From RMS Titanic over a hundred years ago to the groundbreaking visuals of Avatar a decade ago, and now the highly anticipated Alita: Battle Angel, the legendary James Cameron consistently pushes the boundaries of cinematic storytelling.
Interestingly, before the release of Avatar in 2009, James Cameron had already completed the screenplay for Alita: Battle Angel. However, due to the massive success and demand for an Avatar follow-up, the Alita: Battle Angel project was temporarily put on hold.
Alita: Battle Angel, a 2019 American cyberpunk action film inspired by Yukito Kishiro’s manga series Gunnm, was written by James Cameron and directed by Robert Rodriguez. This special sci-fi blockbuster, produced by Weta Digital, presents a thrilling vision of a future where humans and cyborgs coexist. Set in the 26th century after a cataclysmic event, the film explores the gritty reality of Iron City, where the strong prey on the weak. Dr. Dyson Ido, a skilled cybernetics doctor, discovers a discarded cyborg girl in a junkyard. He gives her the name Alita and helps her to adapt to her new life and the dangers of the city.
To stay true to the source material, Alita: Battle Angel not only showcases impressive “mechanical beauty” but also preserves Alita’s unique characteristics as a manga character. The film introduces cutting-edge 3D shooting technology, “Performance Capture,” used by Weta Digital. This approach allows for the creation of full CG character Alita with photo-realistic detail. VFX Supervisor remarked that the visual effects in the film are noticeably more advanced compared to movies like Avatar and Rise of the Planet of the Apes, and that Alita is the most complex digital character Weta Digital has ever brought to life.
Presenting Alita realistically in a cinematic environment required the innovative use of new performance capture technology. This technique, which James Cameron and Weta Digital Studio employed, builds upon the motion capture technology established in Avatar and enables the simultaneous recording of body movements and facial expressions. In a 1:1 scale recreation of the steel city setting, a multitude of cameras captured Rosa Salazar’s facial expressions and body movements. The system reconstructed the markers in 3D which generated a set of virtual skeletons within the CG model. Through this process, the most accurate power and action data were collected from her body to drive the CG model performance. Rosa Salazar’s performance was flawlessly transformed into Alita, with every scar, wrinkle, and micro-expression appearing remarkably consistent. James Cameron expressed that their aim was to ensure the audience perceives the characters as genuine in close-up shots.
Alita’s captivating big eyes and intricate skin
Weta Digital, which has previously produced special effects for The Lord Of The Rings and Caesar from Rise of the Planet of the Apes, found Alita’s facial expressions to be their most challenging and refined work. Her eyes are composed of 50,000 pixels, with the irises alone using 9 million pixels. The production team dedicated an entire year to creating these eyes, with four months devoted to perfecting them. For the first time, Eric Saindon and his team were able to replicate the fibers found in a baby’s eyes. The close-up shots of the eyes weren’t created using dynamic maps; instead, they completely designed the entire eye structure, incorporating elements such as the iris.
The texture of Alita’s face was similarly challenging. To create a realistic face, the team undertook an in-depth analysis of the facial structure and the different layers of skin. The team then incrementally built the head, jaw, muscles, supporting tissues, and finally, the “skin layer.” Over 2,500 facial expressions were designed and finalized to capture the diversity of Alita’s emotions, achieving a visually stunning and harmonious result.
47 hairstyles, 2700 components
To achieve Alita’s lifelike realism, the team developed 47 different hairstyles, with the eyebrows and eyelashes comprising over 132,000 hairs, 2000 eyebrows, and 480 eyelashes, and had nearly 500,000 “peach fuzz” hairs on her face. Engineers implemented a new pore growth technique, which involved implanting fluff into every pore of Alita to enhance the skin’s delicate texture and make the face more realistic.
On screen, audiences see a complex mechanical body. In reality, Alita is a simulation model composed of over 8,000 parts. The internal “mechanical skeleton” is made up of 5,000 pieces of various sizes. The production team also developed 71 digital special effects costumes for Alita, featuring a total of 125 diverse wearing methods. The Berserker’s body incorporates 2700 components, all moving at high speed with each movement of the character.
Approximately 500 million hours of rendering
The film’s Steel City was shot in Austin, Texas, on a practical set spanning 96,000 square feet. This sprawling city, with its elevated design, required an immense amount of computation during the rendering process. To streamline this, the production team chose to employ a relatively recent technology, the “OSM” caching system. This involved using 1 to 20 very basic building structures to define the city’s architecture, and then incorporating additional details such as “pipes.” Using the “OSM” caching system, the buildings could be cached once and then reused in different compositions. This approach reduced memory usage and conserved both time and resources. Each scene in Alita: Battle Angel took approximately 100 hours to render, totaling between 400 and 500 million hours in total (with 500 frames per frame), which is significantly more than Avatar. At Weta Digital, a team of 800 staff members and 30,000 computers worked tirelessly together. Notably, when the weather was cold, staff could even feel the heat radiating from the studio, a clear indication of the high-performance computing at work!
