority of them won’t advance more than a dozen paces before being killed. Inset: The very same spot in 2018. — Clare Olssen
Courtesy Wingnut Films

An extremely important source the artists drew on was a collection of 1,800 still photos Jackson himself took during trips to World War I battle sites and elsewhere — the very same locations filmed in the footage he was using. “Peter went out to Belgium and found many of the locations and photographed them. Those locations are pretty much the same today as they were a hundred years ago,” Stables explains. “So they were very useful, not only for landscape details — the same kinds of trees are still growing — but also the buildings. A lot of really old buildings out there were built around the time of the war. So you can see what types of bricks they used, what type of tiling on the roofs, etc.”

One location he visited was the front lines at the Battle of the Somme, which began on July 1, 1916. “It’s the first day of the battle, and you see the men, the Lancashire Fusiliers, sitting on a bank, just at dawn, waiting,” Olssen describes. “There are four or five shots of them — there’s groups of men, a guy polishing a gun, things like that. Peter found that exact spot, that same earthen bank, which is in an area referred to as the Sunken Lane, a sunken farm track halfway between the British and German trenches. They snuck in there to get in position for the attack.”

When the film premiered in Europe, Jackson invited his colleagues, who had toiled so hard on the project, to visit the Sunken Lane. “He said, ‘Come on, we’re going to go to the Western Front and have a look.’ I can’t tell you how harrowing it was to stand in that exact spot, and walk up over the bank, just as the Lancashire Fusiliers had done. And you go up over the top, over that bank, and walk down about 100 meters, and there’s a cemetery. And the majority of those guys we see in the footage are buried there, because very few of them survived.”

Images of the Western Front today

Editor Jabez Olssesn climbs up the bank and walks over the top, just as the men of the Lancashire Fusiliers did. Inset: A few meters on, he discovers a cemetery and in it grave after grave after grave of Lancashire Fusiliers — all killed July 1, 1916. The images in our film are the last ever taken of these men. — Clare Olssen
Courtesy Wingnut Films

The team has actually looked at trying to connect family members of those seen in the film with the otherwise-anonymous soldiers. “We don’t know who those men were in the footage. But we know the names of the ones that died, and we know when that footage was taken,” Olssen explains. “We’ve got a photograph of every single gravestone. It’s important to not just see these as a bunch of men or as a unit, but the individuals as humans. And if we can figure out who they are, that would be amazing.”

Jackson’s reference stills were immeasurably beneficial to the matte painting team. “We tried to keep everything as much as we possibly could to what Peter shot as reference,” McCoy relates. “He was very big on if it’s a green hill, and it looks beautiful, then it’s a green hill, and it looked beautiful then. Not to make it drab, just because it’s war. It’s still a beautiful hill of green grass in the summer of France.”

The shots also helped with things like sky replacements, Adamou says. “Because of the nature of the film at the time, it didn’t have the latitude that we’ve had in the last 20 years. So you’d get overexposed skies or underexposed shadows. So we did sky replacements, using the skies in Peter’s shots. Those cloud formations end up in the movie, which I think is really cool.”

The team’s Palette Tool was also utilized in matte paintings. “The skill that our matte painters had, going in and painting every blade of grass, and those trees, made it useful, as well, for the matte paintings. So it was really a hybrid approach, because we had all of these rotoscoped shapes over these mattes. It just gave a vibrancy to all of the objects. I would say that a big part of the realism you’re seeing is the result of those matte paintings and the artists behind them.”

The paintings were easily tracked, thanks to the limited motion of the photography of the day. “They didn’t exactly have cameras on drones,” Adamou points out. “At most, you would have a pan. So because the photography was quite limited in the movement of the camera, we were confident that we could track in those matte paintings, and, using certain techniques, give them life.”

Where, as noted before, the image was so degraded that portions of the frame were not present, the painters would, again, draw on the imagery in Jackson’s stills. “There was one shot, for instance, where there was smoke going across the frame, where there was a horse and a cart, galloping across a road. And inside the background there’s bits of smoke around them, and bits of foliage. And the footage was just so degraded that the horse and cow just disappear, for frames, just dropping in and out.” But because the foliage in that area hadn’t changed, the painters were able to make use of that seen in one of Jackson’s stills, or even from other footage, “and we could just put a few bushes inside that bit of the frame, so that where it disappears, it’s as if they’ve simply gone behind something. And it actually feels and looks accurate to the environment. That’s how we would approach the image, when we had those types of challenges, where there was just no part of the image left to put color into.”

