Mar 20

Star Citizen: Around the Verse – Level Design

In this episode of Around the Verse, published on March 16th, 2017, Lead Technical Designer Kirk Tome joins Sandi Gardiner for a detailed look on level design. Studio Director & Global Head of Production Erin Roberts reports on the UK’s progress in this week’s studio update. And stay tuned at the end for a very special montage of ship progress.


  • Sandi Gardiner (VP of Marketing)
  • Kirk Tome (Lead Technical Designer)

SANDI: Hello and welcome to Around the Verse, our weekly look at the development of Star Citizen. I’m Sandi Gardiner. Since Chris Roberts is busy at our Manchester office, I’m joined by Lead Technical Designer, Kirk Tome.

KIRK TOME: It’s awesome to be here Sandi.

SANDI: Our community team is at it again following up their appearance last week as PAX East with a visit to SXSW this week, and this Saturday the Austin studio will be hosting a panel as part of the festival.

KIRK TOME: The team will share details about the Evocati testing group, the importance of the Issue Council, and more! So if you’re attending SXSW you should definitely stop by.

SANDI: Later in the show the design team will reveal what goes into the modular designs for level and layouts in Star Citizen, but first let’s head out to Erin Roberts in the U.K. for our studio update. So, let’s take a look.

Studio Update

With Erin Roberts (Studio Director)

Hi everyone, I’m Erin Roberts. I head up global production at Cloud Imperium Games and I manage our European studios. I’m based here in our Foundry 42 Winslow office which is about ten miles south of Manchester in the northwest of England. We’re now at 201 people in this office and we have a further nine employees at our new small Derby studio in the east midlands, which focuses on mainly facial animations but also some body animation for Squadron 42 and Star Citizen.

Already this year we’ve hired 22 new staff and for a bit fun we worked the combined games development experience of everyone in Foundry 42, UK, which comes to 1510 years. We’ve just finished work on expansion of the first floor of Freedom house in Winslow which is given us back some badly needed space as everyone was crammed onto just two floors for the last few months and it wasn’t great. The new floor is definitely a breath of fresh air and has given us 76 new desk spaces, two conference rooms named Retribution and Gladius and a great new break area called Fortune’s Cross. Some of these names you know, others you’ll be introduced to as Star Citizen grows.

We’ve got a pretty big update for you guys so let’s to some of the details of what we’ve been up to this last month.

On the programming side, we’ve been working on systems to push the quality and immersion of both Squadron 42 and the Persistent Universe. We’ve completed sprint two of the player interaction system which improves how the player interacts with items or picks up objects using the new highlighting and inner thought systems. This will allow a much more intuitive and accessible UI experience for players clearly identifying what they can interact with as well as a clean, smooth experience while doing so.

The new mission system is moving really well, we are now on sprint three and designers are using the new tools to set up missions in the PU. The old flow graph missions which were not scalable to the need of our dynamic universe are going away to be replaced with a tool which can create diverse and systemic missions, giving the player an abundance of different and diverse mission types.

Also tied to this, the design team here is now also using the new Frankfurt developed system editor tool known as Soled, to visually put together our system maps for Star Citizen. Both of these tools will really increase the productivity of our design teams.

We’ve completed two locomotion sprints, the first to blend the walking to full run and back to walk animations sets to create a much more realistic feeling for player inertia whilst using the keyboard. The second sprint was to vastly improve AI path following so characters traverse closed spaces and blend between animations in a much smoother manner.

Our graphics team has been busily working on improving and optimizing the lighting in the game. One part of this was a major upgrade to the quality and accuracy of rectangular lights which is a feature the artists have been requesting due to prevalence of these types of lights in science fiction films. Typically, game support to the rectangular lights is very limited due to their high cost which is why we spent a lot of time optimizing our shaders to make them viable. The diffuse lighting and the reflections are now much closer to real world behavior and the difference this makes for our character lighting is absolutely huge.

On the networking side, the team is finishing off the serialized variable which will reduce network bandwidth for the PU. They finished the new message queue to make the sending and receiving of packets more stable and are finishing off the new multiplayer mega-map, so players can quickly traverse to different game modes without the long load times of the past.

