As is with any information on our transcripts and summaries, everything posted is subject to change by CIG and in some cases may not always be 100% accurate at the time. While we strive for accuracy, mistakes do happen so please let us know if you find something amiss that we didn’t catch. Enjoy the show!
Sandi Gardiner (SG): Hey everyone, welcome to Around the Verse, our weekly look inside Star Citizen’s development. I’m your host Sandi Gardiner and joining me this week is Chris Roberts, welcome back.
Chris Roberts (CR): Thank you, good to be back and just in time vamp it up for halloween.
SG: Yes we did, and you’re visiting both U.K. and Germany offices correct?
CR:Yes I was catching up with our two Foundry 42 offices in Wilmslow up by Manchester and in Frankfurt, and I was also overseeing some performance capture for 3.0 along with Dave Haddock who actually did a lot of directing on some of the 3.0 stuff, he did a great job.
SG: Very cool, and we’ll actually be discussing how performance capture goes from the stage to in engine later in this episode, but before we do let’s take a quick look on what’s going right now in the verse?
CR: Okay well the Super Hornet free flight just wrapped up. It’s pretty cool to see the first impressions from new people who test out Arena Commander and the Alpha PU during these free flight events.
SG: And a big welcome to all of the new backers joining us and thanks to all the veteran citizens helping them get settled in.
CR: Yeah thank you very much. They make our community special.
SG: They do. Looking ahead to our next release, 2.6 from all our studio’s they’re checking in stuff everyday, care to share?
SG: In the hotseat.
CR: Pressure. Well okay we’re playing Star Marine in internal QA everyday. I had quite a fun session actually with Tony Zurovec, Erin, Todd Papy in the German QA team last week when I was in Frankfurt. We have map and game mode adjustments based on internal playtest feedback we’re working on which were applying of course. We’re also working on network synchronization to reduce lag and make these smoother client remote animation. In addition we’re making sure the structural props and glass work properly over the network. Helmet 2.0 which is item 2.0 implementation of the helmet, it’s just gone in with the new hud and widgets.
Combat emotes have been hooked up as well and we’ll allow you to trigger them from the keyboard to allow you to signal your friends, taunt your enemies without having to break your FPS flow.
SG: There you go, and on the Arena Commander front, the designers are pouring through all the feedback we got from the last Evocati build and are taking another pretty big pass based on some of their feedback and some of our thoughts together.
CR: Yes I think people are gonna like what we’re doing. We’re going to go quite a bit further on the next iteration of flight changes and in addition to that we’re working on missiles to give them some much needed attention to give them more variation and personality and how they behave.
SG: Very cool. Missiles are now persistent between respawns and we’re working on the gamemode pickups to rearm or refuel mid combat.
CR: And the UI team is hard at work on the new front end for Star Marine and Arena Commander. So there’s a lot of work there, that one’s going to come in very hot. Of course we’re getting the new ships in and tested for both Crusader and Arena Commander. I know you’re all eager to get your hands on 2.6. Don’t worry, we’re working on finishing up the content and features so we can release 2.6 to the PTU and then live.
SG: Now let’s head over to our Austin office for a Studio Update.
Jake Ross (JR): Hey guys, Jake Ross here, Producer of the Austin studio and I’m here with you this week to talk about what’s happening in Austin. Hard at work working towards 2.6, 3.0 and beyond. First I want to pass it off to Emre Switzer, Lighting Artist here in Austin to talk about lighting updates for the Star Marine maps.
Emre Switzer (ES): So with the lighting in Star Marine, we’re not specifically looking to highlight the players, one of the elements of Star Marine’s FPS is we’re trying to build a more tactical gameplay experience and so while in other first person shooters you can use lighting to contrast the player and make them more visible, with the lighting in the Star Marine, we’re looking at it from a very realistic standpoint.
So we’re saying “OK, there is this pirate system which has been abandoned for quite some time, pirates have come in” – all the original lighting fixtures probably has been, to some extent, destroyed or they no longer functional. So what would the pirates do? How would they go in and light this environment for their purposes? So there is a lot of temporary lighting, there is a lot of lighting that’s not what was originally there – a lot of flickering lights, there is a lot of very moody lighting.
This again means that players are going to have to be a little bit more cautious when they are entering a new playspace, they are going to have to cover their teammates, they are going to have to be a little bit more tactical in how they approach these situations – they can’t just go in there guns blazing and hope to survive. Echo 11 has a lot of these different rooms, connected by these tight intricate corridors so each of these rooms has their own unique feel – maybe they are owned by different factions or ran by different people – but each of their own rooms has a distinct feel about it so it’s also going to help with the teamwork element because you can give callouts pretty easily.
