By Iain Blair
When the first Mission: Impossible film starring Tom Cruise came out in 1996, it began a tradition of bringing on a new A-list director for each movie — Brian De Palma was followed by John Woo, J.J. Abrams and Brad Bird. This changed after Christopher McQuarrie directed 2015’s Rogue Nation. The Oscar winner has been at the MI helm ever since, cementing his and Cruise’s long-standing creative collaboration, which includes Top Gun: Maverick, Edge of Tomorrow, Valkyrie and Jack Reacher.
The pair are back at it with Mission: Impossible – Dead Reckoning Part One, in which Ethan Hunt (Cruise) and his IMF team embark on their most dangerous mission yet: to track down a terrifying new weapon that threatens all of humanity before it falls into the wrong hands.
As you’d expect, the seventh film in the series is brimming with death-defying stunts, exotic locations and cutting-edge visual effects. Here, McQuarrie talks about the challenges of making the film, dealing with the VFX, and how they created the nail-biting train/bridge sequence. (His answers were edited for length.)
What were the main technical challenges of pulling all this together?
Too many to count. Whether it was COVID protocols and shutdowns, engineering challenges, or being in simultaneous prep, production, post and/or promotion on Top Gun: Maverick and two Mission: Impossible installments — compounded by the standards we set for ourselves from the outset — there were no easy days or shots on this film. None.
How tough was the prep and shoot?
Preproduction happened twice, each time under extreme conditions. The first was a rapidly accelerated prep in late 2019. Top Gun: Maverick had just wrapped, and the studio wanted Mission: Impossible for summer of the following year. In order to make that date, I told them I would have to start scouting immediately, with no time to so much as outline. As such, preproduction, scouting and writing a first draft were all happening simultaneously.
Cut to February of 2020. We were in Venice, Italy — two days away from the start of principal photography — and I was under enormous pressure with a first day of shooting full of unknowns. Add to this the fact that Top Gun Maverick was not finished. We’d been hearing reports of some sort of respiratory virus and the next thing we knew, Venice was ground zero for the pandemic in Europe. With Mission shut down and our crew being evacuated, editor Eddie Hamilton and I set up an editing room in our hotel and went back to work on Maverick, working remotely with the rest of the team in LA.
Over the next few days, the Mission crew pivoted to Rome but was asked to leave for London shortly thereafter. The entire movie business had shut down almost overnight. We set daily Zoom calls and gathered all department heads together, which resulted in a unique level of interdepartmental understanding. It revealed a lot. We essentially reprepped the movie, yet we were not in what you would call actual prep because everyone was isolated at home.
So we had two prep phases, neither of which was in any way conventional or complete. The consequences of this were felt in big ways and small until the delivery of the film. As for the shoot, Mission was one of the first films to resume production in the shadow of COVID. We were in uncharted waters, developing many of the protocols that would eventually evolve into the industry standard. Everything was trial and error, compounded by the fact that we were back in Venice, often working in narrow alleys with a full crew… at night.
The city was largely deserted, and my memories of the shoot there are slightly surreal. It was very hard for people to make the adjustment to a new way of living and working. I realized I had to lead by example. I prefer to direct from right next to the camera, which was no longer an option. I had to work in total physical isolation whenever possible. I took to wearing a communication system called a Bolero [from Riedel] — a Wi-Fi-based radio system with six programmable channels. The advantage of Wi-Fi radio is that conversations are much more like on a phone than a walkie talkie, so it’s very fluid.
In time, I was able to run the entire set with three buttons — one channel to my first AD, another to playback and a third to my script supervisor. I would position myself in a corner, facing outward, with my monitor directly in front of me so no one could gather behind me. Before long, the Bolero became my primary means of communicating. You could operate at a distance yet never have to raise your voice to communicate or deal with the frustrations of conventional radio. No matter where my AD was, we could speak quietly as though we were next to one another — a bit like the team in Mission, actually. It remains our standard means of communication.
