Upcoming Events! Community Event Calendar
Social Media Follow us, or not!

Around the Verse: The Process of Procedural Cities Written Thursday 16th of November 2017 at 09:19pm by CanadianSyrup, Sunjammer and Desmarius

As per usual, anything said during the show is subject to change by CIG and may not always be accurate at the time of posting. Also any mistakes you see that I may have missed, please let me know so I can correct them. Enjoy the show!

TL;DR (Too Long; Didn't Read)

Burndown

Shopping/Cargo

  • 197 issues across all categories remain
  • All of the shopping bugs and tasks are dead with over 150 resolved, hopefully moving to feature complete status
  • Item 2.0 ships focus towards completion is starting this Friday
  • Triage for minimal viable product in the next two weeks begins
  • Like with all endeavors at CIG it's an international effort working in coordination
  • Dumper's Depot now includes Size 2 items
  • It's definitely been a challenging time by dialing in the details, and they've definitely encountered the Devil in doing so

Missions

  • Missions has moved from sprints to weekly reviews and continue to use tracking via Jira while still taking into account playthrough feedback and soon should be feature complete
  • Seven missions have been brought to final leaving four to complete
  • Automated turrets have been added to combat landing pad griefers as well as group mission participants have their own “radio bands” to help with mission security from others in the vicinity

Ships/Vehicles

  • Beds are being setup for persistence to be able to log out in them and come back to them.
  • Several ships that were once broken are now properly updated and fixed, Gladiator being one of them.
  • UI integration is going well and brings a bunch of improvements via item 2.0 integration.

Traversal

  • Complete pass on ships velocity regarding SCM has been completed to make them a bit faster.
  • Some bugs that affected the ships and performance were resolved giving a best of both worlds result.

mobiGlas

  • Starmap is in a great place now, but there's still some bugs to squash with some feedback being reviewed and adjustments being made.
  • Work on the inventory app for mobiGlas to make it more intuitive has been ongoing by Chad McKinney to enable players to easily sort through their inventory or their ships inventory to find what they want.

Performance/Stability

  • Many of the improvements that have been worked on for months are now finally coming together to increase stability and performance overall.
  • Evocati have had several builds pushed to them with more enhanced debugging analytics being applied in the newest build for better tracking.

Procedural Cities

  • Alex showed Jake a demo he developed creating Paris procedurally using Lumberyard
  • Over weekend Alex ported the tool to StarEngine and Jake provided Arc Corp assets
  • After about 3 days they had their first version of a city working
  • Then they added the Area 18 landing zone
  • Then they put the whole thing on Yela
  • Different city blocks are defined and laid out: low, high, industrial, commercial, etc.
  • Designers can edit the procedural city as required for art/logic/gameplay reason
  • We'll see different materials depending on planet, theme, branding, weather, etc.
  • Landing zones will have the highest density of assets
  • You won't be able to go everywhere - it will be logical
  • Over time they will be add more interactions, crowd systems, monorails, flying traffic, etc.
  • Jake would like to see some dynamic landing zones/missions pop up
  • The tool now uses prefabs so kits have lighting, VFXs, sounds, splines, etc.
  • They are currently just working on human cities but are looking forward to alien cities

Full Transcript

Intro With Chris Roberts (CEO, Director of Star Citizen and Squadron 42), Sandi Gardiner (VP of Marketing). Timestamped Link.

Sandi Gardiner (SG): Hello, and welcome to another episode of Around the Verse, our weekly look at Star Citizen’s ongoing development. I’m Sandi Gardiner.

Chris Roberts (CR): And I’m Chris Roberts.

SG: On today’s show we’ll take a look behind the scenes to explore the origins of our procedural cities tech and what went into getting it ready to show at CitizenCon.

CR: Yeah, and you definitely won’t want to miss it. It’s a really interesting feature, but first let’s check in with Eric Kieron Davis and the rest of the team for this week’s Burndown.

SG: (Breathless Whisper) Burndown.

Burndown With Jake Ross (Producer), Eric Kieron Davis (Senior Producer), Todd Papy (Design Director), John Pritchett (Senior Physics Engineer), Rob Reininger (Associate DevOps Engineer), Matthew Lightfoot (Associate Producer), Leo Vansteenkiste (Junior Gameplay Programmer), Matthew Webster (Associate Producer). Timestamped Link.