So you “Oh, I’m in this blue room with all these crazy monitors” and all your teammates will know where you are, so I’m really excited for Star Marine – I can’t wait to get in there with all my friends, create a squad of six guys and run around one of these maps and shoot up the enemies.
JR: Alright, so now we’re going to pass it off to the QA Team and the Game Support Team and talk about testing for 2.6 and working with the Evocati.
Justin Binford (JB): 2.6 will bring about a lot of changes and new content, some of which include a refactor to our lobby system, some new missions, an addition to GrimHex and also the new ships – the Hoplite and the Herald. QA has also been heavily been focused on testing ship model changes which includes a balance pass to shields, weapons and missiles. This has been a heavily iterative process where QA will provide feedback, design will make changes then QA will test those changes and provide additional feedback.
Each day we have a live release sync for 2.6 for where we’ll get together, discuss any new blockers critical issues, any outstanding issues and also get an update of the build.
Scott McCrea (SM): Alongside everything that is going into 2.6, we’re also heavily focused on Star Marine, myself and my colleague Bryan Croker are both FPS specialists that work on Star Marine all day, every day. Basically what that means is we utilize checklists, use something called ‘Ad-hoc testing’ which is kinda exploratory testing. One of my favourite ways of testing would be the cross-studio playtests and basically what that does is word goes out that we’re making this cross-studio playtest, everybody from all the studios come in and play against each other, after the playtest we all get the feedback, get the bugs, I take all that information, put it into a single area, collate it down and send it off to people to who need to know about it.
We’re very excited to get Star Marine into PTU soon, that’ll ultimately help us get one step closer to the live release.
JB: There is a lot of new stuff coming in 2.6, some of which, like our flight model changes, are already in the hands of our volunteer Evocati testers which have been a huge help.
Will Leverett (WL): So the Evocati Test Flight Programme is a programme that we started in february of this year and it’s been hugely successful. It’s made up of 800 volunteers right now – they are from 20 different countries, they speak 18 different languages and they cover- they are from five different continents around the world – it’s actually pretty cool. So, what Evocati does is take a build that is out of internal QA but let’s just say it’s not ready for primetime, it’s pretty busted, it’s unfinished, it’s broken and a lot of times the people who play are masochists because they crash a lot – a lot of bugs – things aren’t quite exactly as fun as they should be before they get out to a wider audience.
Benjamin McMonnies (BM): The Evocati process is a pretty difficult process, we’ve vet all them through an algorithm that we use to make sure that they’re active and dedicated in testing through the Issue Council. We also double check them ourselves to make sure they’re not just adding contributions randomly and it’s pretty hard to get into the Evocati Test Flight – you have to be quite dedicated and everyone there really is.
Eric Green (EG): Working with the Evocati has resulted in an enormous amount of feedback via the issue council and via the PTU website, we’ve just had to pour through a massive amount of tickets provided to the Issue Council and they are miraculously not duplicates. They find just an enormous amount of bugs that we can actually find and fix and make sure the servers are more stable when we move over to our Wave 1 PTU process – in addition to that we’ve been provided so much feedback on the gameplay that we’ve actually had to find creative ways to go through the sheer amount of feedback that they’ve been providing so that way we can distill it down and provide it to the producers and designers and they can make changes to the game.
WL: Thinking about some of the successes of the Evocati – I think back to 2.3.0, if you remember that was a long time in waiting, and once it got out it was terrible, it was just crashy, had lots of bugs, and literally was not fun. But also wasn’t ready to go out to the public – that was- I think it was a fun time for the volunteers because they had a common bond that they all went through.
But the reality is that it took a long time to get that build ready before it went out to the PTU and ultimately to the public. So that was a good month of pain for the volunteers but they hung in and they gave us lots of feedback – we’re able to get that stuff fixed and out the door.
JR: OK, thanks guys. Now we’re going to hand it off, finally, to Jason Ely in the Backend Services Team to talk about the updates they’ve got going on in their world.
Jason Ely (JE): Lately we’ve been working on a number of things, we have Iain who has been working on- well he built our service recovery platform which allows the services to easily connect to each other- identify and connect to each other and now he’s been working on admin and customer support tools. While Tom has been working on a lot of things, he’s our #1 firefighter and he also our leaderboard and matchmaking modifications for the upcoming patch. I’ve been working on- let’s see, so far it’s been the hub rewrite that’s been my biggest chore lately – that’s been quite the ordeal, it took about a month to do from beginning to end.
