Mansion
Mansion is set in a bachelor’s residence in the mountains of southern California. It was often rented out to fans of a nearby music festival, but recent earthquake warnings forced a quick evacuation of the mansion.
Domination, Team Deathmatch, Capture the flag
4-6 players.
Pre-planning
Design Intention:
Create a fast-paced close-quarters multiplayer map that replicates the classic 3v3 faceoff game mode. it enables rotation and repositioning through ample cover and multiple paths.
I always like to start with a reference board of environments, architecture, and interesting locations to help inspire and guide my level layout and get a feel for the aesthetic.
First, I blocked out the play space and its surrounding area, thinking of flow and the time to engage. Then, I shaded in cover and elevations, drew out the main path and alternate routes, and plotted out the spawns.
After playtesting and building the level, certain rooms were too big or too open. This made players camp far away, unhappy, and/or the rooms simply never got used. Sorting some of the spaces while adding cover helped fix these issues.
The middle pool area is the heart of the map as players take control of the balcony overlooking it with still enough cover to hang out down below. It is the most open area and also the fastest route to traverse the map. I was pushing for the high-risk high reward feel that works well in objective game modes.
White Boxing
Once I established my references and layout in preproduction, I started my Whitebox in Unreal Engine using basic shapes to get an idea of the space.
When establishing my space for this map, I made sure to pay attention to the dimensions and measurements. I made sure to keep the distances between both sides equal. This way no team would have an advantage right out of the gate. Making sure the map was balanced was one of the biggest challenges I had. With each iteration, I had to make sure the starting spawns and the middle was symmetrical. I intentionally allowed the two side paths to be unique and asymmetrical to give the map flavor and make the space feel lived in.
Playtesting
Base Structures & Flow:
After playtesting, the map turned out to be a bit more complex than initially intended. The spaces were too open or too big. Some didn’t get used at all. I ended up cutting a good portion of the map but made sure there were still plenty of routes players could take.
Incorporating different playstyles:
I’ve been playing shooters for a long time so I know there are lots of playstyles. I found a lot of people preferred the shotgun and assault rifles. This led to some camping since players with the shotgun could keep the assault players out of rooms. To compensate, I made sure most places had counters to campers by having multiple paths or grenade holes. Also, while I left plenty of room for snipers, not all the sightlines were utilized. This could be because snipers cant really push objectives, and on a smaller map players want to stay mobile. There wasn’t a lot of room to stand still and hold angles since players can rotate relatively fast. This didn’t make them irrelevant though. Many players could still be effective as snipers.
Building the Environment
Once I was satisfied with the layout and flow of my white boxing phase, I moved on to lighting and set dressing with my selected asset packs.
Base Structures & Flow:
First, I focused on the base structures of the layout so I could start testing flow to make sure the walls and floors looked good with no awkward colliders or material issues. I especially paid attention to doorways since players would be running through them most often.
I used lots of different colored walls in order to tell the players which room they were in right away, rather than having the house all painted one color.
Filling in rooms:
From here, I went from room to room, filling them in one at a time. I did the first pass thinking about who lived in the room and what was its purpose as I placed objects. This started to bring my level to life.
On my second pass, I started to add detailed objects to break up space.
I did a final pass, where I added lamps and detailed lighting to each room and filled in spaces that needed an extra push.
This approach allowed me to make each room consistent and make the house overall feel alive.
Player Boundaries
Once the inside of the house was finished, I worked on a skybox and the rooftop of the mansion that players could see. I did this to avoid my level looking like a floating house in the sky.
Optimization
Optimization was an ongoing process once I began set dressing. Many assets were needed to achieve the aesthetic I was going for, and since I was using asset packs, kitbashing was the main struggle.
Collision
The first step was making sure the collisions on my new meshes were working as intended. Especially since it is multiplayer, there’s no point in optimizing if the players can’t use the space.