Stereo D would also draw, on occasion, on the original black and white plate they had received, to add some detail back in. “One of the things that can happen during restoration and colorization,” Olssen explains, “is that some detail can drop out, just due to the nature of the process. For instance, you might have a soldier with beard stubble. As part of the restoration process, that stubble might be lost. So Stereo D would colorize the footage, and they might pull back in some of that original detail in the black and white of the stubble.” As part of the restoration process, Stereo D would receive both the original plate and Park Road Post’s restored plate, along with notes about how they had achieved that restoration. “So they didn’t always just do their colorization on top of the restored plate. They wanted to make sure that none of that detail was lost.”

More Story To Tell

As completed shots began to come back from Stereo D, Jackson turned to another key story source, which resulted in a film running three times the original 30 minutes’ worth: the archival audio interviews with the veterans.

As mentioned, Jackson listened to and studied the interviews numerous times. “He would immerse himself in each soldier’s story,” Olssen reveals. “And in doing so, he was able to pick up common elements, rather than hearing little topics in isolation. He was actually hearing the same story and the commonalities between men,” enabling him to find and tell a broad story, while still homing in on specifics. “We might have three guys talking about using tea to shave, for example. Peter put that in the film, because he heard it over and over and over again. He wanted this to be the slice of life of the general soldier on the Western Front.”

He also made use of as many of the 250 to 300 interviewees’ voices as possible. “He could have easily have said, ‘Hey, we’re just going to take these four guys. They’ve got really cool stories. We’ll just illustrate that.’ But he wanted the common experience. That was important to him.” There is, in fact, no writer credit on the film. “Peter felt that the guys told the story. It’s their story to tell.”

600 hours of interviews is a lot of material, so whenever the director began developing a scene with his editor, Jabez, on a topic he knew he wished to cover, he would turn to Olssen and her coordinator, Jess Henry, who had catalogued each of the topics in their transcripts. “We had done a pass on all of the interviews, and cut them up into clips and placed them into folders on some topics that Peter had given us,” she explains. “So if he asked, ‘Hey, guys, can you see what you can find on trench life?’ or ‘food?’ or ‘shaving?’ we could go through the transcripts and find soldiers speaking on those topics and point him in the direction of those clips.”

Once he began reviewing the material, and knowing the amount of rich content available in the interviews, Jackson decided to create a prologue and an epilogue, each about 30 minutes in length, to extend the film to its finished total run time of 94 minutes. The scenes use a combination of additional footage — this time unrestored, uncolorized and un-speed-corrected — set in frames within the screen, accompanied by, in surround, the sound of a projector’s shutter, to give a sense of the films as newsreels of the day (some of which were also included). “Peter just thought, ‘Let’s take people to that time,’” Olssen explains. “Because there were newsreels coming back to Britain, at the time. It was a kind of unprecedented access to the Front Line.”

Adds Matthew Wear, “That black and white footage was just pretty much graded as the colorist would, to make sure that you had good shadow detail, that the highlights weren’t clipped, and that the shots worked together. We kept it honest, pretty neutral. We didn’t want it to appear anything other than nice-looking, well-graded black-and-white footage. And there was no restoration pass. It has its scratches intact.”

Additional footage was made by the inclusion of photographs and posters, which Stables had colorized by the same team of artists who had colorized the photos in The Great War Exhibition.

All of the work for those sections was done by Park Road Post; Stereo D’s work was limited to the colorized, dimensionalized 34-minute mid-section of the film. The transition from that flat, scratchy, speedy footage, with the framing window slowly filling the screen, revealing Stereo D’s first shot, is startling. Says Olssen, “Suddenly, you’re fully immersed in it, as the soldier would be.”

Final Touches

Once restored, colorized and composited, Stereo D’s 34 minutes of footage was sent back to Park Road Post for color-grading. “The color-grading process was integral in serving the realism of the images,” Wear states. “Each element within every shot was assessed and adjusted, according to the uncompromising vision set by Peter Jackson.” Adds Newell, “We’ve kept it grounded. We’ve tried to make it as real as possible, by making sure everything is technically accurate. We didn’t do a creative grade. We tried to keep it real.”

Though Stereo D’s artists had the color key frames to help guide them in their pursuit of that same accuracy, Olssen notes, “Remember, their talented artists were seeing the shots that are selected. They weren’t seeing the cut, because that was evolving throughout the whole process. So when you lined those shots up, of course, you had to do a color grade. So it became a very, very important part of the process, once their colorized scenes came back to Park Road.” Notes Newell, “In a sense, it’s just classic color grading.”

Along with the completed — and flat (not dimensionalized yet) — footage, Stereo D also sent along, as part of the EXR package for each frame, the individual rotoscoped black-and-white mattes used to create the colorization. “We were able to take any of those cutouts, as many as 65 or 70 in one shot, to isolate certain areas within the flattened main image and just adjust those areas,” Wear explains.

