This post is a transcript of 10 for the Developers: Episode 13, material that is the intellectual property of Cloud Imperium Games (CIG) and it’s subsidiaries. INN is a Star Citizen fansite and is not officially affiliated with CIG, but we reprint their materials with permission as a service to the community. INN edits our transcripts for the purpose of making the various show participants easier to understand in writing. Enjoy!
10 for the Developers: Episode 13 – Full Transcript
Brian Chambers (BC): Hey everyone, Brian Chambers from Frankfurt. Jared reached out to me and asked me if I could help out for this edition so here I am. Let’s get right to it. You guys post a question so I dug through, see what I can answer, what I can’t. I’m probably going to mess up names and excuse me for my dull background.
[0:53] Sao Saoldian asks: What are some things about Star Citizen that has drawn yourself and so many other talented people to the Frankfurt studio to work on this project? From your perspective as a studio lead, has the crowdfunded model been a hurdle to overcome in regards to head hunting talent?
First of all it will sound like a fluff piece but it’s not. So honestly having doing this for quite a while the ambition behind this project and the commitment behind all the senior staff, Chris and Erin and the whole team, what they are able to pull together with the community. And also them wanting … pushing boundaries technically on design, forging some new ground in my opinion, and others, is really what pulled us here to take a look and 18 months later here we are still working and pushing on things.
In regards to the crowdfunding, honestly I don’t think it has hurt at all. It’s actually … it gets us noticed from time to time obviously from the amount that we’ve crowdfunded and that draws people’s attention. Which is great: it gives us a little bit more exposure and people know who we are. I do get question occasionally from candidates in regards to crowdfunding and what does that mean and the finances and so on. And I take the time to explain the detail of what we are doing day to day, month to month, midterm, long term, etc. So at most I would say it impacts is they want to know that things are secure if they are going to join us as a team … because it’s … for these guys here we’re hiring guys that are experienced and do this for a living so their career per se.
[2:48] Fosmosis asks: What are the advantages and disadvantages of working with four separate satellite studios in four separate zones as compared to one centralized studio?
I think the biggest hurdle honestly is communication. On the production side, on the senior level, across all sides, we need to make sure the communication is as solid as possible and that we’re communicating in a way that all the people that need to know the information understand where to go and how to get the information; If they have questions. That’s something that I think we’re constantly on the production side, looking at reevaluating.
I don’t think there’s honestly, and this is my opinion, I don’t think there’s one perfect model to it, of how to do it or not how to do it, but it’s putting things in place, having the flexibility to adjust them as they need to be. We’ve done a pretty good job so far, made a good amount of changes even since Frankfurt’s been involved. Stuff been by the UK with Erin stepping up and overseeing all the studios and the production level, it’s really pulled stuff together so we’re all speaking the same language.
There’s definitely some times though, you know, frustrating. I’ll get emails at two in the morning, I’ll need questions answered when people are asleep in LA so it definitely has some disadvantages, but at the same time it also evens itself out because work is being done almost on a 24 hour basis to some extent which keeps things churning. So we’ll leave here at nine in Frankfurt and come back in the morning and we’ll have seen progress and had things checked in and so on from LA or Austin that are vital for us moving forward.
[4:48] Daz asks: Can you explain a bit about your process for creating the first procedural location and what look and feel you are aiming for?
I get asked a lot about procedural planets and rightfully so because it’s cool stuff and I think we made some progress that people weren’t expecting, and kinda dumped it on people on that livestream which was pretty awesome to do. The first procedural planet really had nothing to do with the aesthetics, hey were we trying to pull off aesthetically… nothing.
To be honest with you it was just about proving tech so textures were pulled in of a certain memory footprint that we knew we would most likely have and then the tech was just pushed until we got that up and running. After those first steps then we had an environment artist 100% dedicated to working with all of the tech guys and engineers that were pushing on procedural planets.
They needed to sort out the best approach on how to get rid of seams, how to make those nonexistent. What limitations and parameters should or shouldn’t be set, on and on and on. There’s so much detail that’s going into that, now work is honestly right now as of this moment going on to some extent but a huge progress is being made. We’ve already done tests with different types of geometry and different types of look and different type of atmospherics, to really see what we can come up with.
