‘Upload’ Director on Filming That Meeting of Nathans and AI Guy’s Real World Test Run

[This story contains spoilers for Upload season three, episode five, “Rescue Mission.”]

For Dave Rogers’ turn in the season three Upload directors chair, it was all about perspective.

More from The Hollywood Reporter

The episode sees downloaded Nathan (Robbie Amell) and backup Nathan meeting for the first time in the Gray Zone, the duo working together to get Luke (Kevin Bigley) out of prison in the Gray Zone following a failed attempt at selling an AI Guy (Owen Daniels) for extra cash. “Rescue Mission” also predominantly features a new (and endearing) turn in AI Guy’s testing, helmed by Aleesha (Zainab Johnson).

Meanwhile, Ingrid (Allegra Edwards) is once again battling her insecurities as she reunites with old — and shallow — friends during a VR workout class. And Nora (Andy Allo) finds herself juggling her and backup Nathan’s feelings, while trying to uncover why her real Nathan was “given over the counter fart pills” to stop his head from exploiding on his downloaded body.

Amid all its comedy, the episode itself really plays with perspective — from whose eyes viewers are seeing events unfold, which reality those events are contained in, and ultimately the authenticity and sincerety of that (upload or human) perspective. Camerawork and editing are always key to capturing the emotions and tones of a scene, as well as getting audiences to understand characters’ points of view.

But for this episode, Rogers and his team were operating in overdrive, using tricks both old and new to help them navigate Upload‘s real — and digital — world. The Hollywood Reporter spoke to the episode five director about those tricks, capturing the two Nathans’ first face-to-face meeting, chronicling AI Guy’s real world test run and what inspired two of the episode’s most dystopic tech advancements.

You were charged with directing Robbie Amell as both Nathans for their first in-person meeting. Can you talk about how you approached capturing two Nathans with Robbie?

Robbie and Robbie — they had this great chemistry. (Laughs.) I didn’t have to say much. He played both of them subtly different. There’s posture things and reactions. But in my episode, there’s a little twist, because it focuses very heavily on who the Nathans are and which one is which. I was just excited that I got them implementing this comedy — backup Nathan doesn’t know when he meets Nora for the first time. Seeing her, we did a bit of a play on what happened in season two when Nathan and Nora meet. It’s like an homage to that.

What was also exciting was the technique of shooting, whether we do a lockdown shot and a split screen; we’re moving the camera manually; we’re doing motion control; just shooting over the shoulder with a double and him; or even just him and comping in his own shoulder and back of his head. I didn’t want to do any cheats. Any time you see him, even if it’s an arm, it’s his arm. We just comped everything in. I was talking to the DP at one point about something where we’d had two shoulders, and I was like, “Nope, I can see a little bit of the body double’s hair. It doesn’t match. I want to shoot him and comp the other one. It needs to be a split.”

The body double (Jason McKinnon) was great and just serves — it’s almost like a placeholder for him to get an eye line and to get some timing as he works on his own thing. Basically, he’s wearing his real-life Nathan clothes with a body double, and then he has to go change his clothes for Nathan backup. While he’s doing that, we’re queuing up what was the best take. Then we have the body double standing where the [real-life Nathan] was and [Robbie’s] acting [as the backup], but now we’re playing back his own dialog to him, so the timing is consistent when we get to the edit bay. I was amazed just watching on set, where we had a mixer and stuff like that, seeing that we can overlap [the performances]. It’s like getting two people, even though it’s one person playing the same role.

Was there anything that inspired your approach?

Before we filmed, I watched a bunch of things just to see what are the examples of when someone plays a double. I remember watching Secret Window, and there was one scene where [Johnny Depp’s character, Mort Rainey] is walking, turning, circling. I thought, this is really good. It’s pretty seamless, even for when it was made. It’s not that old, but the technology still does keep getting better and better. I also went back to Back to the Future Part II, which was one of the first motion control [examples], where they had Michael J. Fox playing like four people in a shot, and they could do the one pass, two pass, three pass.

In the Gray Zone, we see Nathan and Luke sporting Princess Leia outfits. There’s also a guard in an outfit I wasn’t able to identify. Who is he dressed as and were these choices already laid out in the script, or did you play around with a few concepts?

They were in the table draft, the Princess Leia and Prince Eric costumes, and Greg was very particular with the look of the characters. With Greg, [I and costuming] worked on what it’s going to look like and how are they going to look in it. Even the fanny pack — how is Nathan going to wear it, and what’s the best way for him to wear it that we can see it? Because sometimes you’re shooting out of order, so if you shoot a scene that’s later in the show, this is where he’s going to be wearing it for the rest of the show. You have to all these details hammered out very, very early on.

But that [henchman] (Petar Gatsby) look was Zangief from Street Fighter. That was what we were going for — the ridiculousness of that. And [who he works for] Zalan (Yasen Peyankov) is just a typical, bad guy gangster in the Gray Zone. It’s funny, I had questions for Greg and we never answered them. I asked, “Are these guys uploaded or are they avatars?” He’s like, “We don’t know yet. Let’s see.” But I asked because it does affect performance. Is this a kid playing as a gangster, or is this an [uploaded] gangster? I took it that they live there. They’re not avatars, but people that have uploaded into these afterlives, and they reside in the Gray Zone.

