How the 'Avengers: Endgame' Visual Effects Came Together
Jen Underdahl, the Visual Effects Producer for Marvel, provides a look at the impressive visual effects featured in "Avengers: Endgame" and "Infinity War."
Avengers: Endgame is the biggest movie of all time. It also had one of the biggest budgets. A movie this size takes a ton of people to put it together and make sure everything looks real and inspiring. To do that, they enlist the help of Jen Underdahl.
Underdahl is Marvel's Visual Effects Producer. It's her job to bring all the visuals of the movie together and make sure the directors and other producers love what hits the screen.
In this video from Wired, Underdahl walks us through the process of creating "Smart Hulk," making Chris Evans look considerably older, and how they were able to digitally create the entirety of the Avengers' time suits.
All pretty cool things!
So watch the video and let's talk after the jump!
How the Avengers: Endgame Visual Effects Came Together
When we wrote about Markus and McFeely, they talked about how they saw Infinity War and Endgame as two separate movies. But in this video, Jen talks about how the VFX for them both blended together into a single project for her.
Part of the reason this happened was that they had to spend a lot of time building and creating Thanos for both films.
You can't just have Josh Brolin show up and do his lines in a mo-cap suit. First, they had to do extensive tests to see what CGI could do for them. So much of what made Thanos real is attributed to the technical discoveries made along the way.
Jen brought WETA in to help them solidify what we saw on the screen.
WETA and Marvel decided they wanted to have Brolin on screen with other actors because they wanted this movie to feel real. And certain expressions and physical cues could only be achieved with Brolin working with the other talent.
So, not only did they have to build a digital form of him, but they also had to build a fake 15-foot suit so eyelines would match and so would his hand and feet movements.
The Hulk Melded Together
The same tech used for Thanos also was used to create the Hulk.
Motion capture suits were used to get Mark's facial expressions, but unlike other Avenger movies, this time they had to blend with the composite of the Hulk to create a more human version of the character.
Lucky for them, they've done the Hulk face so much it just meant redesigning the lines and grimaces they already had using Mark's unique facial characteristics.
In the end, that blend allowed them to be able to seamlessly drop the character into scenes that Mark could act in wearing the suit.
Old Cap and Hungry Thor
Things were not as easy when it came to aging Captain America. While making Steve Rogers less svelte was easy in the first Captain America, because they could just digitally de-muscle Chris Evans' body, this time they didn't have a body to work off because Evans is so young.
They had to bring in an older man who looked similar to Evans.
This person was his wrinkle double...aka skin double.
They then digitally took the wrinkles on that person's face and dropped it onto Evans' face. They also made his neck thinner and his body look thinner so he didn't appear to be a muscular guy in a costume, but a 100+ guy who has lived a life.
But not everything as CGI in these movies.
Thor's fat costume is actually a suit built for Hemsworth.
They stitched him into the costume every day, and then WETA digitally took out the stitches in post. So, all they did was smooth some of the skin.
This blend of practical and CGI is one of the reasons the movie felt so real. It's hard to tell what is computer generated and what's just excellent costuming and makeup. Bridging that gap is the next stage in the future of Hollywood.
I'm excited to see how Marvel will continue to push these boundaries in both visual effects and practical solutions.
Earlier in the week, we talked about how Industrial Light and Magic changed the cinematic world when they developed CGI and used it in Jurassic Park. Those computer-generated dinosaurs marked the beginning of a creative revolution where filmmakers could actually take what was in their heads and put them into their films (and with as much or as little realism as they desired).
I’ve always been drawn to movies that build worlds.
A unique setting can be a powerful storytelling tool and a memorable experience. Would The Thing feel as isolating set in a diner? That said, would creating such locations often seems out of reach for low-budget films. You’re usually stuck with sensible, easy-to-find locations. How many finales happen at a construction site?
There’s a reason for that.
But you don’t need a big budget to create compelling worlds. For Stalker’s Prey 3: A Predator Returns, our the villain lives in a lighthouse on a rocky island in the ocean.
Stalker’s Prey 3 is an ultra-low-budget, campy-fun, made-for-TV movie that aired on Lifetime. We didn’t have much time or money, but wanted to attempt something bigger and more conceptually unique than the traditional suburban stalker story.
The script was originally written for Falkner’s Island.
It's a real place with a lighthouse and a few buildings. When we couldn’t get permission to shoot there, we moved to plan B: creating the location ourselves. I identified the essential components for the story: a lighthouse, a dock, a shoreline, and an abandoned house—al isolated feels. Then, we set out to find their real-world counterparts.
Finding the lighthouse was the hardest part. Coastal lighthouses were accessible but surrounded by things we’d need to remove in post-production, which would be too expensive.
That's when we found the New London Ledge lighthouse, a stunning structure on a concrete block in the harbor.
The New London Ledge Lighthouse
It looked great but was a logistical nightmare with no water, electricity, or outlets for lighting and cameras. Luckily producer Andrew Gernhard found a lighthouse at the end of a long rock jetty, isolated and perfect for filming.
This was our lighthouse.
The exterior was great, but inside the lighthouse was too modernized. So I reimagined the story to place the keeper’s house and lighthouse on opposite ends of an island shaped like a barbell, connected by a rock jetty.
I revised the script to match this new shooting plan, re-working some action to account for the layout. There was no longer a grassy island around the lighthouse to hide on. The entire “island” could be captured in a single frame. The separation offered new opportunities: views from one spot to the other, a linear chase across the jetty, and the fun light from the lighthouse sweeping the keeper’s house once every revolution. Neat!
Next, we needed a keeper’s house. We couldn’t find a suitable one on an island, so we used a house in the city with an unobstructed view of the back across its yard. If we could place this house on our fake island, it would work.