One thing that was interesting is Pascal our environment artist was working on stuff, he’d been playing with this one planet and mind you it’s huge, for a good amount of time, a couple weeks I think. At one point he kind of made some noise and some of the guys had come over and looked, he had actually seen something like a new formation that he hadn’t seen although he’d been looking at this thing for, you know, weeks… in getting to a new part he’d seen a new formation look really, really cool but was completely unique to anything else he’d seen over the last couple of weeks. That’s cool, it’s telling us that the parameters there are working, it’s keeping things new, it’s keeping things random but still staying within the guidelines that the tech is putting forward.
[7:16] WarDog[KRT] asks: How will the definition of procedural planet generation work? Is it only a set of random seeds? Or will an artist define where land and where water is, and the algorithm seed will only define the shape of the coast and the land(mountains, river, etc)?
Honestly it’s a bit of both. First we go “Cool! What type of planet do we want? What are we generating?” And then we’ll obviously adjust the parameters based off of that planet for what we would like procedurally generated. At the same time during the process we can guide it as specific as we need. There’s a lot of artist control over there and I think that’s essential because we want people to come and have certain experiences on these planets. So not only can we define and create new parameters for each type of terrain or even each type of gameplay that’s going to go in there, but then the artist can go through and have really fine level of control.
We would … in a sense we want to focus on more on the quality than the quantity. You’re going to have a tremendous amount of quantity in it and I think you’re going to have so much that you’re not going to see things repeat themselves: which is awesome. But at the same time we know having that control of the artist to go through and fine tune stuff is extremely important. Not only for look but also for gameplay. And then on top of that we’ll create tools for the artists on top of that so now they’re actually playing around with the procedurally generated terrain and so on, to really give it … pull it together as efficiently as possible.
[9:01] Shrike asks: What technological challenges still remain before CIG is comfortable adding a procedural planet mechanic to one of their upcoming patches?
I get that question a lot. Right now we’re in the process of incorporating what we have so far in the procedural technology and everything else. When we’re proving out the tech and building it up and have Pascal working with guys, we have it in a kind of very controlled area. We needed to pull it away from everything else so we could focus on it.
Now that we’ve gotten it a bit more mature, now we’re pulling it together with everything else. We’re seeing really good progress on that, but then we need to see what blows up, what unknown or unforeseen errors have come up and that we need to address them. We knew there would be some issues already when we put it in and we’ve already identified those and it came true with what we thought because our logic made sense. So we’re already working on those and then we’ll churn through it. Once we get all that solid and sorted so it’s actually a good decent play experience for you guys, then we think that’s the time to hand it over.
Some of the stuff on the procedurals, it’s been interesting for us working on it. There’s a build with a couple different procedural planets in there and you’re on one, you’re there taking off in your ship, you get up, you fly away. Five minutes of flying through space later I see a distant planet and then another five minutes to actually get there, to get to that planet, which was nuts, you just start flying and we’re going as fast as we can and we’re going, “Alright, cool we’ve got this, wait, wait, wait”, and we’re just sitting there tapping our feet, tapping our feet. So that’s another thing we’re looking at is: you have the distances, but we also need to make sure that it makes sense for the game, I guess you could say that the distances are there and so on, but that’s for design and Chris and all those guys to sort it out.
[11:32] Drugh asks: Cinematics and animation have played a big part in Wing Commander series, but tend to be quite passive. Will we have a better blend of the player being in the room and simply have animated characters or is it going to be MGS style, long cut scenes of little interaction.
You’ll have a good honest mix of interaction with the cutscenes. You won’t have these long, drawn out cutscenes that are noninteractive all over the place. The cutscenes were designed, written and shot in a way to give a large amount of different variety and how they play out. Some do take control fully away from the player but for very specific reason, if that makes sense. There’s also a large amount that are fully interactive, and then there’s hybrids that put the player in a defined path but the player still has some sense of movement and sense of control.
So, I really do think once these all come together you’re going to get a whole good mix and it’s really going to feel very fluid and seamless. We also have added elements to the scenes where they branch, but branch in a very relaxing, casual, sensible way so the player has some choice but it’s not like, you know, things pause and the AI roams until you push a button. So, I think even those with different interaction that’s kind of casual, and the mix, it’s even get to make them feel a little more fluid.
[13:17] Xitix asks: Since the Frankfurt studio is still growing, are there any plans to build a german CS team there? Or is so would it be placed in Manchester? Otherwise which department (in Frankfurt) will grow in the future?