I noticed lots of meshes would be blocking bullets due to their simple collisions. Rather than making complex collisions, I substituted my own basic shapes that were set to “hidden in-game.”
I did this as well for places I didn’t want the player to go and to keep them from looking like they were floating. The invisible walls allowed me to make sure the players couldn’t get out of the map or into weird places. They also made sure the players could still shoot at each other without weird hitboxes that broke immersion.
Finally, I went and turned off collision on meshes that didn’t need it. This would ultimately lower build times and make sure players weren’t getting stuck on small meaningless objects.
Lighting
For my lighting, I really didn’t need to make it complicated. I used a directional light to cast shadows on the house and in windows.
Then I used the skylight with cast shadows turned off so I could lighten the whole place up.
I then added detailed lighting on lamps and other lights that would cast shadows.
Finally, I went back and made sure my shadows were showing up properly after building and added spotlights in certain situations to help.
Lightmap Density
Next was optimizing meshes from all the asset packs I used, and this work was badly needed.
A lot of the meshes’ lightmap density resolutions were extremely high with red, orange, and yellow spectrum. I toned them down accordingly to be mostly blues and greens to reduce lighting build times.
Also, I was mindful of areas where I needed the lightmaps resolutions to be stronger for better-looking shadows.
Take-Aways
Things I learned:
Asset Packs are not created equal:
I used several asset packs to fit the aesthetic I was going for. Certain asset packs needed to be increased in size to fit with my established scale. Others came with no collisions, and some just didn’t work at all. This was a particular problem when using doors, doorframes, and window assets. I kitbashed a lot and am happy with the results, but it was a struggle to find all the right parts while still keeping the mansion fresh and modern.
Scope:
I tried to design a 2v2 gunfight map. As I started building, I realized the technical challenges of doing a 1 life best of 7 round system with random weapons across multiple servers was a little too much for me in the amount of time I had. I pivoted to a 3v3 map which would better suit my original layout design. The increase of 2 extra people made the map bigger overall and took more time to fill in. However, it allowed for the gameplay to be more balanced and worked very well with the asset pack’s base functionality.
Multiplayer Blueprinting:
I created a ladder that initially worked in the editor, but not when I playtested in a build with others. This is because I needed to replicate it across the server to make sure it updated for all the player’s computers. There were no starting spawns in the pack so I had to create one so that players weren’t spawning in the middle of the map. The challenge was to disable all the other spawns until all the players get into the map and leave the starting area.
Measuring the space is important:
I started the spawns on each team at an equal distance from the 0,0,0 origin and then built the starting rooms symmetrically. Initially doing this saved me so much time down the road. I was able to write down and then reference the coordinates to make sure objects were spaced appropriately. Adding in the objectives was generally an easy process knowing no side had an immediate advantage.
Things I’d do differently:
Spend less time on the technical preplanning:
I tried to do the 2v2 gunfight initially which crashed my scope on the tech side with this multiplayer asset pack. I had to pivot and was still able to make a great product I am happy with. If I spent more time planning this level out instead of my technical planning it would be even better I’m sure.
Not download as many asset packs:
I organized the project a lot, but using fewer assets would’ve definitely helped my project. Certain packs I barely used certain assert packs, but I kept them in case I needed them. I hindsight, I could have eliminated some and not wasted time downloading and organizing them.
Things I’d Like to do if I had 6 more months:
The biggest things I would do are add more game modes and do more playtesting. There is no such thing as too much playtesting when it comes to multiplayer. The longer people play a level the more they figure out, and I designed this map to be played for hours. The main goal is to have the game continue to be fun when replayed. This is where additional game modes help out since many people like playing different modes, and it can help keep a level fresh. I designed for 4 game modes so far but only could playtest 3. I would like to have many game modes available on this map and see how players evolve with it and learn new techniques and tricks.
It should be noted that the asset pack designer is going to deliver several exciting updates in
June/July. The updates will allow for easier blueprinting and incorporation of more complex game
modes.