Eric Kieron Davis (ED): Welcome back to Burndown, our weekly show dedicated to reviewing the progress on the release of Star Citizen Alpha 3.0. Progress continues towards the release of PTU as we focus on the category shopping, missions, ships and vehicles, traversal, mobiGlas and overall performance and stability. As of right now we’re looking at 197 issues across all categories, so let’s check in with the team  to see how each are progressing.

Todd Papy (TP): We’re continuing to close out 3.0. Basically I think that we’ve killed all of the shopping bugs and tasks. We go to review today or tomorrow. So, at least by the end of this week and then from there Paul and myself and the UI team are focused on getting the Item 2.0 ships and all the UI that’s associated with that up to snuff and ready to go.  And so then, we’ll be ready hit PTU with that, and then from there we’re also focusing on PMA and some of these other features in order to basically close out all the Tier 0s. In the next week or two I will be triaging all of the basically 3.0 bugs that we have … that we’re currently focused on and making sure that we can refine that to as small of a list and as concise of a list of what we would consider minimal viable product to put that out to PTU and to live.

[2:14] Shopping/Cargo    

Jake Ross (JR): Shopping this week … we’ve made a lot of progress on shopping. We have a review schedule for tomorrow morning with Chris and all the directors to go over the progress. We’ve resolved over 150 bugs, since this past couple of weeks, and we’re going to review that progress. Hopefully get a minimal amount of feedback, so we can get super close to calling this thing feature complete.

So, it’s not just us here in Austin that are working on the shopping experience. You know we’ve got Spencer Johnson in LA that is working through bugs on his end, and then so we’re coordinating with him to make sure that he has everything he needs from us on that. We’ve also got just extraneous little issues like the NPC’s heads jittering and moving back and forth has to … we’ve got Rich Kay in the UK who’s working on that, and just lot of different, weird, random bugs at this point that we’re just trying to identify causes for and get knocked out just to make this thing super polished and super smooth.

Robert Gaither (RG): As people in our Evocati might have seen, we had a similar shop - by similar I mean an identical copy of Dumper’s - in both Levski and Grim HEX that we’ve now divided out and gotten two different shopkeepers in and kinda mixed things up a bit. We got Size 2 items now. The empty back room is now full of beautiful Size 2 items that we have on some new displays we got from Ben and his team. Also we’ve got all sorts of art that had rotations all over the place. We had to go back and make sure that everything was placed properly/looked pretty on the shelves, and that we’re having all the final stuff in the game or all the final assets in the build.

Robert Reininger (RR): There’s definitely been challenges going through this. Shopkeepers, that’s generally our first usage of the subsumption engine in general, so getting to know that has been a challenge. The shopkeepers, although what you’re going to see is fairly simple compared to the overall AI you’re going to see in the long run, is still just a good learning experience for us. Getting some of the other bugs worked out of the system, interactions with objects, getting the try on camera to sit at the right thing to focus on the items you’re trying on . . . things like that. Just kind of dialing in the knobs so to speak has just been taking its toll and taking its time so … it’s been good.

 

Ashram Kain (AK): Interestingly this week we’ve had to deal with a number of bugs related to how we do occlusion in the cargo grids. For instance if you were to store a ship in your cargo bay, how do we fill up boxes around that ship bay so that we don’t block you so you can’t keep the ship there, but still give you the option of storing cargo inside your ship? It’s been very interesting stuff.

[5:02] Missions

Janine Irmler (JI): So what has happened recently is that we transitioned from doing sprints to weekly reviews to close out the work for missions for 3.0. We already have a lot of missions available in Evocati, but they’re a few still missing. So, what we are doing is wrapping up the outstanding missions and tasks, which we track through Jira.

So the team is working through the outstanding Jira list and to keep it organized we define the focus at the beginning of each week as well as which missions we want to review, and we’re doing the review at the end of each week. In addition to that we are also doing daily things to ensure we stay on top of blockers or any issues that could arise, and of course also to track progress.

