Star Citizen Live: Working With Gas

1 May 202001:02:101,638 linesWatch on YouTube ↗

Transcript — every row opens the video at that moment

00:00:02
Hi everybody. Welcome to another edition of Star Citizen Live Game Dev. I'm your host Jared Huckaby. And if you've never seen a Star Citizen Live Game Dev before, it's where we take about an hour out of our Friday, the end of our week, and explore a little bit about the process that one of our our developers or group of our developers do. In the past, we've created characters, we've created creatures, uh crotch spiders, which we won't talk about. Uh we've done spaceship concepts. Uh today, we are we we are going where no game dev has gone before into the realm of VFX and exploring some of the work that our team is doing developing our gas cloud tech for uh well, for Star Citizen and Squadron 42. And joining me on the show right now are two people that I was going to say need no introduction, but I'm pretty
00:00:49
sure this is the first time they've been on the show, so I guess we should do an introduction. I always do an introduction anyway. Uh Jacob and Oliver from the VFX team, how you guys doing? Yeah, good thanks. So, uh welcome to the show. This is your first time on the show. Uh let's tell tell folks a little bit about who you are and what you do for Star Citizen. We'll just start at the top and go down. Uh so, Jacob, who are you? What do you do? Uh yeah, um I'm Jacob. I'm a VFX technical artist. Working CIG for S42 and PU. And and what is it what does a technical VFX artist do? Um oof, that's a I I guess I could confidently say that we mostly work within the program called Houdini as it's I guess it's what we work in most of the time. It's a bit like 3DS Max and Maya, but just a
00:01:38
bit different. It's a bit more technical and a bit more uh More suited for particles. Artist. Yeah, yeah, it's a bit more of a simulation-based program, which makes it great for um dealing with VDBs and gas cloud. All right. And how about you, Oliver? Uh so, I'm Oliver. I'm a delete visual effects technical artist. Um yeah, uh as Jake said, like we work on like Houdini pipeline and we also support the visual effects teams and other teams around the company working on the PU and Squadron and like helping them with the tools and the processes and we're going to go through that one of those processes today. Awesome. So, we folks who have been following the project for a while know that we started our our gas tech our gas gas cloud tech for start for Squadron 42
00:02:28
developing a location called the coil. Folks who have been following have saw that first in the the vertical slice that we did a little while ago or a lot of while ago. And now we are looking we're currently looking at ways to to bring that tech into the persistent universe right now through I know that you guys are experimenting with the the the Grange points and stuff like that. So, why don't you go ahead and start your screen share? I'll get rid of the cameras and you can show us a little bit about what you've been working on. So, um Yes, so we showed at last CitizenCon the jump point and what we're going to do today is I show how like that jump point was created. We did it with in Houdini using like the gas cloud tech and we are going to like talk about the
00:03:16
processes that we make with it and also like doing the parent and child relationship and that is where we have a large gas cloud for just like doing the very like long distance far away shots and then where we put a a smaller one in the middle where we get that really close to it and we see all the detail. Um, and that's what we're going to try and do within an hour today. Um Before you get started, let's go ahead and move my little camera up the corner here. It's just I'm just looking at myself and slowly there we go. That way we can see the work. And you're going to have to explain what we're looking at right now. This doesn't look like a gas cloud. It doesn't look like a gas cloud. So, um in Houdini we start with um geometry and like our workflow has like evolved over quite long
00:04:04
uh quite with like everyone who's worked on it has changed slightly. Um one of the things that Jake brought to the gas cloud workflow was how we use geometry and like Houdini to make uh the big um gas clouds we see. Uh the first thing is that we're going to look at this pig head. It's a classic Houdini primitive. Uh scale it up for um just size so it fits in um the world better and we're going to convert to convert it to this fog volume. Um this is now can we can export that to game um and we can see it here. Um this as you can see is the pig head as a volume. Um but it's really base it's very solid. It's it's quite nice how we can see it really
00:04:51
quickly, but we want to make it really nice and cool. Um so we're going to just start by like we're seeing it right now where it we don't actually have any shadowing on it. So, if we put the sun shadows on it, we can now see that we get some detail in here. Um and we also have like a very like simple light that we have at the moment. Um so what's like really great about this tech is that we can offer like all these shapes that we want and put them in and that's really fantastic and that's something which we haven't really like been able to do in games before. Um so let's go back into Houdini and start off making something. Yeah, just a Just for the folks that are in the chat right now, that the pig head is just a standard a standard piece of test geometry that's within Houdini.