Some items may simply have required more refinement in this second pass, Newell points out. “They might put all trousers in one rotoscope matte, and then we might have to isolate certain trousers to be a different color from other trousers. The variations on khaki are quite amazing,” depending on how long a soldier had been wearing them or their rank, etc. “We’d either split them off, or we’d use keys within the rotoscope matte and could build up more subtlety that way, within the image.”

A rack of World War I uniforms in the DI suite.

Vintage World War I uniforms were available for reference in the DI suite.
Courtesy Wingnut Films

Once again, Jackson’s collection of historical artifacts — and Jackson himself, along with Connor — were present at the color-grading sessions, to help guide the colorists’ work. “We were spoiled in terms of the amount of reference material we were given,” Newell says. “We didn’t really want for anything at all. We had an expert in the room telling us exactly how it should be.”

Racks of uniforms and equipment and tables covered with hats, webbing, ribbons, bandoliers, etc., were always at the ready. “And sometimes, we could divide our screen, which is five meters in width, and put the image we’re grading on one side and just white light on the other. And then we could literally hold up jackets or greatcoats and check that they looked correct under the same color temperature light.” For grading environments, Jackson’s location photographs were always available, allowing adjustments to be made in a similar fashion.

And, as Stereo D did to restore missing detail, Wear and Newell could draw on the original restored black-and-white footage to find whatever their instinct told them could bring yet more to the image. “One of the things about colorization is that some subtlety of light and shade and face tone roll off into black and white, and ambience is sometimes lost,” Wear explains. “Jon and I found a technique where we could apply the luminance from the black and white back into the colorized version, to readjust the luminance of each pixel set. And that brought the image back alive and added a reality and depth.”

“Those guys are incredibly clever,” Olssen notes. “They would pull detail back through from the black and white, if they needed to.”

Once that 2D color-grading adjustment was completed, the footage was sent back to Stereo D, where they would perform their well-known dimensionalization skills, turning them into 3D. “They had already done all of their dimensionalization, in a sense, during their rotoscoping process for colorization,” Wear notes.

PRP’s own stereo artist, Antonis Voutsinos, then did a final convergence adjustment across the entire 94-minute film. “We do a 3D pass, a convergence adjustment, just to create consistency from shot to shot,” Newell explains. “So that you don’t get a huge amount of depth in one shot, followed by one that is really shallow. There are subtle changes that we can make in order to create consistency for the viewer, so that it isn’t too jarring a jump between shots.” Voutsinos also created whatever dimensionality was required in the prologue and epilogue, mostly appearing in the framed black-and-white footage, which inset somewhat from the plane of the screen as if one is peering through a window.

James Carroll

Foley artist James Carroll at work
Courtesy Wingnut Films

The final element of the film, of course, is the soundtrack, mixed by re-recording mixer Phil Heywood. “Just as the soldiers saw the war in color,” Jackson said, “they certainly didn’t experience it silently.” Sound effects editor Brent Burge had the sounds of period weapons and artillery recorded at a military base in New Zealand, which his team incorporated faithfully. “We wanted, as best we could, to give the impression of the sounds the soldiers would have been hearing,” noted the director.

The aforementioned archival interviews, skillfully crafted from Jackson’s selects by dialogue editors Emile De Le Ray and Martin Kwok, were to be used as the film’s narration. After beginning to view the final footage, Jackson saw something else. “Peter wanted to be very, very faithful to the footage,” Olssen explains, “and when he saw the men’s lips moving, he said, ‘Well, we need to look at that.’”

Forensic lip-readers were hired to read the lips of the men on-screen and tell the team what was being said, as best as could be deciphered. “Then we had our military historians identify, from shoulder patches, etc., what regiment that soldier belonged to and where they were from.” Actors from those very regions in England were then hired to record the ADR, using the correct accents and dialects. “We even looked at how or if accents and dialects changed over time, which was interesting.”

That was just one of countless details the production was proud to include to help bring the time-ravaged footage to life, both to the eyes and the ears. “We work on roughly eight to 10 day-and-date movies a year, and every movie we work on is special,” says Adamou. “But this is a different kind of movie. It has historical implications. People usually think of World War I as a silent war, a flat, black-and-white war. And you go from looking at these pictures as memories, or fragments of memories, and the minute they are restored, and you add color, and then you dimensionalize them, they become real people.”

“It was really important to Peter that we were faithful to these men,” Olssen states, “because they’re human beings. Whatever detail we could find out about these men, we did. And that’s what you see.”

They Shall Not Grow Old is available now on digital platforms and arrives May 6 on DVD and Blu-ray in North America.