Much later, I was contact-traced and forced to work from home, directing over Zoom (reports that I’d been hospitalized were not true). While initially frustrating, I eventually acclimated and realized I had developed a new and powerful tool. When I returned to set, I set up a remote feed to a splinter unit shooting inserts on a separate stage and connected to them via a fourth channel on Bolero. I was able to direct two units at once.
Later still, I was on-location supervising two other splinter units: one at Longcross and another at Pinewood. Eddie Hamilton was connected to set via the same system and could watch and give notes in real time, whether he was cutting at Longcross or in Soho. This is just one example of how an initial compromise to our workflow would eventually become an advantage. Many of these discoveries are still in use — things we might never have discovered without the difficulties of COVID. One of many Mission mottos: Disaster is an opportunity to excel.
Obviously, you had to integrate all the VFX and post into the shoot right away. What was involved?
Visual effects were extremely unconventional on Dead Reckoning. It was initially very difficult to review and approve effects on the big screen simply because we were not able to gather in a screening room together. This slowed things down considerably because I could not confidently approve effects on a small screen.
Because we were shooting over such a long period, with numerous starts and stops, we would often have VFX stand down to keep the budget in check. As such, there was no advantage to such a long production schedule. VFX were slammed right up until the last minute. We finished the last shot the day before the premiere in Rome.
Did you do a lot of previz and postviz?
Only two sequences. Because we’ve shot more than our share of car chases, the only part of the Rome chase that was previsualized was the Spanish Steps, owing to the hugely complex problem of how to destroy the steps without ever physically touching them.
We built three sections of the steps on the Longcross backlot, and they stood there for well over a year. The steps portion of the sequence is comprised of countless specific shots, all of which involved a huge effort regardless of their size or length. The Fiat was incredibly resilient, whereas the big SUV was a diva, breaking down after every other take. The splinter unit would be on the steps occasionally, shooting background players as elements.
The Spanish Steps became an albatross around our necks and something that amounted to an unbelievably elaborate means of getting Hayley [Atwell, who played Grace] and Tom to switch places. I was terrified the punch line was simply too subtle and that all that effort had been for nothing. The first time we screened the film for a test audience, the gag elicited a huge response. I’m still shocked that it works. I evaluate all gags now on this basis: What does it cost versus what does it achieve? It’s not to say I wouldn’t stop myself from doing such an elaborate gag again, but I would certainly be more honest with myself about what I was doing.
As compared to Rome, the train wreck was previsualized extensively. While the physics changed, as they always do when things get real, the basic sequence of events never did. Production in Mission is very fluid, with a great amount of improvisation, experimentation and creative discovery on-set. Yet the train was a massive undertaking that involved enormous resources and engineering.
The sequences had to be both predictable and flexible. This started with a previz involving the broad strokes of the wreck, with the understanding that the actors would be finding performance on the day. The essence of the train sequence is suspense, not action. That means characters struggling through an environment that is working against them. It was the tendency of the previz team to animate the characters rushing through the sets as quickly as possible. It took numerous iterations to convince them to slow things down. The concept of slow action was simply not in their realm of experience.
The earliest previz is dated February 2020. Certain nerds who would speculate on alleged sources of inspiration for the sequence should make note of that date. The train sequence is not inspired by or derived from anything other than the decision to wreck a train simply because Buster Keaton, John Frankenheimer and David Lean all did it in films I greatly admire. I simply wanted to try my hand at it.
The film was cut by your go-to editor, Eddie Hamilton. Was he on the set? How did that relationship work?
Editing was down in Soho, first in offices on Dean Street and later on at Clipstone Street, though Eddie Hamilton would often work remotely in a shipping container on our makeshift backlot. Because of the uptick in streaming production prior to the pandemic, Pinewood, Shepperton and Leavesden were largely unavailable to us, and we had to build soundstages and facilities from the ground up at Longcross.
In late summer of ‘22, we shut down production to focus on a rough cut of the film. Eddie and I took our families on a working vacation for five weeks, moving to an island off the coast of Maine. It was the only real “break” we had during the nearly three years of production. It was also the first time we were ever able to see the bulk of the film at a run.