Aleesha’s focused on taking a physical clone of AI Guy for a real world test run. It’s filmed from multiple perspectives. How did you juggle those POVs and which to take through that sequence?

With Owen, we were shooting a lot of coverage, and like with the two Nathans, a scene would be written like it’s just two people walking, but it’s not two people walking. It’s one person walking that I have to shoot twice every angle. So we broke this up into two days, and another day we shot him meeting the person who doesn’t have a home. We had to shoot him where we have cameras for real coverage, then him with a camera on his face to move, and a P.O.V. of what he’s seeing. You can’t shoot that at the same time, and [with] the P.O.V., you don’t put it on [the actor] to do it. You have to have a camera person and [the actor] nearby.

That’s [on top of] shooting a whole other scene with people watching him, and I’ve got to shoot multiple coverage on them and coverage on the monitors to cohesively tell the whole story. Because they are all interconnected. Their reactions and comic commentary are very important, what Aleesha’s dealing with. So that’s basically what we did. We’d shoot it, have regular coverage; put a camera on him; run it until we were happy; have the scene with the camera in front of his face; and even get it reverse it to get his point of view to match what he’s seeing when looking around or crossing the street.

And when we’re seeing Aleesha’s experience through a screen, that’s a choice, just like pulling out on him on the street. It’s asking, “What’s the funniest? What’s the best coverage?” When he’s popping out on the street, and he’s seeing people, we want to see what’s on his face. You get the wide coverage to see how he is in the world. I shot crossing the street from both angles, so we had that choice. We could see his face as he’s crossing and from behind. There was something nice about him from behind that you don’t expect what’s going to happen. He’s walking along, we get to him bumping into someone, and he gets threatened and runs.

I like seeing the world and seeing him. He’s so physical and active. But then [in that scene] when he relaxes after he’s out of breath, I like seeing his face. It’s comedic. For that, we set it up that you could see them watching him [from the office] and the heart rate monitor. He’s human now. And we did a great thing once he relaxes, that he just snaps right out of it. Owen held on to those elements of who his character is. He doesn’t have this experience, and he’s not even ready to mimic that complete look.

How did the emotional tone of various parts of that sequence — like the charging station moment versus his iteraction with someone on the street living behind an electronic wall — impact which point of view you used?

The scene where he sees the charging station — we’ve introduced a little bit more sexual stuff in the show that we didn’t do on The Office — is the kind of thing where we can get a little R-rated, pushing the envelope with this guy, so shooting it is just funny. I don’t think you want to see it all. I feel like seeing what we’ve established with this camera on his face is enough — the movement in the point of view. We get that he’s going forward and backwards on this thing, and there are sound effects. The thing that you hear is this squeaking, almost an attempt at a squeaking bed just to say we’ll sell it, have some kind of sound effect that shows he’s interacting with this thing.

When we get to his final scene, there are moments where he’s very sensitive, and I want to say he does have empathy. I don’t know if it’s been programmed or what, but I wanted to show all of that. There’s an episode — I think it was the season one finale — where he gets upgraded, and he goes to talk to Nathan, and he’s like “Hey, are you OK?” I told Owen, it was such a great performance because he really sold that he’s been upgraded. [AI Guy] is much more natural and comfortable. So meeting this person, he’s aware of the difference. You’re on the street, this is terrible, and the elements are out here. I feel like a lot of this he has within him — his own development from learning things, experiencing interaction with people in Lake View. And now that he’s in the real world, it’s that much more prevalent because he’s interacting with real people in the real world.

These scene addresses, in a typically tech dystopian Upload way, how we might handle the visibility of the houseless. There’s a digital curtain that creates the image of diners so you can’t see the man inside. How did you design that?

I kept asking Greg, “What is this thing?” We went back and forth: Should it just be a cloth? Should it be something else” I said, if it’s just a cloth, this is just low-tech. This exists today. So should it be an illusion? We really worked hard to make this thing. I even showed him the season two marketing for Picard. They did stuff with him at bus stations where one side it of it was glass — they had a window — but they would shoot the street and then they would have him beam in. I said that’s one thing to consider. We should make it look the same, have it be the street, but we also want to make it look different so it’s not confusing. There’s even this couple walking behind it. Her arms should be cut off. We shouldn’t see her arm because she’s behind this screen. It was tricky, you know. We weren’t even quite clear on what was behind them. Is there a restaurant? Is it a bar? Is it functioning? Is it open? Is it closed? There were many questions. But the performances from the actors — Owen and Walcott Morgann — were so great. It was just such a sweet moment with a little bit of comedy in there that puts the hat on it.

You have another virtual element in this episode with Ingrid, as she tries to reconnect with old friends while at a VR workout class where everyone is a horse. What inspired that setting?

They already knew it was going to be a workout video and the original description was like Red Dead Redemption, so we built a couple of different backgrounds. Eventually we just focused on the horses. But we get there on the day and I have no idea what a horse moves like. (Laughs.) But I’m talking to Allegra and we had someone else who was the trainer. I asked, “What moves should we do? What can we do?” Allegra had great ones. So we come up with moves, and [the trainer] will do stuff and they’ll mimic him. Everybody’s there with their VR glasses on — and I realize they can’t see the guy, the instructor. So it ended up being just watch [him] and then when we start, do your own thing that’s horsey.

Best of The Hollywood Reporter

Click here to read the full article.