There’s been a bit … as you know there’s already a CS team in the UK. I don’t know how many people are there but I know there were more people last time I visited recently. All primarily English speaking: I think there’s some other languages in there too, but I don’t know them off the top of my head. There has been some early, early discussion – really early – of having native German speakers. Obviously being here in Germany and interacting with the German fans, although there’s a large amount of English that’s spoken, things are always, in my opinion, a little more comfortable in their native language. So with such a large German community, and I know the communities are large everywhere, but with such a large German community that’s where we’ve had those early discussions. That position could be here in Frankfurt, I think, which would make sense because they could interact with directly with people here too if needed. But it could also be in the UK embedded with the other team. So … yeah … I mean that’s still to be seen.
As far as Frankfurt studio and growing, I’d say in most departments we’re still growing. We’re still looking for a few more people on … for cutscenes. As an overview we’re hoping to drive cutscenes, we have a large portion on design here, on producing weapons, on AI code, and with engine code. With weapons we’re fully sorted out as far as team size goes. And different departments are at a slightly different state or various, right? What we’re pretty … we’re getting pretty close to to capacity. As I mentioned briefly, recently in an ATV, we’re going to do an expansion so we can build out the rest of the team and everybody can fit here comfortably. But yeah, as a general overview most disciplines we’re hiring for whether that’s art, whether that’s code, QA, etc. But we’re doing pretty good filling up with what we need.
Also important to point out, we’re not hiring people to hire people. We have a specific plan in place. We have a headcount plan based off of the work that’s needed, based off of our dates and so on. So all of the hiring, all the planning, all of the expansions that we doing and the building up of the teams is all based on a specific growth and schedule that we’re doing. Alright.
[16:30] Stadulator asks: I’m very curious about where CIG is with their AI, any news for us?
Nothing extremely new to report this second, I can say we’re making good progress. AI is one of the bits I try to give an update on and I know it’s fairly high level on the AtV every week but I try to give an idea of what the guys are pushing on and what they’re working on. There is an overall large amount going on with AI right now, with subsumption system and how we’re making sure that things are set out today that are going to be best used for the future, right.
We need to take the time, it’s in everyone’s best interest to build a strong a foundation as we can on things like AI so that can set us up to handle things we want to implement, you know, in the next few months, in the next six months. Even ideas we have that are, you know, loose on paper that we wouldn’t want to put in once the game is already out and mature, stuff that we’d still like to push on that we think would be cool for players.
So, we actually have here recently it’s coming up not to date this video but it’s going to come up in the next few days, where I think we’re going to have nine different people out here to Frankfurt office including Tony from Austin and Chris and Sean Tracey, so on, so on and so on. A good amount of people from the UK, with a large amount of discussion based on AI. So, we know AI’s important, we’re making huge advancements, we’re always adding new bits but we’re also taking the time to make sure like I said that we build that foundation as solid as we need to.
[18:31] KuruptU4Fun1 asks: Is building the weapons for the game something that gets prioritized and therefore built faster in order to get a specific patch over building it over the course of time and when it’s ready it’s put in then?
As in a lot of the things that I manage, honestly it’s a bit of both. We know what ships and weapons will be released, we know what new releases they’ll be in or what patches and so on, so we do our best to line those schedules up. Obviously if we’re going to put a ship out, we want to try to get a weapon functional on it, or as many or all of them if you can. Sometimes the ships move a little faster in production, sometimes the weapons move a little faster in production, but we do our best together to line them up to give the best experience.
Internally we have a good amount of weapons we’re working on. We have four full time weapon artists, we have our lead and then three other guys are helping him out. So there’s stuff that’s coming out of them on a daily and weekly basis, again making good progress.
[19:51] FloppyDonkey asks: Considering Squadron 42 will be story driven, how much if any of the procedural worlds tech can we expect in the single player campaign?
For that we’re honestly still sorting it out, I’ll try to explain. Since Squadron is primarily linear and it’s got a defined experience but sometimes it’s loose based on players actions and so on. Procedural planets are not absolutely necessary, right. At the same time everything’s on the same code base so we have those in there and we have them functioning meaning we have the tech functioning the way it needs to which it will be.