Animation has been very busy also, weapon reload, firing, hand pose, select, and deselect work has been worked on for the P4AR, P8SC, P8AR, Devastator shotgun, railgun, Gallant and Arrowhead. Previous work has been done for the oxygen and stamina sprint as well as feedback on the female rig so we can lockdown final posing. Other work includes a no weapon locomotion pass update, stop to sprint update, and the prone combat animation pass.

Our facial animators in our Darby studio have been focusing on a lot of S42 work to bring our characters in story to life as well as work on the PU for 3.0 to support fixes, bartenders, shopkeepers and general population wild lines.

The concept team’s work is ongoing on the Aegis Reclaimer’s interiors. The team has worked on a second pass on weapons to improve reload visuals and add detail where needed and work has been ongoing on new ship weapons also. There has also been a lot of concept work for both our PU and Squadron 42 environments to give our artists strong direction on our planetary landscapes, habitations and landing locations but also for our space environments and space stations…And it is looking really cool.

Moving on to the environment team, has lots of ongoing work for Squadron 42 but has also started early work on the truck stop experience including the interior modularity to show the variety of locations we’ll be able to place in the PU. The team has been working to keep the art style consistent while also accommodating all the functionality required by design. The planetary service outposts are just finishing their initial art sprint and the base building set is complete. The team have all the elements needed to create small outposts in multiple configurations which are being set up so they can be distributed across different landscapes.

Now we have our building blocks, we can start adding the details which  give them flavor and detail. Also with the service outpost the team is developing how our shaders will react when we place these architectural elements in various biomes. We are looking into a system which will help give us believable systemic integration without having to invest lots of bespoke art time.

Lastly, the environment team has been investing time in creating the visual targets for our space look and feel. Not only do we want to add lots of detail to our locations, moons and planets but also we want space itself to be exciting and interesting to explore, whether traveling through anything from a nebula or dense asteroid field to a space storm or anomaly.

The visual effects team has been focusing on a lot of planning to support our new planetary environments, including atmospheric flight effects and modular procedurally generated surface spaces. Work has been done on thruster and damage effects for the Constellation Aquila, high tech damage effects library updates building on last month’s explosion template and also further polish to ballistic SMG weapons.

The UI team has been working incredibly hard over the last months putting together what you guys have already seen with the new front end interfaces from 2.6 and are still strongly plowing ahead with the needs of both the PU and S42. This month’s work has progressed on our new kiosk shopping interface. Proven out by our prototype which allows us to make sure it works in all our locations and shop types. Also work is continuing on improving all our in-game HUD UI where they’re walking around or on a ship.

The audio team, as always, are supporting all the sprints and tie into and support most of the work the other teams do. This month the standout tasks include fixing up performance issues and tool improvements. Audio for new ships included the Dragonfly, Connie Aquila, Prospector, and Buccaneer.

Work on the music composition for both Squadron 42 and the PU, speech processing, fixes to weapon audio and finally Foley work so the right noises could be heard from different material types.

Anyway, that’s it from the U.K. office this month. I hope you all enjoyed the update and it gave you some insight into how much is going on not only in this office, but all throughout CIG as a lot of these features we talk about are a collaboration between teams spanning sometimes all five of our studios. Hopefully you get an interesting glimpse of why the team is working so hard to create a universe and level of immersion never seen in a computer game before. This is truly why Star Citizen is the Best Damn Space Sim Ever.

Once again I’d like to thank all our subscribers for helping us put together these updates together and of course everyone in the community for your incredible support. You are all powering us on to make this groundbreaking universe and dream come alive. It’s really appreciated by everyone here at Cloud Imperium Games. Thank you all, and I’ll see you in the ‘verse.

Back in the Studio

SANDI: Thanks Erin! It’s wonderful to see all the ways the Manchester office works with other studios to improve the Persistent Universe.

KIRK TOME: Yeah, I can’t wait to experience all the different missions and locations that will be available as the PU grows.