Much of what we do in the backend is never really seen directly by the players but- in rewriting the hub server, much of what we did was I took time and simplified the threading model, and converted it over to support our new architecture – obviously, which is how we handle messages and routing which greatly reduces the amount of memory moves we have to do from one network port to the other network port.
We’ve also removed some of the state and isolated some of the state into one thread which makes communication a lot easier. Performance wise – it’s nothing that the player is going to know – it’s more of a throughput issue from an overall larger perspective. We can move more data to more clients without exhausting the machine and running down.
Originally, it was the first service written, actually by Tom Sawyer back a number of years ago and it was the most reliable server at the time – it ran like three and a half months solid with no problems. Over time, before I got here, it grew and it went through the growing pains – it had too many cooks in the kitchen. Some complexities that were introduced that were unnecessary and so it didn’t grow gracefully and so when I came on we started the new service architecture which brought us persistence and GIM – we always knew that we were going to have to address our hub server but it was so- it had grown into something it wasn’t meant to be, it had a lot of state, so it took quite a bit of time to address and so we just didn’t have that opportunity so we saved the best for last.
Tom Sawyer (TS): The Arena Commander and Star Marine UIs are getting a major facelift this round and we’re really trying to polish the UIs as much as possible and make them more fully functional. The thing I work on the most is the C++ code that the UI system interact with the server backend and there’s information in lobbies that need to be shared between all the players when you are in the lobby – what map they are in, what game mode, the character loadout for Star Marine, or their ship selection for Arena Commander – so I focus on propagating that information and syncing it up between all the clients. I kinda consider ourself like a second generation matchmaking system.
It’s where you form a lobby or a group of players or a party of your friends and you can chat with each other but at some point you hit ‘match me into a game’ and at that point then your request goes to the backend servers and the match makers take a look at all the game instances that are currently running and tries to do a best fit match for your group to join a game.
If there is no game open, obviously, then the general instance manager will pull a server from the open pull and assign you to that and then other players can be matched to that over time but a third generation matchmaking system is what you see in Battlefield or Call of Duty where you just get this massive super-lobby and you see players come and go, being matched into it and you see a timer going from two minutes to zero and if it’s got enough players then it just chucks all the players to game instances.
We’re more like Overwatch or Heroes of the Storm, where you form your group first then you match into a game.
JR: Alright, thanks guys, that’s all we have for you this week. Thanks, and see you around.
SG: We’ve mentioned the backend servers and lobby system a number of times recently. What makes these updates so important for 2.6 and beyond?
CR: Well we’re really focusing on the user experience so we’re trying to make the ability for players to connect with their friends, to play matches, customise their ship or their character much easier and quicker as well as check out the standings and do it all from the same place so it’ll just be a better experience.
SG: Speaking of connecting players, it’s time for this week’s Community Update.
Take it away Tyler.
Tyler Witkin (TW): Hey everyone, Tyler Witkin, Community Manager in the Austin, Texas studio here to bring you this week’s Community Update.
First and foremost I hope you all had an epic Halloween and I wanted to give a quick shoutout to Kinshadow and RedFang87 for their Star Citizen themed pumpkin carvings which were posted on our website earlier this week, very cool stuff.
Just a reminder the Esperia Prowler art contest is still going on. We’re going to be accepting submissions through November 11th and announcing the winners at the anniversary livestream on November 18th. We’re really excited to see what you guys come up with.
In other news, Star Citizen has been nominated for most wanted game for the Golden Joystick Awards. There’s only one more day to voite so if you want to follow the progress, you can go to Gamesradar.com/goldenjoystickawards.
And now it’s time for this weeks MVP. I want to give a huge congratulations Lundfoci for his extraordinary efforts in creating Starship42.com. This is a one stop shop where you can go in and explore the ins and outs of all of our ships in a detailed 3D viewer. I encourage you to check it out and you’ll quickly find you’re losing hours of your day exploring this website. So congratulations again, Lundfoci, you’re this week’s MVP.
Lastly the week would not be complete without Reverse the Verse so make sure to tune in live at 10am Pacific Daylight time at twitch.tv/cigcommunity where we’re going to talk about everything you saw on today’s episode.
Thank you all for your support and we’ll see you in the verse.
CR: Thanks Tyler, awesome stuff. For more on the latest, make sure to check out our Community Hub on the site, there you’ll find our social media feed, featured forums posts, links to Star Citizen streamers, information on up and coming Bar Citizen events and much more.