Our last official day of photography was in April of 2023, just two months before the premiere. Eddie and I were still making small editorial changes while we were in the mix. So as not to change the length of the reels and disrupt the sound team, we made sure all changes fit within the exiting length of the reel down to the frame. While color-timing the film in Dolby Vision, I noticed a flaw in one VFX shot that had escaped us at lower resolution. Thus, the last VFX shot was completed the day before the premiere.
There are some amazing set pieces, like the train/bridge sequence. How was that done? Break it down for us.
For the train itself, it’s probably easiest to break it down into interior and exterior sets, starting with the locomotive. I wanted it to look romantic but also a little mean. Because there was no existing locomotive that we could destroy, we had to build one from scratch. SFX supervisor Neil Corbould and his team built a 70-ton monster locomotive.
The complications of building a working steam engine were incredibly complex and expensive. It was deemed cheaper to build a shell with a diesel engine housed in the coal tender that trails behind the locomotive itself. This created problems later when we discovered that the lack of weight in the front of the train meant there was less tonnage to keep the front wheels pressed down on the track. As a result, the locomotive had a limited top safe speed. The train could go much faster, but it wasn’t safe to do so. This led to endless frustration. The bulk of the train cars are actual working ones that we rented and dressed for the sequence yet had to return intact.
We owned two cars that could be modified and destroyed. We named them Butch and Sundance, one of which was kitted out for stunt rigs and fighting, the other of which was doomed to be dropped off the bridge when the time came to wreck the train. We shot the bulk of Ethan and Gabriel’s fight, along with the majority of the wide establishing shots, on a working railroad in Norway, control of which was given to us for a few weeks. The wreck was shot at a quarry in the UK with the use of a track and partial bridge we built for that purpose. We also built a decoupler rig so that Tom and Hayley could shoot the action of uncoupling the locomotive from the rest of the train practically.
The entire rig was being pushed by a conventional diesel engine and shot from cameras on the rig itself or from a pursuit vehicle driving parallel to the train. We shot that particular action on an unused track in the UK. The section of track that had a usable road beside it for shooting with the pursuit vehicle was only about two miles long, meaning the actors had very little space to shoot the action at speed before we had to reset.
Moving the entire show back to one was incredibly time-consuming. Despite this, we finished the UK portion of the exterior train action (just about everything other than the fight itself) in just four days. The interiors were all shot on the stage or backlot at Longcross. The trains were designed so that interior colors (red, blue, white, sleeper) could be swapped in a matter of days. The baggage and kitchen cars were more particular in their construction, so they couldn’t be refitted. There was only enough space for two train interiors on the stage, so choices had to be made as to what interiors would be used at any given time. Because we were still working out the action inside the train, this made planning very complicated.
Two massive gimbals were built on the Longcross backlot. One gimbal was about 30 feet high and could support a single train car, able to tilt 30 degrees up and down or 15 degrees side to side. The car was designed to be fitted with a blue or white interior as needed. Later, we swapped this out for a kitchen car, which was waiting off to one side. The second gimbal was designed to tilt from level to a full 90 degrees inverted. This had a blue interior only and was designed specifically for the sequence’s climax. A rail system was fixed to the roof of the set, enabling us to attach a camera that could travel with the actors as they climbed or quickly positioned to shoot basic coverage of their stunt work.
Finally, a purpose-built, fixed vertical car was attached to the back of the Spanish Steps set. This is where we shot the bulk of the character work between Tom and Hayley as well as the bits involving the piano before it fell. A splinter unit stayed behind to shoot all the extremely close inserts. Shooting on both vertical sets was extremely slow and particularly challenging for the actors and camera operators. That’s just an overview of the engineering component of the sequence. The shooting and editing of the sequence would fill volumes.
Apart from ILM doing the bulk of the VFX, I noticed in the end credits that you had a lot of vendors, including BlueBolt, Rodeo FX, Lola, SDFX, Yannix, beloFX, Alchemy X, One of Us, Cheap Shot, Atomic Arts, Untold Studios, Blind and Territory. How many shots were there?