Could we use it so we can pull things together more efficiently? Absolutely, but I don’t think it would be something… and I could be wrong but my guess would be that when you know you go back in, you’re going to have a different experience or see different terrain, it would just be used to set the defined terrain at that point.
[21:07] Typhi asks: This has been a question alongside the whole community (at least on Reddit) for awhile. With new API’s becoming the standards in the future (DX12 and Vulkan), I wonder what CIG’s approach is on this topic. We want a future-proofed game which has the best of the very best graphics, in our eyes this can only be done with early integration of DX12/Vulkan so that you know what the limits are of the new technology and that you can build around that.
This one I actually … because I wanted to get you guys solid, solid, solid answer … a lot of the tech guys, like I said before, they’re really not interested in getting on camera but they will write stuff up for me if I ask and explain. They totally dig the community they’re just not fans of being on camera. So I asked one of our tech directors, Mr Carsten Wenzel, and he actually wrote me up some stuff so I’m going to sit here and actually read this so if I’m not looking at the camera I am reading it from paper. So his response …
“We’re actively looking to adopt the new graphics APIs. As you know DX12 and Vulcan differ quite significantly from previous API designs such as DX11 and OpenGL 4.5. They try to map very closely to current graphics hardware architecture, as well as provide applications more low level control of video memory. In order to take advantage, full advantage of them we started a massive refactor of the renderer and 3D engine in order more efficiently prepare and submit items to be drawn, well beyond the scope of the massively instanced rendering optimisation done by us.”
“Think of a departure from a monolithic render thread to more lightweight jobs that prepare individual render passes in parallel via the job system. The hope is this will allow us more flexibility in terms of draw calls in a lot of areas without complicated runtime merging of materials and meshes, or tedious offline optimisation passes on the content creation side, for example, such as character attachments.“
“This refactor will lay the foundation for all future rendering and graphics work. On top of that we’ll be reworking the shader system and resource management to better map those new APIs and take advantage of new features exposed by the new APIs and shader languages. Especially memory management is going to be a fairly complex matter and far more trivial but crucial in order to get the best performance. Don’t underestimate the support work of video drivers do on DX11 and OpenGL.”
“Additionally we’ll embrace asynchronous compute, taking advantage of the heterogeneous multi-GPU set ups. On-board plus discrete GPU is another possibility that might yield extra gains. That is instead of alternate frame rendering, each GPU works on different tasks for a given frame. This need some initial research first though to see how it can work and if the performance gains are actually worth the effort that they put into it.”
“In summary, we agree that adopting these API’s is going to be essential moving forward but at the same time we want to take our time to get stuff done properly so you’ll actually see a benefit over DX11, visually and in terms of performance.”
So answer yes we’re absolutely concerned with future-proofing. I wanted to be able to read that and although I probably messed up pronunciation a little bit. The guys that are working on stuff here they know their stuff. We want to set this up in the right way that we know it going to be able to milk every thing we can. That we get the best performance. That we can support the best hardware and so on. So the guys are really taking their time and putting their energy and their big brain tech sweat into getting this where it needs to be.
[25:36] MisterCole asks: Thanks for stepping up to the plate yet again to answer our questions. Compared to your earlier work with animal and motion capture for Roughnecks: Starships Troopers Chronicles, how has the technology advanced in the last few decades and are there any drawbacks compared to the earlier methods?
Tech’s made huge steps since I first started. I mean tech’s made huge steps in the last few years. When I did the work on the initial season roughnecks, optical mo-cap wasn’t really used in production. I think you had some of the larger film effects companies using it, but I don’t think any smaller houses or games were at all. At that point we had a magnetic system, flock of birds I think it was called, staged size was limited to 12×12, 12 feet by 12 feet. It was supposed to be wireless, but it would sense these wireless data packets and occasionally the data packets would drop, which would make the motion completely blow up or freeze so a lot of times we had to run tethered.
So you had somebody in these funky suits with all these gyros on them and then even on the software side there was a lot of scripting to get stuff to work. Mapping something, the actual motion to character, it would take a full day to set up a character and get them running when now you can do it now in a couple clicks.
Back in the day when we finished stuff and we looked at it, we were so proud seeing characters moving in the environment. When I go back and look at those nowadays, you know you can see the high frequency noise, I kind of cringe, you could see that high frequency noise in the animation, you can see floaty feet, contact isn’t really there where it needs to be, but you know that was the day and for the time of what we did I think it was really cool.