00:05:39
Yes. Yeah. All right. Uh and it's a bit of like a Houdini trope, so I had to show it. I'm sorry everyone, but It's a it's a Houdini meme. It's It's not something that's going into Star Citizen before before I see the Reddit thread and then I have to get that email from CR. What did you say we were putting in the game today, Jared? it's good to know that the style modeling to create this gas cloud is there's so many different ways of doing it. This is just one of the ways to do it and I think until we we plan to do all of this. I haven't actually seen um modeling gas clouds in this way using using essentially the layers. Um you'll probably see
00:06:27
the result of that quite quite soon. Yeah, so it's kind of interesting cuz what I'm doing at the moment is I'm trying to like create a polygon of the big forms that we're going to use later on. Um and we're going to do this quite rough to start and then maybe if we have more time we're going to add in uh the detail later on. But at the moment we're just like playing around with like the shapes that we can get and make something quite interesting to look at because um we found that sort of like the best things that we can like see that we can fly through are like these very like abstract forms as well. Um and it's it does like change from location to location. Um so like for instance the the jump gate that we saw had that we
00:07:16
called it the Owl Nebula because it had this like very magnetic theme to it which had looked like owl wings so it's I saw them brought up and then we couldn't get away from like calling it that. Um and that's sort of like based around the um the lore of how like the jump points work. Um so like for each different thing that we do we're sort of going to be playing with different methods of creating things. Now, right up the bat, um we we we are dealing with glass uh glass you see gas clouds uh today. They're they're they're a space uh in in space like nebula and stuff like that. Uh what we're talk what the tech that we're talking about today isn't what's going to be used for volumetric planetary clouds and it's not what what folks can look forward to for Crusader for
00:08:05
instance which is which is a gas giant. Those are those are different initiatives. Yeah, and it's really important to understand that they're all really different challenges as well because like the thing with the gas clouds is that they can actually be bigger than the planets um and and then like also just trying to like tell clouds across a planet is a completely different thing. Uh and it's a really interesting like to have all these different technical challenges that we're having to deal with at the moment. Yeah, so what we're talk what we're what we're exploring this today this week is uh gas clouds in space nebula the likes likes of that as far as uh atmospheric stuff that's not what we're talking about today. It's not what we're showing today. This isn't the process for that and for creating a gas giant uh planet
00:08:53
like you said or again different initiative different tech. So that's uh that's not what we're doing here today. Just to set some expectations early on in the show. Cool. So yeah, what you started with a with a simple strip and then you've just been mutilating this thing. Yeah, this is one of the big challenges when when it comes to modeling the the larger forms of the uh gas cloud is you're trying to you're trying to model um like essentially like an abstract shape trying to model chaos in in in a sense cuz you just want really random shapes big forms and and running it can be quite tricky cuz you're trying to
00:09:41
like create randomness when instinctively as a as an artist, you're trying to create something with with actual shape. Okay. So, this thing you're doing in the lower right in in the engine, uh we we we refer to something that looks like this as like flow graph. Is this does this function similar to flow graph? Ish. Oh, also, Ollie, uh poly extrude, you need to help with the back. Thank you. It's why he's here. We said that was going to happen when we invited him. It's like, "Oh, this just means I get to backseat drive his work, don't I?" Sorry, yeah. It's about it all week. It's my time. Sorry, what was the question? No, I was I was saying, so so so what
00:10:29
are we doing? What are these steps that you're that you're doing here on the lower right-hand corner? The transform, the blast, the mountain, the bend. What do each one of these do? You don't necessarily need to hit each one, but Yeah. What's happening with these? So, um Houdini is a node-based program, unlike 3ds Max and Maya, where you apply a function to it. So, it's like the your polygons, and then you do an extrude to create a cube, for example. Houdini works in a similar sense, but instead of keeping things in a function and just having code in the background, say you want to perform this action, Houdini actually has all of these in nodes, so it's totally non-destructive. So, if Ollie went all the way to the back of his um to the top of his network, you you would start from your you'll be you'll be able to see the
00:11:15
original grid that he created from. Yeah. So, it's just totally non-destructive, and it's actually one of the big benefits of Houdini. Is that non-destructive workflow? Yeah. Yeah, cuz then you can start sampling data at different points of of the network, uh which is actually what we need to do and actually how is is the only way we can overcome some of the technical hurdles when we start dealing with things as complex as the coil. Um we need to know and and extract data from our models at different points of its creation. So, we need to know maybe the size of of a feature and then as before and after you performed an action on it so we can make corrections if need be.
