Being Realistic: 3D Models and Modelers for VR Hoover Dam

Each artifact and environment in our game requires time and creativity from digital artists – specifically 3D modelers. In the VR Hoover Dam game, players will find themselves surrounded by objects they can see and interact with; it’s the 3D modelers who bring these things to life.

A photograph of Winthrop Davis sitting in front of his ham radio setup, and a closer photograph of a B&W radio transmitter. Reference photographs such as these are what our 3D modelers rely on to create our virtual world. As the cabin setting is one of the first that players will explore, the ham radio was an early 3D model which Michael Swartz has worked on.
Based on reference photographs, our modelers use basic shapes– polygons– to construct artifacts and environments throughout the game. Above are two early steps of the modeling process that our ham radios went through: from here, our modelers apply textures, double-check their sizing, and finally release the asset into the game.

What is a 3D model?

A 3D model is a three-dimensional object that takes up space in a game. Everything that a player can see in our game—from the cabin, to furniture, trees, keys, books, ladders, doors, fences, towers, trucks and other machinery—has been created by a 3D modeler. In order to design these items, modelers usually start with simple shapes like cylinders and cubes, or geometric coordinates in the form of edges and vertices. Modelers build using polygons– or “polys”– until they have a recognizable shape that resembles a door, or flower, or anything else you can see. Our virtual reality environments rely entirely on 3D modeling.

With a project that’s meant to feel as real as ours, and with as many historical artifacts as we have in the game, reference images are essential to our 3D modeler’s toolkit. Dr. Anthony Arrigo, author of the book Imaging Hoover Dam: The Making of a Cultural Icon, has been researching the dam for almost a decade, and his collections of photographs accumulated over the years are the starting point for the 3D modeling team.

One early example of a model that team member Michael Swartz worked on was Winthrop Davis’s ham radio. Starting from a family photograph of Davis sitting at his ham radio sometime in the 1960s, the team scoured the internet, eBay, and ham radio enthusiast message boards to identify each individual piece and collect as many photographs of them as possible. Swartz, a professor of animation at Northeastern University in Boston, then used those images to create 3D models of an exact-replica of Davis’ ham radio setup that the player will see in the game.

Most of Swartz’s design career has centered around photo-real assets, and so Michael especially helped to guide our team’s sense of a realistic virtual environment from the start. Still, Swartz notes that “at the end of the day, a project like this is never going to have enough research images. And so there has to be a bridge between what we can find, and then what we can use our imagination to fill the gaps with.”

Two shots of the back and front (respectively) of Wint’s cabin. To build this, our modelers needed to create the shape and proportions for one of the logs and individually then stack them atop one another to “build” the cabin. Every detail of the cabin down to the shingles has been scrutinized based on a reference image.
Early shot of one of the suspension bridges modeler Mya Ramirez worked on early in the project. The shapes and lengths of the wiring running alongside suspension bridges such as this were one of the trickier parts of this build.
Reference images for the cabin and suspension bridges.

Mya Ramirez, an undergraduate at the University of UMass Dartmouth, knew she wanted to work within the arts since she was little and was fascinated by the animation in her shows on CartoonNetwork. Mya, who is responsible for a number of models in the game including bridges, buildings, the water bags, the employment office and more, says that thinking about the coordinates, the geometry, and all of the logical thinking a 3D modeler is required to do in order to meld their work seamlessly into the virtual world can certainly be intimidating. But when you’re excited about what you’re creating, and you trust the team around you to teach you more while you work, “even the hard stuff feels like a helpful, fun challenge.”

James Ristaino, another undergraduate at the UMass Dartmouth working on the team, took a Virtual Reality class with Professor Michael Swartz which opened his eyes to his own passion for the gaming industry. Ristaino jumped from ‘noob’ to pro at shaping 3D models, UV mapping and texturing them, and importing them into Unreal Engine, the software which runs VR Hoover Dam.

Textures and UV Mapping

Another example of a 3D model that players will encounter in the game is Winthrop Davis’ cabin. The cabin in the game is recreated as closely as possible to his actual cabin in Toledo, Washington where Davis lived after his time in Nevada. The cabin is an incredibly complicated asset because it’s so big, and because players are able to view it up close. The cabin also has a lot of textures. There are textures for the inside and the outside, and all of the different smaller models that are attached to any given room: every little detail from the kitchen sink, to the wood floor, to the roof, the porch; everything we see requires a custom texture.

What’s texture? If you think about anything you see in a game, it has a specific color to it. Grass is green, wood is brown; but more than colors, the models in a game like VR Hoover Dam have textures that allow thin, shiny blades of grass to look different from the grainy, rough texture of bark on a tree. Using textures, a 3D modeler can “paint” their models to look even more realistic.

Modelers might spend well-over 100 hours on a single model, like Wint’s cabin. But by using a technique called “UV mapping,” modelers can design an outer layer of an object and then “wrap” it around a model to cover it with a texture.

Since we’re working with three-dimensional models, UV mapping is necessary to ensure that a model looks realistic from any angle a player might view it from. Think about a cardboard box: we want to cover it entirely with wrapping paper without overlapping or misaligning any edges of the paper. If you cut too much or too little paper, the overall design on the outside of the paper becomes slightly distorted, and while the recipient of our gift may be forgiving, a 3D modeler’s job is to make sure their textures line up perfectly. Otherwise, our realistic models can become distorted-looking, and disrupt a player’s immersion within a scene.

These shots show three bulldozer models in various stages of creation. The first image is constructed but not textured; the last two images portray ‘dozers with different levels of rust and dirtiness from use. It will be important to litter the canyon site with historically-accurate construction vehicles such as these, but also to ensure they are not merely copies of each other. Styling variations of the same artifact contributes to the immersive quality of the environment.

“Finishing” a Model

Overall, our goals for our models revolve mainly on their looking and functioning realistically in the game. When a player is able to forget that they are playing a game; when they get lost in the scenery looking out across the Dam, or pause to look around and admire where they’re standing, or pick up the camera that looks exactly like a camera from the 1930s, even when inspecting it extremely closely, then we’re a happy development team.

With any creative process, an artist can find a number of reasons to keep working on their art, given the chance. Our modelers appreciate the parameters they’re given to work toward. In this case, our team luckily has a number of historical photos to use to confidently design each environment. However, all 3D models require its modeler to balance a sense of realism with creativity. Finding this balance is the constant struggle with a project like this: we’re trying to be historically accurate in every step, but sometimes we just have to do our best and ultimately make it appealing and fun to look at.

Before digging into the canyon site environment, it was necessary for our modelers to map the coordinate locations of where gameplay would take place. Although our environment is expansive, the actual locations of gameplay are localized to certain areas along the cliffs.
Even Wint needs his own model! Thanks to several photographs of Wint at many different points in his life, we have been able to experiment with ol’ Wint: his hat, his beard, and even his voice are important decisions that our team has needed to make.

Beginning this March, VR Hoover Dam project has initiated its beginning stages of user testing. Check back for an update on our time spent with local Keith Middle Schoolers of New Bedford, MA, as well as an update on the feedback we gained when team member Michelle Turk (UNLV) introduced her students to the game.

We look forward to sharing these exciting next steps with you! Please reach out with questions or comments, and please follow us on Instagram for more news and updates.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s