While we’re on the topic of the expanding PU, designing levels and layouts for Star Citizens is unlike level designing for any other game. While most games only want to take a player from point A to point B, Star Citizen needs to feel like a livable place so traditional game design techniques don’t always work.

SANDI: Which is why we sat down with the Game Director, Todd Papy and Lead Level Designer, Andreas Johansson. Up next they’ll share how their level design process is unique to Star Citizen.

Behind the Scenes: Level Design

  • Todd Papy (Design Director)
  • Andreas Johansson (PU Level Designer)

TODD PAPY: So, today we kind of want to talk about our level design process and in particular towards the Persistent Universe and how that differs from what we consider traditional level design processes.

Both of us come from a very traditional level design process, which is you’re building everything very bespoke. You start with what the level goal is, where it starts, where it ends and then everything getting from the start to end is tailor made and the path is completely chosen and tweaked and tuned by the designer as well as the artists until the product ships.

With all of the challenges that we need to build and solar systems that we need to build in S42 as well as Persistent Universe, we couldn’t use our traditional experience from there. So, we started talking about a modular system and what that will allow us to do is build archetypes of things and then from there go through and start switching out different modular pieces. So, why don’t you kind of run us through like a truck stop or…

ANDREAS JOHANSSON: Yeah, our greatest challenge is how do we populate the universe that is the size of our game with enough content to feel alive, right? We have a hundred solar systems in the game. We might have 50, might have a hundred space stations and we have a hundred space stations we’re looking at 10,000 locations which we have to build. And with the small team we have, four level designers, it would take us about 650 years to build that the traditional way, which was…

TODD PAPY: It’s totally achievable in our lifetime.

ANDREAS JOHANSSON: I mean it’s a long-term job situation, right? [Laughter]

TODD PAPY: Yeah, exactly.

ANDREAS JOHANSSON: So, I mean the only way we can really do that is like Todd said is with a modular system.

We do build our locations with a tile set, which is small pieces of walls and corners and doors that we put together into rooms, but this is still not fast enough. We have to find a quicker way to do this. So, the way we can approach this is looking into grouping these smaller tile sets into bigger entities, rooms. We have kitchens. We have toilets. We have locker rooms. We have lobbies. We have everything that you can imagine that you need on a space station to…

TODD PAPY: …To make it feel believable …

ANDREAS JOHANSSON: …Yeah, exactly. So, instead of building every location unique we build these rooms we populate the big library of assets and then we use these assets to put together the station itself in a much quicker way.

TODD PAPY: Correct, but even then when we’re talking about a base level of an archetype or something like that. We’re talking about a very neutral feel and look to that so that when we start adding these modular pieces in a hub section or something like that, you don’t really notice you know like if, if those base assets are in that neutral sets then you really start to noticing the repetition and…


TODD PAPY: …And everything.

ANDREAS JOHANSSON: So when we build things we build everything using a template tile set which is a complete that’s was like a neutral. It has no textures and nothing special on it, and then we build out the basic shapes.

We define the purpose of the room, and then we can convert that into many types of themes, low tech, high tech, all the things we have. But we also have another level on top of that which is the content of the room itself.

So, when they do a game that is an MMO, almost all MMOs use some kind of modular system or some kind of tile set to build their locations. You would always see repetition at some point. You will go into locations and say, “Bob, I’ve been in this room before”, and we want to get away from this, so we’ve gone through many iterations of the modular system and try to figure out how can we alleviate this in the best way possible. How can we make sure that even if it is the fifth time you go through this very room it still feels different?

TODD PAPY: And we have a couple of different ways of doing that…


TODD PAPY: …Like Andreas said we have the tile sets, so what we would consider low tech alpha, bravo, charlie, as well as high tech and then we have supermodernism and a few other tiles sets that will still need to be built and worked on.

Then from there we have wear and tear associated with that. So, there’s dirt passes. You know, how pristine is this truck stop? Is it in the middle of Crusader, so therefore there is a lot of money that’s around that one so that one will be in very pristine condition, versus something out in the middle of nowhere, and it’s super run down, and it feels more like a mom and pop or Route 66 kind of truck stop or something like that.