Additionally to the reviews we’re also doing playthroughs where directors or leads for example can experience the missions themselves, they can play through them themselves, and provide additional feedback to the team as well on top of the reviews we have.

So, this feedback ... we get all of the feedback, and it really helps the team to improve and balance the missions further. So this is ??? we have. We have made lots of progress doing all of these reviews and playthroughs, and now we’re just closing out the last bits so we anticipate being feature complete in the very near future.

Luke Pressley (LP): This week we’ve been focusing on taking the rest of our missions to final. We’ve got seven more taken to feature complete, and I think that leaves us with probably only about four more to take there, and they’re very close. So, this leaves us able to focus on bugs and the smaller features that we’ve been leaving until last. For instance, we’ve added automated turrets to some of our space stations that’ll better defend against people hanging around and kind of griefing - shooting ships on landing pads, and we’ve also managed to get it now so that when you join a mission you also join a kind of radio band for a sort of … that means like if you join a multiplayer mission you and your friends will be within just one kind of radio band and hear the dialog from that mission, and it won’t spill out to the rest of the server. Oh, and people nearby won’t like come across your mission dialog flying backwards and forwards as well which really neatens things up.         

[7:38] Ships/Vehicles

JR: So another feature I’m tracking is Item 2.0, UI integration, and ship setup so it’s getting all the ships setup and utilizing all the new features in item 2.0 support that the engineering team has implemented.

We’ve got L.A. tech design and U.K. tech design all working hard to setup the ships, getting the beds setup to so that you can get into the beds and log out and it saves where you’re out and you can log back into your ship.

We’ve got the lighting setup, the operator's seats, we’ve got a whole lot of different, too many to name at this point. It’s just a lot to do, but this week we’re hoping to make some significant progress there to be able to call a lot of these ships feature complete.

The Gladiator which was previously broken is now functioning, it’s working, it’s flyable. You can actually get in and out of it without breaking the animations and stuff at this point.

We’re excited about the progress the ships have made especially on the setup side, but also on the UI front as well, getting the UI integrated into all of the different ships and getting that balance is going to be a key part of the remaining effort for item 2.0

[8:59] Traversal  

Matthew Lightfoot (ML): So as we talked about last week guys, some of the focuses of the IFCS team was supporting the AI’s team with bug fixing and then spline following. So we’ve managed to close out quite a few bugs now. We’ve got one left which is going to extend the start of this week and before John and start on the spline following tasks.

John Pritchett (JP): So I’m out here at Foundry 42 Manchester so that I can work with Andrew Nicholson, David Colson on flight model. So helping with the tuning and helping David work on, he’s been working with me recently so he’s doing the work on the gravlev system and optimizing IFCS and things like that.

Andrew Nicholson (AN): So last week we managed to complete, or I managed to complete a full pass on the ship tunings for the SCM, velocities to make things a little faster. It took a long time to get through all those ships, but I’m really pleased with how that’s worked out. Gives us more velocity to play with especially when switching to the higher velocities with the afterburner.

This week the main focus is the gravlev simulation for the hoverbikes and atmospheric tunings and the drag simulation in atmospheric flight.

David Colson (DC): We discovered a little bit before we went to Evocati last week and we found that every so often ships would lose control. You would roll and roll to one side and the plane would start spinning uncontrollably and you weren’t able to recover very easily. We quickly discovered that some performance improvements that were added a few days earlier caused this issue, we weren’t exactly sure why, but we needed to release to Evocati so we removed the performance improvements which fixed the ships losing control, but we still want these performance improvements.

So I went back a couple days later, took a deeper look actually at the performance improvements and we discovered that I had made a mistake and there was some force couples in the way the ship rotates that weren’t balanced very well as a result of the performance improvements that when the ship tried to make certain maneuvers, it wasn’t balanced and it would just uncontrollably spin and not be able to get out. So that’s been fixed and the performance improvements have gone back in so we get a little bit more performance on ships and they don’t lose control anymore which is nice.

So we ended up getting the best of both world with this bug. We got performance improvements that make the game 3.0 as smooth as possible for players and we also fixed any issues with ships losing control which is the best possible outcome we could have got from this.