Today you have huge volumes, you’re capturing numerous people at once, you’re capturing body, you’re capturing face, you’re capturing fingers, we’re recording audio; we’re tracking props, huge volumes if we want to use them. The software is also leaps and leaps and bounds. It’s almost to the point to say kids nowadays have it easy, right? But it’s true, back in the day it was a lot more archaic, but it was still incredibly new and exciting back then just as the new stuff is coming online now so it’s cool.
[28:27] [BN] Kaji Arkale asks: Uncharted 4 placed the level so high in term of quality (animation, lighting, level of details even in large open area, content streaming, etc) and on a PS4, a hardware where CR clearly stated in the past that SC couldn’t run on. Does that change the team mind on the power console can have and can we expect that you at least reconsider a console release on the long term? On a technical point of view, I mean.
Uncharted 4 looks awesome. I think it was the day or so after it came out, we grabbed it and had it here at the studio and I told the team “hey, let’s go in the kitchen and talk to Chris about it and so on”. When I say kitchen, we had a couch in there with a console, so we all went in there and were playing it. They did an awesome job, it looks good, it plays nice.
Generally in console development and I can just say this from experience: the games look nicer the longer the console exists. You understand how to use the hardware better, you understand how to go around or through or above or under the limitations that the hardware gives you. You find new tips and new tricks on how to put things together, manage memory/stream things, so on and so on and so on.
So as you can see, if you look historically, even PS2 and so on, late generation PS2 games looked awesome comparatively, right? And it’s just because people got more mature on how they can use the hardware. With that said there’s still a ceiling, there’s still a ceiling on that hardware on what’s possible. You have technical limitations. If you’re applying the same learning and then same logic to a piece of hardware that has a higher ceiling, in theory you would be able to produce more or do things better or do things more fluid. It doesn’t necessarily mean that it would look better which is clearly what you’re after right?
Uncharted 4 looked great and it was based on a certain hardware and we had harder hardware, could you draw more characters on screen? Could you put a little more love into the atmospherics? Could you have more on hair, cloth animations, just all those as examples right?
So yeah I think the team over there did an awesome job pulling it off, it looks amazing, yet that doesn’t point directly that Star Citizen could run. As far as, “hey would you consider it in the future?” It’s not really a point to think about right now or put energy on because the teams fully focused on delivering Star Citizen and Squadron 42 on PC like we promised to the community, to the backers, to the fans and so on, so that’s really where all of our energy is.
Brian’s Fire Round
So I’m gonna grab a bunch of these, last some, maybe another 10 and answer with one word. I already kind of sorted the ones I can answer with one. So this will be Brian’s Fire Round.
[31:53] Leonard Sapalla asks: In AtV weekly updates you’re well known for saying a lot without actually giving anything away.
Thanks Leonard, really appreciate that. He also goes on the ask:
Any updates on VR/head tracking support?
[32:16] Salotorn asks: Can you talk about current plans for spinning motions of planets and moons, will they have realistic spin or will they all stand still?
[32:28] Aragorn [BH] asks: How does expansion of the Frankfurt facility going and has office space limitation prevented you from hiring new people?
Good and nope.
[32:45] Lorik Eterness asks: I’ve been wondering for a very long time, how do you take care of that luxurious hair?
[32:55] Destrike asks: In the first procedural video it was somewhat easy to recognize where procedural planet met the handcrafted landing area, regarding the blending and hiding of those areas would it be more of an automatic or manual process?
[33:16] Saber asks: If a real gun fired from an Idris in Sol, might a Prospector hull be breeched in an Orion?
[33:30] Almack asks: Are there any innovative technologies or gameplay systems being researched/worked on currently that haven’t been heard about yet but you really need to share with us that will have a major impact on the game if they can be made to work.
[33:52] TheHound asks: About animations, are the get in and out pilot seats settled for every seat or are they meant to evolve?
[34:04] Kaji Arkalel asks: Which shampoo do you use to keep your hair so nice?
Thanks for sitting through this last 35 minutes, appreciate it. We at Frankfurt appreciate all the support, appreciate the support for the game, for all the studios. It’s fun talking with you guys and doing this when we have the time so thanks again and if I did an ok job maybe they’ll ask me to do it again, cool. Take care.