Then from there we’ve got like what he was talking about with the actual props that go inside there and then we also have the different levels of power. Piping that can go in there, so we can make it feel more like a derelict or it’s up and running and it’s got a certain amount of hustle and bustle associated with it based off the AI, or what is happening in the solar system at that time.

ANDREAS JOHANSSON: Yeah, so if you go back to for example the props that Todd mentioned that is something that is a very useful tool for us to change up the feeling of a location. The way we’re moving right now is that we won’t actually create, let’s say, a bathroom and fill it with all of the props that it needs. What we’ll do is we’ll create the shape of a bathroom and we’ll fill it with area shapes and volumes which can spawn in assets in different constellations. That means that you might go into exactly the same bathroom in one station and it looks one way. You go into the same bathroom in another station and there is a different amount of toilets. There is a different amount of sinks. The mirrors are in a different place. It looks different.

So, by using these kind of systems we can get a great amount of variation from our smaller component rooms and make the stations feel unique and different even though you’re actually playing with a similar room which you’ve seen before.

TODD PAPY: In the process, in us building this, I think we really started on this about a year and a half ago. And at that point it was just Andreas and another level designer that was really focused on this. And now we have the other resources really coming in.

Meaning art is coming in and working on how do we do this from an internal perspective, but also from an external perspective on the space stations.  And how do we make these things have unique silhouettes associated with them so that when you’re coming up to it you get a very clear read of what you are coming too, and the size of it, and what is it mean to be. Is it meant to be a very large space station or is it meant to be a small, little space station out in the middle of nowhere?

Then from there we also have a tech artists/coder that is working with us to build our modular system. This will allow us to run this tool which will allow us take these different shapes and combine them together with unique points of interest in those. And then basically randomize it, run through it, play it and see what it feels like. Randomize it again. And then this will allow us to generate the stations as quick as we possibly can because we do have a lot of content that we need fill. It’s just building up these tools, and building up these rooms, and the pipeline to actually flip the switch and really pump these things out.

ANDREAS JOHANSSON: One of the processes we’re looking into is we won’t have a level designer sitting and dragging rooms in the construction of a space station. That’s not how we are going to build it. Because we have the libraries of all the rooms that we can use in a space station. We have all the procedural generation of the props that exist in the rooms.

The way we are going to build it is we are going to create a basic flowchart of a station which indicates where rooms are supposed to be. You start with a hub, you come from elevator and you go into a hub, you might go into another corridor that has like a locker room or has a diner attached to it.

And we build this flow in a scripting tool and then from that flow we can generate a seed and we generate the location, basically the editor will generate a location, based on that flow.

So, taking the seed we can generate many, many different space stations with the same flow that will look completely different because we get different types of bathrooms. We specified it needs a bathroom so in one location it gets a small one and then another one has a medium one.

TODD PAPY: …Different stores…

ANDREAS JOHANSSON: …Different stores. It has different props in the rooms when you go into the location.

So, a space station will always look the same but you can have another space station based on the same seed that looks completely different. So, it might have the same hub but they actually feel different because they have different props and they have different layout of doors and connectors.

So, with that a level designer could technically throw out 20, 30 space stations in a day.


ANDREAS JOHANSSON: But of course we have to go in…


ANDREAS JOHANSSON: …And double check all that information. We can’t just generate, ship to the Persistent Universe and it’s like “job’s done!”. That doesn’t work like that. We have to go in and verify the layout. So even though we can generate a large amount of seeds and a large amount of different stations we still have to do the proper work. We still have to go through, check the consistency of everything, see that it works, see that we don’t walk into a room and it’s a door into space and everyone has a very bad day. That would be pretty terrible.

TODD PAPY: We don’t know exactly how this will play out, I mean we have our ideas and I think once we’re fully through the R&D phase and can actually generate 15, 20 different truck stops we start seeing where the repetition happens and then at that point it will be trying to work out how can we cut down that repetition.