With our anchoring elements set, we filled in the details. Swimming at the lighthouse wasn’t safe, so we needed a matching shoreline. I picked a rock jetty on the beach where we were already filming. Underwater shots would be done in a pool with black cloth backdrops. Boat scenes would be filmed at a marina that we were also using for other scenes, giving us easy access to boats. This is the sort of planning that lets a movie get more out of its budget, using locations for multiple purposes.
Planning angles was the next step. Whether using real or fake locations, the shooting plan needs to capture the action and also communicate geography to the audience. Think of it like taking pictures of the inside of a house intended to explain the layout. If you’ve chosen your angles right, after you show someone the pictures, they would be able to answer the question “how would you get from the upstairs bedroom to the kitchen closet?” Then, stage your story on top of those angles.
I thought about which angles would best explain our fictional setting and found the closest approximation of them that I could piece together from real-world spots. Of course, I found the version of that approximation that could also be achieved with our time and equipment.
Too often, even in big-budget films, we see big wide shots followed by tight, disorienting coverage that could have been shot anywhere. That’s not a good solution and it feels conceptually lazy. Story geography should be reinforced throughout the movie, especially if it’s important later. Story geography needs to be really clear because your audience is not actively trying to put it together. To reference my previous analogy, the person you’re showing the photos to doesn’t know that they will be asked a question after seeing them. The audience of your movie needs to understand the space as a side effect of watching the story unfold. In this movie, I wanted the space well understood so that when we get to the finale, we know what the hero’s options are she tries to escape, run, and hide.
Starting with the house, I decided the back porch would face the jetty and lighthouse. This minimized compositing. When the camera was on the porch looking toward the door, no compositing was needed at all. It was just a house.
When looking toward the lighthouse, we shot on the jetty, also without compositing.
With the camera on the porch looking at the lighthouse, we placed a green screen in the yard to comp the lighthouse beyond the porch, which itself obscured the seam where locations met below the frame.
We also needed to see the house from the lighthouse direction. For close shots, we filmed the house from the yard and cut it out, putting the real house on a virtual island.
For distant shots, we made a digital version of the house because there was no real-world way to get a sufficiently distant perspective on the house. I created a digital island and planned coverage at the lighthouse to show the house sparingly, avoiding a heavy VFX load.
We also needed aerial shots. This would help establish the big picture and also to match the style of the rest of the movie. To achieve them, I used the same digital house, shortening the jetty and placing the house at the new “end.” Drone shots from the lighthouse side kept the house away from the camera, making the island and seam between real and fake simple. I mostly used 2D pictures of rocks.
Next came the special stuff.
The movie opens with a 10-page shark attack sequence at night, where kids sneak onto the island, have a bonfire, swim, get attacked by a shark, and are rescued by the villain’s boat. This big set piece is both inciting incident and designed to establish the the geography of the location.
There’s a lot of detail about this sequence in my video, but the basic approach was the same. I broke down the things that needed to happen into places that we could shoot them. While we were allowed to swim at the lighthouse, it would have been dangerous and difficult for a whole host of reasons, so I opted to double the rocks of the jetty with another jetty on a beach we were using elsewhere in the movie.
For shots in the water, we used the ocean right off of the same beach next to the jetty, using the jetty for a light placement. We shot in a pool with black fabric backdrops to get some more controlled water shots, including underwater shots. And we shot at a dock to get the boat actions as well as climbing out of the water.
A dinghy was the target of the swimming contest. We hauled it to each location: the beach, pool, and the lighthouse. Lighting can help audiences understand geography. In this case, we put a spinning orange safety light in the dingy. The lighthouse had a spinning green signal light. These two colored lights were easy to spot, but also motivated extremely identifiable lighting. A character with green on one side and orange on the other would be identifiable as between those two objects in the story space.
Some minimal compositing linked it all up. I put the lighthouse into wide shots at the beach jetty and the dinghy into shots looking out to sea.
There are loads of other tricks too of course, like staging action to ensure your characters move actively between the key spaces to force the audience to link them. I really wish I had been able to do more of this in this project.
In it’s most basic for, for instance, you could show a host character walk from a kitchen to the front hallway to say goodbye to a guest. The guest picks up a coat on the way out the door, revealing the living room. And then the host walks to the bedroom, establishing that the windows face the driveway. Now the audience knows which direction the kitchen windows face, and that someone in the living room could see anyone coming or going.
The most important step is tracking your geography.
Na important note: you may have it all planned out, but the crew and cast won’t know the plan. Explain the artificial space to everyone ahead of time.
I drew a map and, at each location and in each new setup, explained where things were in the story space. I checked eye lines and light positions to ensure everything is properly situated. This one scene was shot across five locations, five days, and 13 physical configurations with different geographical cheats.
It’s important that everyone trust that you have a plan, so do your homework.
At its core, this is about simple filmmaking tricks adding up to build a coherent space. I’m not saying that this illusion was perfect or that it looked photorealistic, but it was understandable to the audience and let us do bigger and cooler things on a movie that otherwise would have had none of it. These techniques aren't tricks only visual effects professionals use. It’s work I was able to do my myself with commonly available software in the same production period as any other Lifetime film.
In prep, decide to “assemble” the location.
Then determine which angles to show and identify how to create them from real-world elements. Use eyelines, motivated light, landmarks, and other cheats to build as many links between your real shots and the points in your scene’s story geography as possible. Complete the illusion with some simple compositing, hiding seams and weak spots. Be flexible and revise the plan as needed.
Try to find ways to use locations for more than one purpose to get more on-screen variation out of fewer physical moves. Low-budget filmmaking requires adaptability.
Know your tools and be ready to move your goalposts.
Falkner’s Island image credit Greg Thompson/USFWS (creative commons attribution 2.0 license
The New London Ledge Lighthouse
This was our lighthouse.