[11:38] mobiGlas

Leo Vansteenkiste (LV): Alright so last week on the Starmap it was quite alright, there were still a few issues. We had some cases where we lost the input, we had a few crashes, those have been fixed. The crashes were kind of hard to fix because they’re kind of random, they’re not reproducible so we really had to find in the code where things can go wrong and then try to fix them, send them into QA and see if they have been fixed and so far so good.

We also had Dot has been reviewing the mobiGlas Starmap. Checking for the last few details. He has a small list to send back with things we still need to fix, but really small polish tasks.

So with Starmap it’s the first iteration of the system and we would really like to have it polished as this now because it’s not the last iteration of the Starmap so we would really like to see people play with it and get some feedback on it and improve it and getting some feedback on what people think about it and seeing how people actually interact with the system.

Mark White (MW): So for the Starmap polish sprint, over the last week we’ve kind of been doing the last few bug fixes. On the Thursday or Friday of the last week we’ve got it to the point where we’re now happy with it on the UI side so that’s now, this morning gone over to Todd. Todd’s reviewed it this morning, he just sent over a list of I think five or six bugs so now we’re going to estimate those, get them into the schedule, see where they fit and then once those bugs are done again we’ll go back to Todd and hopefully that will be it and the Starmap will be completed, signed off, and ready for 3.0 live.

Chad McKinney (CMK): So right now there’s kind of two things i‘m working on right now with persistence. One is just general bug fixing. We’ve done a lot of work to get persistent behavior into the game so this includes persisting locations, persisting inventory, item port attachments and so on. Now that we’re getting these features into the hands of the backers, the Evocati were starting to see some very rare and hard to catch bugs come in and so I’m spending time tracking those things done like very rarely a certain ship might not show up in your inventory whenever you go to an ASOP terminal, things like that.

Also I’m finishing some out some feature work having to deal with persistence, mainly the deal with the inventory and the inventory app in mobiGlas so that the players have a way of looking at all of the items that they own and some useful ways of querying those items. So you can potentially have a lot of things that are... potentially hundreds of things attached to a ship, there’s all kinds of weapons you can have, there’s commodities, there’s lot of things in the game now that you want to be able to look at your inventory and be able to rummage through it and see okay this ship has this gun and this ship has this particular cargo commodity on it and I want to transfer it to a certain location so being able to be able to both track and modify the persistent values as well as query usefully into those values is something that I’m finishing out the feature work currently on.

[15:16] Performance/Stability

Ricky Jutley (RJ): So for the high level agenda today in order to go to Evocati. The kind of big ticket items that we’re going to run through with webster and the team are how much we’ve triaged over the course of the day. The fact that Todd and Paul have obviously worked with QA to be able to do that and then also have that triage is starting to affect how we’ve condensed our Evocati push into PTU push and putting those tasks and bugs together through a condensed list of what we need to deliver going into next week and also how we push the Evocati again today.

Following on from that Brennon will start running through the cloud of major issues that we’ve seen, the kind of stuff that’s blocking us from doing another push today as well as the real nasties that are going to affect the Evocati and for Will Everett to obviously let the Evocati know the general impacts we’re going to have and then we’ll actually get a discussion of the issues that we’re going to see. During the course of the day though QA have probably noted as being worth to raise and then following on from that we can jump into the player count meeting afterword.

TP: In order to get out to PTU we’re focusing on stability, more focusing on performance. Now obviously when we were released it out to the Evocati there was a whole different setup and a lot of times they’re seeing things we’re not seeing from our internal playtests or internal QA department because they’re going about things differently, different machine setups so on and so forth. So it’s kind of eye opening when we start seeing some of those bugs come in, but the guys have been focusing on those and trying to squash those as fast as possible.

Associate Producer Matthew Webster (APMW): Once again we made another push to Evocati on Friday. We had a bug report from Will sent this morning which seems to suggest that playability is taking a bit of a step backwards. So QA are going through the email that Will sent on this, responding later today with any bugs and tasks relating to what’s being called out there. We’ll also aim for QA to provide a crash report as well. We should highlight the main crashes that are affecting Evocati.

