One of the year’s most visually striking movies, the sci-fi adventure Alita: Battle Angel, is a film more than a decade in the making. It tells the story of a cyborg revived by a kind scientist in the year 2563, who struggles with the mysteries of her past while trying to make a new life for herself, and it does so with an impressive blend of CG characters and environments with live-action actors.
Adapted from the manga series Gunnm by Yukito Kishiro, Alita is a long-delayed passion project for co-writer and producer James Cameron and director Robert Rodriguez that stars Maze Runner actress Rosa Salazar as the film’s titular heroine. Salazar’s impressive performance as Alita — and the visual effects artistry that translated that performance to her CG character — gave the story’s cyborg protagonist an impressive level of humanity and earned high praise from audiences.
Digital Trends spoke to the visual effects team for Alita, one of the contenders for an Academy Award nomination in the Best Visual Effects category at this year’s ceremony. That VFX team includes visual effects supervisors Eric Saindon (of Weta Digital), Richard Baneham (of Lightstorm Entertainment), and Richard Hollander (Lightstorm), as well as animation supervisor Mike Cozens (Weta).
Digital Trends: What made this the right time for Alita to finally get made after so long in development? Did the technology finally reach a certain point to make it possible?
Eric Saindon: It wasn’t as much of a technology thing as it was getting the script to the point that it could be made into a movie, because the technology has been evolving all along. We’re now able to capture a performance like Rosa’s and put it onto a CG character, and this is actually the first time Weta has worked on a humanoid character.
We did Gollum [in the Lord of the Rings films], we did the Na’vi [in Avatar], and we’ve done apes [in the Planet of the Apes franchise], but we’ve never really had to put a CG human on camera for the hour and a half that Alita is on screen.
Richard Baneham: On Avatar, we had a lot more latitude, [because] we only had a handful of shots where we had the CG characters in the live-action world. In this case, it is incredibly important to believe that Alita is present in every scene, so to achieve that, we had to make sure that Rosa was represented in every single shot of Alita. It really was our job to protect Rosa’s choices and make sure they get to screen.
Pantomime is your enemy when it comes to CG characters. So any outside forces or interactive forces — whether it’s a character making contact with an object or other actors — must all be fully represented on both the photographic side and then on the CG side, too. So we relied heavily on Rosa and her performance.
There’s typically a lot of testing when it comes to a project as ambitious as Alita, just to make sure it’s even possible. Was there a moment early on when you realized translating Rosa’s performance was indeed going to work?
Richard Hollander: It all started with Rosa Salazar’s performance. Once we saw what she was going to do with the character, and what she had planned for the character in the live-action shoot, that was the beginning of knowing we could do this.
Saindon: There’s a scene where Rosa grabbed an orange and bit into it, and she made some crazy facial expressions. When shooting that, I remember thinking, “This is never going to work. We’re never going to be able to get this.” And then we got the scene in and we put the performance onto Alita. When we looked at the two side by side later, with Rosa and Alita each biting the orange, it was exactly the same performance.
That’s the moment I thought, “Holy crap, we can get this to work — and it is going to look great.”
What contributed to translating Rosa’s performance to Alita so well?
Mike Cozens: Because of the amount of dialogue and intimate drama Alita is involved with in this film, we had to spend a lot of time refining how the process works with phonemes ( a unit of sound that distinguishes one word from another in a particular language) and speech shapes — specifically, understanding how the mouth behaves. Understanding the ligaments and muscle structure on the mouth was really critical for communicating the subtle, yet really important facial performance details that make the character come alive.
Baneham: It’s all about the idiosyncrasies of every individual actor’s face. Biomechanically, our muscles are mapped similarly from face to face, but the order in which the muscles fire and the synapses that cause that fire are unique and different to every single face. Getting into those types of subtleties was really important.
Rosa has a very, very expressive range on a conscious and subconscious level. There’s the conscious performance she brought to the role, and then there’s the subconscious movement that happens in her face while she’s performing. There’s a huge amount of idiosyncratic behavior there, and while it creates a huge challenge to capture, that’s really where we mine the gold.
The single most important thing is not what a character is saying, but what the character is thinking. The mechanics of the phonemes are completely necessary, but it’s the inner monologue we read from a performance that allows our audience to connect with a character.
Cozens: There’s a moment in the film when Alita is cornered by Grewishka (Jackie Earle Haley), and he’s telling her some of her backstory, but she doesn’t want to hear it. There were multiple layers of things happening with Rosa as she performed that we needed to recognize and make clear on Alita. It’s those layers that draw you into the story.
When the first trailer and photos for the film were released, there was a lot of talk about Alita’s oversize eyes and whether they’d be too off-putting for audiences. It didn’t end up being a problem at all, though. How did you find the right balance of human and nonhuman characteristics with the character?
Saindon: It came down to seeing Alita with other characters. The balance came from putting Alita side-by-side with other characters and not looking at the size of her eyes, but the makeup of her eyes — the size of her iris and the amount of sclera you see. It’s about seeing the same proportionate amount of sclera in Alita’s eyes as you see in Rosa’s eyes, because that way, you know the performance Rosa gave is going to be the same for Alita.
Cozens: It sounds like such an inconsequential detail, but it’s critical in reading a performance. A performance is, in the end, all about these subtle details. It’s all about the relationship of an iris to a lower eyelid in millimeters. A small twitch on the face and a microexpression — those are the things that make the character come alive.
Hollander: Our job is to not lose any part of her performance. There’s magic in the artistry involved throughout, but the cornerstone is always the performance.
Cozens: Absolutely. The actor’s performance originates in a single moment in time. It’s like an inspired, improvised piece of music you would see on stage. The post-production and visual effects process happens over weeks and months with hundreds of artists, but our goal is to not step on or diminish the energy and spark from that initial performance in any way.
One of 10 films contending for a nomination in the Best Visual Effects category at this year’s Academy Awards, Alita: Battle Angel premiered in theaters February 14, 2019.
- How Dune’s visual effects made an unfilmable epic possible
- How GPU-fueled visual effects built Black Widow’s Red Room, then blew it up
- How visual effects brought The One and Only Ivan’s animal stars into our world
- How visual effects took Love and Monsters’ sweet creature feature to the Oscars
- How visual effects made Mulan’s battles bigger and its hero fly higher