There were 2,500. While Mission is almost entirely practical, there is a great deal of VFX work — removing camera rigs and safety gear, etc. The underwater submarine shots are entirely CGI, which was a very big point of concern for me. I am deeply skeptical of purely CGI shots, particularly when they are the opening and closing shots of the film.
Special effects supervisor Alex Wuttke assured me it could be made to work, and I was satisfied with the end result. The shot above the ice at the end of the film is practical, captured when we filmed the Arctic portions of Dead Reckoning Part Two in March of ’23, just a few months before completion of Part One.
What was the most difficult VFX sequence/shot to do and why? Was it Tom’s motorcycle jump off a mountain top?
With over 2,500 shots, many of which were extremely ambitious, it’s hard to pick just one. The motorcycle jump, despite being shot on the first day of principal photography, was one of the very last shots to make final. The shot in the trailer is not the finished product.
The nature of the camera move and the fact that the ramp had to be smooth, yet the terrain had to feel natural, made for an insane number of frame-by-frame complications — compounded by our insistence that Tom himself never be in any way manipulated. No digital doubling was permitted.
The shape of the ramp was a challenge as well. Most versions were difficult to perceive in the wide establishing shots. We finally settled on the asymmetrical ridge in the final film for this reason — it allowed us to add a directional shadow for perspective. Despite countless iterations, there were still a few frames that I was not satisfied with — a judder in the original plate caused by the intersection of a lateral camera move just as Tom hits the upward slope of the ramp. Finally, Eddie and I added a simulated camera bobble — as if the camera-ship encountered turbulence. Just a tiny, barely perceptible 22-frame bump to hide this unfixable anomaly. Unless I point it out, you can’t see it.
I haven’t even gone into the countless graphics shot in the film, all of which are precisely honed down to the millisecond, conveying non stop story that has to be conveyed in shots often as short as 13 frames. They all have to work in a way that requires no effort from the audience. I drove the graphics people crazy, but their work speaks for itself. And there was, of course, the Entity itself, which went through several iterations before we settled on just the right design, behavior, color and sound. It really didn’t take its final form until shortly before we delivered picture.
What about the importance of music and sound to you as a filmmaker?
Both are vital. More important than dialogue. I’ve seen music alone increase a test score by 20 points. I consider the final mix to be a full third of the movie’s emotional impact, along with character and cinematography. I also don’t trust my ears. I have hearing loss and cannot perceive certain sounds as acutely as my team. This is why my mixes are the way they are. It’s not enough for me to hear it. I have to feel it.
Where did you do the DI, and who was the colorist?
The colorist was Asa Shoul, and we did it at Warner Bros. De Lane Lea in Soho. We’ve worked with Asa many times and have a good shorthand with him. Scenes like the Committee meeting and the safe house were hardest, as those environments are essentially expressly lit boxes filled with exposition.
Asa found a great texture for both. And a grain pass was essential. Dolby Vision has extremely rich color and the deepest of deep blacks, which is great. What proved to be a problem was white. The heavy use of source light and back lights in Venice made for extremely high contrast between light and dark in many shots. The backlight was simply overpowering — like staring into the actual source. I’d never seen anything quite like it. Going through every shot would have been extremely time-consuming.
Asa managed to create a universal pass to decrease the high end of the white spectrum in those sequences, which brought the intensity under control. When you watch the scene in the safe house after the Venice foot chase, you can note the occasional cut where the shift from light to dark is quite impactful. Keeping in mind that these have been pulled back, you can get a sense of how jarring some of the cuts in Venice would have been before we toned it down.
Did the film turn out the way you hoped it would?
There are things in the film of which I am immensely proud — particularly the performances of the extraordinary cast. And I am in awe of our incredible crew. The adversity they faced over such an extended period of time is unprecedented. They are simply the best in the business. That said, every film ultimately leaves me feeling the same way: I can do better.
Industry insider Iain Blair has been interviewing the biggest directors in Hollywood and around the world for years. He is a regular contributor to Variety and has written for such outlets as Reuters, The Chicago Tribune, The Los Angeles Times and the Boston Globe.