Steven Brennon (SB): We currently have 35 issues left on our Evocati fix version. These are bugs specifically that we want fixed. Around two thirds of them are crash reports that we’ve gathered from the public crash handler. So when the Evocati get a crash and hit submit, it goes through towards the nav QA via pubmap. So our main priority is getting stability up and we want to have a very stable build before we go to PTU and to  ensure things are looking good.

At the moment I am working on a method of prioritizing these crashes because we’ve been getting a lot in and we need to work on which are the worst ones so I’m working on a method of looking at the percentage of these crashes that have based on the total concurrency on the Evocati servers and once we’ve got that in place we’ll be able to get the worst of the wost fixed and then start working backwards from there.

SB: In summary then we have one issue that we would very much like fixed, that is the crash on maps and it’s looking like we can get the fix in for that we should kick off a build with that in and have that as our release candidate and if not we should kick a build immediately which will have muhammad's extra debugging information in it and that will be our release candidate instead, but that would be for ashrim to find out from Max, he has to find out where that’s up to. That’s the plan for tonight so fingers crossed guys.

Clive Johnson (CJ): In the last week or so we’ve not had that many bugs back, so the network team has been mainly focusing on optimisation and we’ve been tackling a couple of areas. We’ve been looking at what in the game code makes our marks or serialized variables dirty a lot is generally in there for generating a lot of traffic, trying to find ways to cut down the need for them to generate that  traffic in the first place.

We’ve been looking at ways of analyzing the traffic that’s coming in and seeing if we can call any of the traffic, reduce any of it so that we don’t need to send it and we’ve been looking at the network code itself, seeing if we can make it work a bit faster and that’s pretty much it really.

For us on the network it seems the combination of months of work, we’ve refactored a lot of the stuff, we’ve spent a lot of time optimising the code, trying to make it go as fast as we can and looked at increasing player accounts. 3.0 means all of that coming together for the first time in a kind of finished polished product and I’m looking forward to seeing how that actually works in live.

Francesco Dimizio (FD): as you may know a couple of months of ago we introduced this new technology which is called waff data which allowed us to reduce data completion times quite a bit, down to 10 minutes, everybody’s happy about it, but of course during the process we came across a bunch of bugs which was kind of expected of course and basically I’m going to talk you through on of the problems we’ve had.

So this file alob.dba dictates kind of like how our animations are going to make it into the build and in this specific case for example it means that these four animations here are going to show up in the build in this file alob.dba. Now the first build was fine, but then we realized when people were crouching these four animations either by like adding a new one, moving a new one or touching one of the assistant ones the build wasn’t really working and then we realized the new dba was being compiled fine because the content was changing, the dba was to be compiled again, but like was not being patched and it wasn’t being uploaded to the register.

So as soon as we realized this we fixed it and everything went much much smoother. We were able to close a bunch of Jira’s, like five, six for PTU right away, which we were super happy about it and we are confident that this is a good platform everybody to work on be it like technical point of view or designer point of view, we are super happy about it. Waff is being used for a couple month now, every branch is using it, super happy about it and yeah, it’s going to be a big help for PTU.

Christopher Bolte (CB): Good optimisation for the character on the server, like not updating NPC animations which unnecessary. Animations as is apart of ships, the player con is working mostly nicely with the network parallelisation. So basically with all the things that scale differently with the number of players and what the players do so we can just increase the player mesh slow, improve it slow and repeat and repeat.

On the other side we looked into infusing our CPU con which is supported because for Amazon

We get really powerful PCs to run it. With 76 and we’re testing with 72 cores, just have to check if we actually gain enough advantage from moving the CPU we’ve gone to.test soonish and then would see in the timeframe to release it soon because we don’t want to optimise for the next two years before we give it out, we have to find the right compromise to give it out basically.

TP: Now that we really have our tier zero’s, it allows the different departments to focus on basically closing this out because now we don’t have all the necessarily oh okay I have to code this for two weeks and come back or I go to design this and come back. Now it really is just tightening the screws down and making sure that everything is polished as we possibly  can make it.

EKD: Since last week, the team has checked in over 736 updates to the 3.0 branch in perforce in every one of those categories. We are making steady progress to get this into your hands as quickly as possible. See you next time here on Burndown.

Back to Studio With Chris Roberts (CEO, Director of Star Citizen and Squadron 42), Sandi Gardiner (VP of Marketing). Timestamped Link.