And then the other thing is since we’ll have so many different archetypes ideally the players’ not running into this area over and over and over again.

You know we’re still very early in the process.


TODD PAPY: We’re in white box phase. And really the artists have started to come in now and are breaking down these tile-sets for the satellites, for the planetary outposts as well as for the truck stop. Those are the three main ones that we’re focusing on right now and each one of them requires their own unique thought process on the modularity and how those things work. The overall idea is still the same: it’s just how do those things connect together is very different.

ANDREAS JOHANSSON: I just want to reiterate we’re not planning on shipping 50 space stations per day. That’s not possible.


ANDREAS JOHANSSON: It’s the possibility of generating the layouts.


ANDREAS JOHANSSON: The actual stations will take quite a bit longer because you still have to make sure that everything exists and everything works.

TODD PAPY: With this modular system this is really to build out what we would consider our lower or even mid-tier space stations: we will still have very bespoke layouts. So, for example if you think about Grim Hex or you think about Hurston or Area 18 these are very bespoke layouts that we go through and do handcraft. In the future I think all of us would like to see if this system would work in actually building a city and doing it procedurally but this is really to build up the other 95% of the content that is in the universe besides just these big bespoke landing areas.

ANDREAS JOHANSSON: The way work has evolved over time form being not only level designers but more kind of spatial architects where, when in a traditional game when you build a level you have to think about gameplay, you have to think about covers and flow and where the enemy’s going to spawn and all this kind of thing. This is not what we are doing when we are building our locations because we want this to feel real, and want it to feel believable. It should be a place where people live and work for months at time.

When we build these locations we have to think about how does rooms connect together? What is the flow? How would people build this?

It’s much more thinking about the space as a living area for people instead of a gameplay space. And that is why the role of a designer, especially the level designer, has evolved into being a little bit more than just gameplay orientated.

TODD PAPY: It’s less about path from point of interest to point of interest to point of interest. This is about making sure that these areas feel absolutely believable. And that you understand that there was a thought process behind creating this.

ANDREAS JOHANSSON: Also being a space game it gives a little bit extra because if we go to space in the future as the human race we would build things modular…modularly.  I can’t even say that word even though I work with it a lot. So, it is not that strange of an approach to take. You would need things to function between different stations for repairs and for expansions of other things.

You might end up going to a location in space and it’s a pirate base but you can clearly see at some point this was an old mining outpost that after the ore ran out it got transitioned over to a nightclub for mercenaries and then it got taken over by pirates and now you have this…

TODD PAPY: …Which is basically what Grim Hex is…

ANDREAS JOHANSSON: Petty much. It’s all party there.

TODD PAPY: This is actually where we’re at in the level design process and I know it’s take us a long time to get to this point but there’s a lot of research and development that have going into the process and how these areas are constructed and definitely hasn’t been a simple task.

ANDREAS JOHANSSON: Measure twice, cut once. Right?

TODD PAPY: Exactly.

Back in the Studio

KIRK TOME: I appreciate the amount of detail they put into designing each truck stop and space station, especially with the assistance of a modular system.

SANDI: I agree and it’s really going to help the PU feel authentic.

And that’s it for this episode of ATV, but before we go we’d like to thank all of our subscribers whose contributions allow us to make shows like this, Citizens of the Stars, Bugsmashers, and Loremakers. If you’re not a subscriber and are interested in learning more, click on the link in the description below.

KIRK TOME: Of course, Star Citizen wouldn’t exist without our backers so a big thank you for everything, we wouldn’t be here with you.

SANDI: No we wouldn’t. And please join us tomorrow at 12 Pacific for Star Citizen: Happy Hour. Jared Huckaby, Tyler Witkin, and Community Streamer, Gritspitter will playing Star Marine with fans in Alpha 2.6.1.

KIRK TOME: Thanks for watching and we’ll see you.

BOTH: Around the Verse!


Special Feature: Ships

ERIN ROBERTS: Oh wait, didn’t I mention? The U.K. makes ships too. Hope you enjoy!

About the Author:

Leave a Reply