SG: Lots of stuff. On the top of the show we talked a little bit about our most recent performance capture session. P-cap is a big part of making the Star Citizen universe come alive. Recording voice, face, and body all at the same time gives us a nuanced performance from the actors that we couldn’t get any other way, even if the headcam gets a little uncomfortable.
CR: Yeah you would know that it can be uncomfortable.
So anyway, it’s a great tool, but one it involves a lot of attention to detail to get right. So it’s not just capturing someone, it’s about understanding the precise metrics of what they’re interacting with. So the animation will work together with props and the sets will look natural in pretty much any situation.
SG: For more on that, let’s go back to our Austin studio for behind the scenes look at how we create all these animation assets.
Brian Brewer (BB): Today we’re going to talk about animations and how we get them from concept into game. And the process at which we do that is Design comes to us and says “Hey we want some thing” so we say “Okay, how can we do this?”
One of the things that we have to think about whenever we shoot motion capture or something that we need for the universe are metrics. How … what are the measurements? Are … is this character interacting an object? Is this character interacting with other objects? Is the character interacting with the player. There are a lot of thought processes that go into this to figure out exactly what we need.
For example, if a character is setting at a mess hall we need to get the metrics of that mess hall table. We then have to go out on stage and we then have to build out that mess hall table to approximate dimensions that you seen in game. This leads to a standardisation across all mess hall tables in game so we don’t end up having 20 different mess hall tables that we now have to … instead of 50 animations we’re now doing 150 animations in order to accommodate every mess hall table that we have. Or bed or chair or bench or locker. You name it! Whatever you can interact with there’s a metric for it and a standardisation.
So we go out on the stage, we build out these props or set pieces, and then actors are given very specific directions on how they are supposed to approach these objects. These are also part of the metrics. For example, when you want to sit down on a mess hall table the actor has to approach that mess hall table and their left foot has to hit at approximately 100cm out, or 1m out, from the bench. Once they hit that mark – which actors are usually pretty good at hitting marks – they can do whatever they want to sit down onto that bench. And that’s where we get the fun acting that you see on the NPCs.
David Peng (DP): Hey guys, I’m working on Squadron 42. These are assets that are going to be in Squadron 42 but eventually we’ll be able to implement these into the Persistent Universe.
We had to do a couple of things, just as Bryan said, we had to make sure that all the tables are now to a proper metric that we can reuse the performances. Now that we’ve got this stuff in place … it’s really cool because … and these obviously, these are blocked, these actually come from one of our more organic cinematic, so even though maybe this one wasn’t intended to … we weren’t intending to use this for end game, there’s always an opportunity to look at some of that stuff and go “Oh you know what I can extract and make that something that has some reusability and we can apply this to the Persistent Universe.”
So here’s another one. This is the hungry guy comes up and is waiting in line for the mess hall. So this is all from and initial shoot … we actually have another one that’s coming in that has a lot more stuff in it.
BB: Once we get the data back. The first thing we have to do is we have to go through and have to analyse all the data. We gotta make sure the data handed to us from the mocap stage is clean and that there’s no problems with it. We then go through a process of usually cutting the animation apart into different segments. When you use an object, like you want to sit down in a chair, you have to break it into at least three segments: an enter, so you actually sit down onto the chair; a cycle, where you are cycling like I am right now on a chair; and then an exit where you actually get up and you exit away from the chair. And these are what we call animations sets. These animation sets are very important – to have a complete set – in order to have an AI fully be able to go in and use an object.
And these animation sets are definitely expandable. We can … we don’t to have just one cycle, of the guy sitting on the chair, the guy can be like leaning on his hand or scratching his neck or picking his nose or adjusting his earbuds or whatever you … whatever we can come up with that character can do in that chair we’ll shoot or animate or create. Whether it’s motion capture or whether it’s hand key. And then whenever it goes into game code will sit the character, the AI, down into the seat and then randomly play these actions that you see.
Vanessa Landeros (VL): Hi, my name’s Vanessa Landeros, and today I’m going to show you what we’ve been working on. And this guy will look a little funny because he doesn’t have his textures on his hands, but he’s basically this is what you’re going to see when the background characters are in their beds sleeping or fussing around. Like tossing and turning here and there. Little fun stuff.
We want to make sure we get a real experience, as real as we can get in the game, so we have these little minute details which you probably wouldn’t see in any normal game, but this is what makes Star Citizen so real and and we’re working towards something really, really believable.