CR: You can follow our development processes in more detail by checking out the production schedule report on our website. Updates are posted every Friday.

SG: Now let’s turn our attention to this week’s feature. The team new that creating ArcCorp, a planet that’s almost completely covered in buildings would be a challenge ...

CR: Definitely.

SG: … as hand placing the number of structures required would be an enormous task.

CR: Yeah, with ArcCorp and the other cities that we need to fill out in the universe in mind Lead Technical Artist Alex Remottii built a tool that used tiles to generate procedural cities. These tiles could then be tailored to contain a variety of areas like whether the buildings were industrial or residential.  

SG: You saw this tech in action at CitizenCon, but it took a lot of work to get to that point. Let’s check out what went into getting it ready for showtime with today’s feature.

Procedural Cities With Jake Gainey (Lead Environmental Artist), Alex Remotti (Lead Technical Artist). Timestamped Link.

Jake Gainey (JG): Alex was showing me some tech. He’s quite passionate about some of the stuff he does so he’s always looking to share it with members of the team and with myself. So he was showing me this procedural Paris demo that he did, which was creating Paris procedurally through Lumberyard. And he mentioned that he was looking to incorporate that into our game, specifically to create ArcCorp at first.

Alex Remotti (AR): Yeah.

JG: And straightaway that was like “Wow! Okay, we can do something with that.” I knew, having worked on ArcCorp in the past, I knew we had assets available that we could use and plug into that system straight away. So I said to Alex “Listen, let me go away this weekend and just work something out for you. I’ll ...” He gave me the template. And he gave me the Max files - the source data - and I just went in there with the assets that we’d made before and just tried to get something together for him on the Monday.

AR: On Monday I got his assets and there were just some very, very slight changes to do to the bevel, to the scale, to the precision of the things and I ported the tool during the weekend to the … commercial Lumberyard to our own Lumberyard. It didn’t take that much effort because obviously it’s the same engine. And then in about three days we got a first version of the city working. He’d worked a lot on illumination, pollutions …

JG: Yeah, yeah.

AR: In other cases we set up the scene and did video with the real Area 18 inside the city …

JG: Yeah, yeah.

AR: … and send it off to Chris …

JG: ‘Cause that was important wasn’t it? So we built those city blocks. We built the city but then Alex was saying “Well it would be cool now to actually get Arc Corp - the Arc Corp that players have been running round in - put that onto the surface of this thing -  put it in context - because previously people in Arc Corp you were looking at the backdrop and going “Okay, that’s a cool backdrop” but now we’re looking at the backdrop and it’s a city.

AR: Yeah.

JG: It’s an actual city.

AR: With the right scale. You can fly through it. It has trenches. It has landing areas. They’re not still detailed but they give you a glance of what actually a proper city is like. And it will let you put that city on a real planet. So the other thing I did after is - we didn’t have Arc Crop - we just put it on Yela with an atmosphere …

JG: Yeah.

AR: … and flyed around. Zoomed in. Zoomed out.

JG: Yeah.

AR: So that’s where the idea of the demo comes: to show to people that actually this is a real city surrounding our headquarters.

JG: Yeah Alex, he took the city, and he went “Okay I’m just going to put this on a planet now” …

AR: Yeah.

JG: ... and did a video just flying out …

AR: Yeah.

JG: … into the atmosphere …

AR: Yeah.

JG: … of the planet. And it was like “Oh wow this thing … “

AR: It worked!

JG: “... this thing actually works!” No that was … that was really good fun to work on that kind of early stage.

AR: Yeah.

JG: For me anyway it was good. And then our first video that … so I … all I did really was I just gave Alex the assets and then you went away and put it into your procedural system. And the great thing about the system is it’s not like you press a button and you come back five minutes later and the city’s built. You actually see it dynamically …

AR: Building and plotting ...

JG: … building a city.

AR: … the city down.

JG: So there’s all these assets that I’ve given Alex flying at the screen building this city.

AR: Yeah.

JG: Dynamically.

AR: Yeah.

JG: And that was the first video that you did.

AR: Yeah. Yeah, yeah.

JG: Capturing that process.

