We stream high end result to hololens from Unity so imagine this through an AR HMD!
Very informative thread, thank you! Inspiring and reassuring to see the process behind such amazing renders. I started using UE4 a month ago, for a work in my firm and I'm still struggling with some concepts! I wish to see more article like this! Thibault (from France)
Your post is very good and meaningful. How long will it take you to write and post? thank you for sharing slither io
David Ballard has shared his breakdown of the multiplayer level design process using an example of Bill’s Town level from The Last of Us.
Bill’s Town Level
The level I will use to help demonstrate the multiplayer level design process from The Last of Us is Bill’s Town. It is a competitive multiplayer arena map originally designed and converted from a cooperative objective based map. It is a rectangular map consisting of 4 main building structures: Garage, Clock Tower/Town Hall, Church, and Office/Diner.
There were 2 main co-op objective modes: Hostage Rescue and Cart Push. Hostage Rescue entailed players starting at one base on the map and fighting their way towards the second base where an NPC hostage was being held. Cart Push required the players to push a supply cart from one base to the other while fighting off enemy NPCs. Coop was ultimately cut from the game and with limited time and resources, I had to edit the level designs to accommodate competitive gameplay.
Representation of the Player
The process begins with a representation of the player. It is the key to discovering the relationships the player will identify with in-game. For The Last of Us multiplayer, the level design theory was driven by a 3 lane concept. This concept states that between point A and point B, there are 3 lanes of traversal available. On a macro scale, this translates to 2 bases connected by 3 major lanes. On a micro scale, this means that at any point of cover, the player has 3 options of movement; forward, left and right. (Note: this was a departure from Uncharted’s additional vertical concepts). There were edge case exceptions, though III was the main guiding ruleset. Part of the fun of designing with this restriction was making the 3 paths vary in challenges, gameplay, strategy, art, and moods.
Narrative & Reference
The next pieces of the puzzle were narrative and reference. They informed context, environment shape language, scale, gameplay, technical limitations, and several other production needs. I also had already-developed single-player assets available to utilize as well.
Blocking Out the General Space
The next stage of the process is blocking out the general space. This includes defining major locations, traversal routes, scale, level boundaries, lines of sight, and landmarks. I sketched out a 2D drawing of the expected layout, which changed accordingly as working in 3D informed me of relationships not identifiable in 2D. Player movement speed greatly determined the distances between major structures. Additionally, I designed each of the 3 major routes to have varying gameplay; the middle street was open for long-range combat, Town Hall was mid-range, and Office/Diner was close quarters. Some thought was given to cover though I tried to remain in a macro mindset in the initial stage. Like a painter, I start with large brush strokes and move to smaller detail strokes.
After discussing potential production needs with the lead artist, gameplay goals with the lead designer, and schedules with the producer, adjustment decisions were made before diving into development.
Lighting needed to be brought into the process early on. It was needed for testing purposes and informed technical limitations and artistic decisions. Lighting also affected the design process by defining combat spaces and was important to creating the level by informing player experience. The texture artist needed the lighting implemented early as well to inform their texture and material creation.
The working relationship with the lighter would continue until shipping the game. It was a two-way lane of communication where both of us would approach each other with ideas, needs, and solutions. Lead artists, the art director, and anyone who had suggestions would give us feedback from art reviews and playtests.
As I refined the design, I remained conscious of artistic synergy while incorporating my art development knowledge. The level looked like a mixture of blockmesh and art elements, (Fig. 1). In coop, I had to contextualize enemy spawn closets the player couldn’t reach, (Fig. 2). The lead artist had an idea of adding a clock tower on top of the town hall for artistic purposes which helped with landmarking, (Fig. 3). When I converted the map to support competitive play, I wanted to add a bridge between the town hall and office/diner. Using the narrative, I added makeshift bridges that connected on top of a semi-truck container, (Fig. 4). Some of my decisions were made to support navigation via landmarking, like dilapidating a roof to see the clock tower when ascending an interior stairwell, (Fig. 5), or framing the church steeple from a piece of cover, (Fig. 6). I modified existing single-player structures to support multiplayer gameplay, (Fig. 7). And as I was also the environment artist, I designed structure shapes using as much modularity as possible in mind, (Fig. 8).
Towards the end, I addressed the wide, or non-playable areas including vistas. This couldn’t be addressed until the level design was finalized as only then were areas visible to players identified. These were also areas that could be artistically optimized since players would not be able to get close. Design-wise, I had to make sense of these spaces early on to support the believability of the environment. For example, addressing how a road curves around a corner and out of sight, justifying why there’s a road in the level and contextualizing the environment as a town.
Adding Assets to the Level
After numerous playtests, hundreds of feedback notes, and dozens of iterations, I got to set dress the level. Figures 1, 2 and 3 show varying distances to the map to illustrate how many individual assets were hand placed in this step.
Every section of the map had unique looks/assets which added to the difficulty of getting this level in memory. In single-player, each location would stream-in the necessary textures, however, they had to all be loaded simultaneously for multiplayer. Fig. 4 is an example of a unique set of assets in only one part of the map.
Both the texture artist and I would take turns vertex painting meshes to blend materials like pavement to grass and brick to plaster, (Figs. 5 & 6). A lot of the fun of set dressing was telling stories about the history of the environment, like placing bricks on the ground next to a building, indicating the direction they collapsed and a possible event that occurred to make them do so, (Fig. 7). It was also fun to demonstrate how plants grew in relationship to the human elements, (Fig. 8).
In conclusion, designing Bill’s Town was incredibly fun and rewarding. There were many lessons learned during development that informed the design process for the subsequent levels. Each location presented their own unique design, art, and technical challenges, keeping development fresh. Additionally, each level’s design got better and better, building upon the successes of the previously developed maps.
Effective collaboration with people from every department was required to finish these maps and ship the game. Input came from every corner, from art and design to QA and company presidents. Every person who interacted with me added value to the final product.