And then the NPCs are going to be able to get in and out of their beds too. So they’re going to be able to walk in, lay down … and the player will be able to get to do this too. Just we need to make sure we get ins and outs, also some of every angle so we can capture those possibilities, the character if he want’s to enter from the front he can enter from the front – all those need separate animations.
Going to be able to see a janitor in game use a mop and a bucket. He’s going to clean up after all the Star Citizens. Potentially this guy is going to have a walk cycle of his own – he’s not carrying his bucket right now – but you’ll just see him drudge a bucket around with his mop. It’ll be fun once we get this guy up and going. Just that little extra detail that makes the game as cool as it is. So – put it on fullscreen for you there – when he’s done with his area he’ll just pick up his bucket and he’ll go.
BB: So when we pass these animations off on Design, Design has something called a “usable” which is an object that’s placed in the universe that has these little entry nodes and attach nodes. And we have usables for pretty much everything that you can interact with. AI will say “I want to use that chair.” So Design will have place that useable node on a chair, AI will navigate to that chair, align its … by the time it gets to that chair it will have aligned itself to the entry nodes, attach, sit down and use that object until it’s ready to get up and leave. And while that object’s being used other AI will be like “I want to use a chair. Well that one’s taken, what’s the next available chair?” Then they’ll nav to wherever the next useable chair, or they’ll wait, or they’ll change their mind and they’ll go off some place else.
Sometimes we get requests that we don’t have motion capture for or we don’t have time to go shoot motion capture, or maybe the motion capture state is tied up doing cinematics for us at the time, and it will be something that we have to get out right away. So what we’ll do is we’ll hand key whatever that motion is. Sometimes it’s temporary and we’ll replace it later with proper motion capture. Sometimes it’s not super significant and so we’ll go in there and we’ll just hand key it and make it look really nice.
DP: So we started working on PAW tool animations. High work zones, low work zones and then we have floor work zones. So with this character we’ve got … we’re going though and he’s still work in progress so we’ve got transitions that we need to clean up, things like that, but here you can go through and you can see a full performance. We’ve got finger animations. This is actually one of the performances where we didn’t have time to shoot the mocap so we’ve actually gone through and hand keyed these. So this is the high work zone. Just so you can see there’s a little bit of difference: he pulls out his PAW tool and flicks it. And what’s cool about this is that this isn’t a unique, one off animation. This has high reusability. You can use this on ships, you can use this on buildings, you can use it pretty much anywhere.
For weapon’s vendors we wanted it so can actually pick the weapons up, take them off and look at them. So from third person this is what you would typically see. So it looks like a character picking up and examining it. Some of these we didn’t have time to go through and mocap or we just needed something … some things we realise after shooting “Oh man we need this but we gotta get this thing turned around”. So that a situation where I’ll have to go in and I’ll have to … I shoot video reference. I’ll get a general idea of what we need – here’s where I’m just checking a weapon – and then this is the hand keyed result of that. So this is tailored to … for first person for player.
BB: I’m going to say the … probably the most challenging aspect of all of this is consistency in whatever it is that we’re doing. For example, if a character needs to go and use a datapad and go around and inspect things with that datapad, that datapad becomes a metric. And it become standardised. And that all datapad in the game now match this in particular datapad. It’s letting Design know that when a character is going to sit down and use mobiGlas, that animation can be used across the board and that everybody’s aware of those animations and so nobody’s re-requesting things that had already been finished.
VL: And that’s just a taste of what players will see in the upcoming releases. Hope you guys enjoyed.
CR: That’ll do it for this episode. Remember to tune in tomorrow at 10am PST, 5PM GMT For Reverse the Verse.
SG: And we’d also like to thank all of our subscribers as always. You guys make it possible to produce all of this content from ATV, RTV, Bugsmashers, Loremakers and more. Thank you very much for your continued support and guess what? We are in the Guinness Book of Records.
CR: Yeah look at that.
SG: We’re in the records again, third time in a row! Very exciting, do you recognize this ship?
CR: Yes and also we’d like to… this is a bunch of nice gifts we got during Citizencon. So there’s a few cool model of X-Wing there and alcohol that I’m scared to drink. Belt buckles from Operation pitchfork.
SG: Das Kartell
CR: And Das Kartell, Guard Frequency.
SG: That’s very cool.
CR: All these are very cool, and a very, very cool challenge coin here. So thank you all guys and thank you for supporting us, we couldn’t do this without you guys and I would say until next week.
SG: We will see you…
CR/SG: Around the Verse!