AR: Yeah. You can actually ...

JG: That was such a cool process.

AR: … walk in the nothing - like in Matrix - and you see the city coming to you …

JG: Yeah.

AR: … and building around you ...

JG: It’s kind of like …

AR: … on the fly.

JG: … fantastic. It’s really cool. And then, yeah, then you’re on the landing pad in Arc Crop looking out at it and …

AR: Yeah.

JG: … you did another video where you’re in - I think you’re in the Arc Crop headquarters …

AR: Yeah.

JG: … looking out the window there.

AR: With the city coming.

JG: With the city building in the background. It’s a shame the players don’t see that.

AR: Yeah. They will not.

JG: Of course they won’t see that but it was a really cool process.

Wai-Hung Wan (WHW): So we have city tiles where we define a specific block, I guess, if you look at a city block we’ll say “Okay here we can have high buildings, low buildings. High rise flats.” And then from those collections of city titles they get arranged into a city. The procedural system will define high and low areas; industrial areas where it’s more ragged, more dense; and then downtown areas where it’s more flat areas but high rise skyscrapers. And it’s my job to make those high and low areas interesting.

So the point where we’re generating completely procedurally that’s where it’s all done by the tool. I’m hands off at that point. Well I’ll be there doing subtle - very subtle - edits. So you may have … one example that was pointed to me this morning was we can have an industrial area but the art direction may be “Oh, but can we have four chimney stacks over here” so that would be a very subtle edit on my part. But everything else: that’ll be procedural.

You’ll want rulesets in the procedural system but there may be times when a ruleset simply isn’t enough. You really don’t want, say, a hospital next to an oil refinery. That’s probably too close for comfort. So you’ll want to be able to separate those two areas to some degree. And that’s the subtlety of the editing.

But if you’re looking at thousand thousand cities potentially you might have to accept a strange arrangement.

AR: So we have … it’s more like teams - Alpha, Bravo, Low Tech, High Tech - and depending on the planet, depending on the theme, depending on the branding, and depending on the weather you will see different materials, different kits and different conditions. But that’s more on art direction.

JG: Yeah. Absolutely yeah. Is it similar to the outposts where you can have … if it’s there for a set period of time?

AR: Yeah [inaudiable] set time.

JG: Yeah, yeah. Great. Because that was really cool the way you could bring in more dirt. And I imagine Hurston would be …

AR: Very sandy.

JG: … very sandy ...

AR: Yeah.

JG: … and mucky and things. And maybe Arc Corp’s a little cleaner …

AR: Polluted and wear …

JG: … well polluted.

AR: … wear down.

JG: And then you go to Microtech - New Babbage - and it’s just pristine.

AR: Yeah. Exactly.

JG: Perfect.

AR: The level difference - we have two different resolutions. One is FPS - first person resolution - which is 512.

JG: Yeah. Like landing zones specific.

AR: He did some research and when to … which one was the space station?

JG: Oh, Olisar.

AR: Yeah.

JG: So the way we build our exterior content is such that you … we’re building these huge, almost kilometer long, stations and within those stations you’ve got landing pads. And the focus of the artwork and the texel rate of the resolution is around the landing pads. And that’s where the highest density of assets will be as well. But then obviously because these stations are so big we can’t put the same amount of detail across the whole station.

So it’s a very similar approach that we’ve done with the city stuff. You’ve got the city as a whole and then you’ve got the landing zones within that city. And that’s where we concentrate the highest density of artwork is around those landing zones because that’s where the players will be spending the most amount of their time. That’s where they’ll be landing. That’s where they’ll be knocking around with other players. The rest of the stuff? They’ll be flying around it. They’ll be moving across it in a much quicker way.

AR: Over time we intend to add, with a certain logic, locations and kits of the cities where you can land and have NPCs, have shops, have interaction marker points. But that obviously with the logic - you can’t literally come here and land with a helicopter everywhere. There are places …

JG: Yeah.

AR: …  where you can go, …

JG: Yeah, yeah.

AR: … places where you cannot go. You’re encapsulated. You walk out and go in some parts and also driving in some specific parts but not everywhere.

JG: Yeah.

AR: Because everywhere doesn’t actually make sense.

