‘A massive 5 million polygon scan of the actual cave in Thailand was brought to life so we could extend our practical sets’

The underwater, cave builds and invisible visual effects of ‘Thirteen Lives’.

The rescue of members of a junior football team and their assistant coach from the Tham Luang Nang Non cave in Thailand gripped the world in 2018.

For Ron Howard’s Thirteen Lives, which tells the story of the rescue, in order to depict the cave environment, surrounding location and other key moments–such as the emaciation of the boys after their long ordeal–the filmmakers looked to visual effects to help enhance what they had filmed practically.

The visual effects team, led by VFX supervisor Jason Billington, used scans of the actual cave system for particular scenes, meticulously tracked the bodies and lights of divers to help re-create particulate and lighting for underwater shots, and worked to show the impact directly on the boys being trapped in the flooded caves for days.

Billington describes the work, in particular the VFX involving water, in this befores & afters interview.

b&a: Water of course was something so important to depict in the film—what were the different kind of water sims or 2D water solutions needed for the film? What was the toughest water kind of shot you had to achieve? In particular, how did VFX come into play for underwater diving scenes?

Jason Billington: About 50% of our VFX work was to create the underwater environment, flow, and particulates. These underwater sequences were some of the most complicated scenes to create. There were a couple of physical sets built for shooting that included a replica of the actual underwater tunnels.

The replica tunnels were placed in a water tank and filled with water so the actors could swim, move, and dive as if they were in the actual cave system in Thailand. The underwater sets were also built with moveable walls to allow cameras and crew to manoeuvre around within the sets. However, it was still a very tight spot to be, and only limited crew were allowed in the underwater sets.

VFX would use the plates shot on set, and add water flow/current, murk, and particulates. The particulates being dirt, small rocks, leaves, moss, and other random bits which would be found in a cave. These variations of debris were used to help visually show the current, but also the dangers and lack of clarity of the water they were diving in. We could dial up or down the amount of each ingredient based on the requirements needed within the story. One sequence could be quite murky and turbulent as we see in the first dive from Rick and John. Or it could be a lot calmer and slower, as seen when the Thailand Navy Seal, Saman, unfortunately dies from losing his breathing gear.


Due to the nature of a film crew shooting underwater consistently, the plates were very inconsistent with the clarity. One day the tank could be nice and clear, and the next day be very murky, and un-usable. While shooting, we tried to work with these conditions, shooting wide when it was at its clearest, and shooting close in when it was murky. Due to this inconsistent water visibility, VFX needed to not only visually control the speed and flow of the underwater in VFX, but also balance out the clarity of the water across the sequences to maintain continuity and the flow of the story.

With over 350 underwater shots, we had to be smart how we used our tools. To help the particulates and murk be correctly integrated, we needed to not only do camera tracks on all the shots, but also body and headlight tracking on the divers within each plate. The headlight tracking in particular, was key to the entire integration and look of the particulates and murk. By tracking the divers’ lights and body, we were able to simulate the particulate movement and flow around the diver, and also use the headlight tracking for lighting and comp integration. The only lighting within these tunnels was from the headlights, so the debris needed to be mysterious, appear and disappear into the darkness outside of the divers’ lights.

In addition to the flow and debris simulation, we also had very specific shot based 3d particle simulation. For key hero shots within the scenes, we would use the body tracking from the performance of the actor within the plate and generate very specific particle simulations. This was used for any extra debris or collisions from foot placements, or collisions with the cave wall or floor, kicking up dirt and debris from the contacted surfaces, enriching the emersion and claustrophobia of the environment.

b&a: How did you tackle realistic digital cave builds or extensions — how did the VFX team use real photography, scans, photogrammetry etc for these?

Jason Billington: One of our most significant above water VFX requirements included a full realistic CG rebuild of the Thailand cave main entrance chamber. As seen in most of the photos we all recognize from the actual rescue a few years ago. The main entrance chamber (known as Chamber 1) was huge. The interior chamber of the cave entrance required a complete CG rebuild to match the actual cave. It’s an enormous and grand cathedral like introduction to the challenge that lay ahead for the rescuers.

Environment build.

To achieve this grand scale, we had two practical set builds on set. One set was the exterior cave opening, the other set was the other end of the chamber, the smaller tighter entrance to the underwater tunnels. VFX was tasked with rebuilding the entire chamber 1, to not only extend the set walls up higher, and put a lid on the ceiling, but also to connect the cave entrance with the tunnel entrance sets to show a seamless chamber 1 environment.

A massive 5 million polygon scan of the actual cave in Thailand was brought to life so we could extend our practical sets. We kept as realistic as possible to the actual cave, including the detail of stalactites we see in the chamber ceiling, and the wetness drips that are continuously flowing around the entire chamber. In addition to the cave environment, VFX also needed to add volunteers and people moving around, both at the bottom of the chamber 1 cave as well as the entrance. Adding crowd to the cave involved a combination of shooting extras on a green screen, as well as CG crowd generation.

b&a: Can you talk about what was filmed and what needed to be augmented for scenes outside of the caves where ’operations/headquarters’ were based (especially for aerial views of these)?

Jason Billington: The opening reveal of the cave was a very complex composite that required CG renders of the cave, plate extensions, multiple plate passes, dmp projections, and CG assets. On set, there was a partial build of the cave entrance and stairs for the film crew and actors to work and film around. VFX then combined the partial cave entrance set piece, with a CG set extension, and blended it with multiple mountain range plates shot in Thailand.


The interior chamber of the cave entrance required a complete CG rebuild to match the actual cave. It’s an enormous and grand cathedral like introduction to the challenge that lay ahead. VFX recreated the entire chamber 1 using a massive 5 million polygon scan of the actual cave, and we then brought it to life with detailed stalactites & atmospheric elements like moisture and continuously dripping surfaces.

A particular aerial shot of basecamp was one of our biggest single shots on the show, looking down on the exterior basecamp. It needed to show not only the terrain and conditions the rescuers were located within, but also the sheer volume of people that were involved in the rescue. This shot involved stitching together two drone takes from two separate shoot days about a month apart, as well as adding in about 1500 digital extras, tents, vehicles, trees and terrain modifications. This shot took almost the entire post schedule to complete. Getting the crowd to feel full but natural in the way each and every volunteer would be reacting to the news of the boys being found required many iterations. The crowd also needed to feel organic and have a natural movement to the way they walked or ran.


b&a: The work is full of so many great invisible effects, but are there any particular VFX shots that you think audiences may be surprised to find out were VFX shots?

Jason Billington: I hope the audience has no idea that VFX were used in this film. Realistic, seamless, and invisible was the goal of this one. However, I think the VFX shots that may surprise people the most might be the emaciation shots of the boys. Particularly the first time when Rick and John finally find them on day 10. We needed to make this shocking to the audience yet, realistic.

We did not want them to appear too skinny that they looked like they would have been in a prison war camp. They had no access to food, and limited access to water. During this time, starving people see their metabolic rate slow down considerably, decreasing by 20-25%. They lose stores of key minerals, including phosphorus, magnesium, and potassium. Due to this period of no sunlight and no food, VFX needed to help them appear hungry, sick, and a little shocking to the viewer.

We worked with the makeup team, to add shadow & dirt in key areas of their faces, body and arms. In post, we further enhanced the onset makeup and augmentation with digital paint over techniques, often applied with 2d tracking, and additionally VFX would digitally pinch or warp in the key areas of the face and body to enhance the emaciation look, picking up on the natural bone structure (cheek bones, collar bones, ribs, etc), and color correcting the skin to appear as if it had lost its colour(sallow skin).

Leave a Reply