JG: It would be like trying to fly a helicopter over London …

AR: Yeah and just land …

JG: And landing anywhere!

AR: Yeah.

JG: I think the military would have something to say about that maybe.

AR: Yeah.

JG: It’s a good approach to it.

AR: Then … now we have some hangars - locations - that we … we do initially for the cities in the headquarters …

JG: Yeah.

AR: … where we have NPCs, you have customs, and you have interactions. And then over time we will add …

JG: Yeah.

AR: ... more interactions, more crowd system, monorails, flying traffic …

JG: A lot of subsumption.  

AR: Yeah.

JG: Yeah.

AR: Yeah.

JG: There was some stuff I was talking Alex about. I don’t know this will happen but I always imagined if you were flying over Arc Corp - the procedural city - you’ve got that landing zone - Area 18 - which everyone knows and visits and goes to quite a lot. But then every now and again one of the top of these skyscrapers, you just see one of the landing pads open up - you see a dude standing on there and you can land on it and go talk to him and there’s a mission from there and then you can go off again. But the next time you fly across Arc Corp it might not be there - it’s opened up somewhere else.

AR: So once I ... the tool has been submitted - is part of the build since a bit more than a month - has been only used by me and Stefan in DE. Working on Lorville to refine it and define how we were dividing our assets and what we needed. So we went from only using graphic content like meshes - CGF - to using prefabs so every kit can use lighting, VFXs, sounds, splines, spawn points, ATCs, and every kind of entity that you need to have an alive, interactable entity.

WHW: We are just working on human cities at the moment. I would love to see how we tackle alien cities. Is that going to be completely random? Are we going to have some or a greater degree of refinement by hand? I don’t know yet. I would hope even on an alien civilisation they have some degree of control and they would make logical, intelligent choices about where they would place specific buildings - even recreational facilities - so each time you visit that location it will look exactly the same as you left it.

You could potentially shoot someone there and if the undertakers haven’t come along the body may still be there. Pock marks. Broken windows. No I think a crime scene sounds interesting. I’d like to see that develop. I’d like to see the whole timeline evolve as you go away. It could be very well that you see buildings decay. Paving slabs deteriorate over time. Pot holes in the roads. That … that would be something to look forward to.

Outro With Chris Roberts (CEO, Director of Star Citizen and Squadron 42), Sandi Gardiner (VP of Marketing). Timestamped Link.

SG: You really get a sense of how many tiles are needed to build out a location like ArcCorp when watching the buildings get generated in real time.

CR: Yeah, well this tool is essential for achieving the massive scale of cities that we want in the game. As you saw though, the team still needs to go in and hand craft some areas, and the ability to easily generate large cities and adjust them to our liking is a huge benefit going forward in building out the universe of Star Citizen.

SG: It sure is, and that’s all for today’s episode. Thanks to all of our subscribers for your support. One of the many shows you help produce is Happy Hour, which you can catch tomorrow at 12:00 PM PST. This week Jared will be joined by xenolinguist Mr. Britton Watkins and Cherie Heiberg to discuss the Xi’An language.

CR: Yep. That should be pretty interesting, and a big thanks to all our backers for helping bring Star Citizen to life. Developments like the city procedural tech are only possible, because of your support and interest in crafting a universe unlike any other. So, thank you for allowing us to push the boundaries.

SG: Until next week we’ll see you ...

SG & CR: … Around the Verse.

CanadianSyrup

Director of Transcripts

A polite Canadian who takes pride in making other peoples day brighter. He enjoys waffles with Maplesyrup, making delicious puns and striving for perfection in screaming at the T.V. during hockey games.

Sunjammer

Contributor

For whatever reason, this author doesn't have a bio yet. Maybe they're a mystery. Maybe they're an ALIEN. Maybe they're about to get a paddlin' for not filling one out. Who knows, it's a myyyyyyystery!

Desmarius

Transcriber

When he's not pun-ishing his patients or moderating Twitch streams, he's at Relay pun-ishing their patience as a member of the transcription staff. Otherwise, he can be found under a rock somewhere in deep East Texas listening to the dulcet tones of a banjo and pondering the meaning of life.

"If you can't do something smart, do something right." - Sheperd Book