00:12:05
And and you have you have some you have some kind of live link between Houdini and the engine here. Well, no, we don't actually, but How are we seeing the What's updating that fast? It just it just updates really well, which is like really really convenient. so, where did the dust cloud go? I actually scaled this object up, which is why we I have some slow down. But, uh what is like really good to just do is just to view these things um in editor like as really quick as we can. Um so, this this is quite big actually. You did scale up by 50 and then you did the output back. Yeah, we we don't have any uh reference
00:12:53
item to show the scale here. Yeah, and that's that's something which is uh yeah, something we need to figure out really. Um so, what I'm doing at the moment is just sort of figuring out what the lighting is going to be and I'm also going to make this a little softer to start with. But, like what we have at the moment is really harsh cubes. Um and let's just improve this lighting a lot more. Where did my light goes over here? Cool. But um our next stages will be to make this volume look a lot more interesting from what it is. What's like
00:13:40
fun is that we can actually get in these like quite quick shapes to start with. Um and then also to add the order like low the high frequency detail and then the Yeah. start looking for locations to actually put something new and like What what we're going to do is put like a station in and then we're going to dress up around that station as well. So, cool. Yeah, so folks in the chat asking what gas gas cloud tech could be used for. While we as we've stated this isn't what's going to be used for something like Crusader or or even for atmospheric clouds, this can be used to make the space around a star system much more interesting.
00:14:27
And much more dynamic. This this is about This is about making space more than just black empty space. Mhm. Yeah, this is Think of it like celestial bodies, not not planets, you know, you got maybe asteroid fields and then you have a gas cloud covering it to just make it a bit more derelict ship in the middle of a of a nebula or maybe I don't know, spilled oil around a gas and it forms around the ship. Yeah, nebulas are are are real easy go-to. Like the Mutara Nebula in Star Trek II. I always have to bring it back to Star Trek. Bring it back to Star Trek. Yeah, on any any given day On any given day I'll bring it back to Star Wars or Star Trek. It's going to be one of
00:15:15
those. Uh Uh, so no Gambril, not gas cloud tech for Crusader. That's a different tech is what we're saying. It's different, they're not the same. So, this not being the thing that we're using for Crusader does not mean that Crusader is 2D. It's It's that it's a different tech that's being used for that. Yeah. Yeah, we we kind of ruled out small note. We explored VDPs and kind of said, "Hm, not not quite the tech that we should be using." Yep. Yep. They're different technologies, different solutions for different problems. And yeah, we didn't like Yeah, like it's in polarity, wasn't it? And we did try it. Um, but for what it is it's not the right thing. Um, and that's
00:16:04
And that sometimes does happen. And that's also just a great learning experience really of what we're doing. So. Uh, let's do it like that. Okay. So, we have a simple base form at the moment. Um, what I'm going to do is I'll start adding some noise in. Um, and this is going to start providing us with some actual like detail on the on the VDP itself. Because at the moment we just got these like large abstract forms, but we really want to sort of distort it more, add uh, more noise within itself, and just like um, make it just a lot more interesting cuz at the moment it's nice, but we could do better. I think what we want to do
00:16:56
Ooh. Getting Uh, Cad and Flint, that's certainly an interesting question, but these aren't the These aren't the folks who'd be able to to answer that. Remember folks, these are our VFX artists. So, questions today should be about process and about art. Uh, if you're looking for gameplay applications, those are usually the realm of designers. Yeah, we we we make the the tools for artists to easily well make an easier process for them to actually model this. Like the So right now what Ollie's got is a very flat base form to build the larger shapes. But as you can see this is a cloud detail. Uh, doesn't mean it's going to be used for clouds. Um, it's actually just a
00:17:45
Inside of this node is a more complex network which we've simplified to make the UI just show like values for artists to tweak to tweak and to easily create um, detail on these gas clouds without them having to constantly figure out how I'm going to add detail. Me and Ollie will kind of go in and and explore different ways to do it and then package these up in a much more friendly way. Yeah, so at the moment I'm just trying to add more large scale offsets distortions to the volume itself. Um, and then we'll add more of the like the higher frequency stuff on the on the gas cloud.
00:18:32
Um, So Uh, Lord Styles has a has a tricky question. How can gas clouds or nebula even exist within a star system due to the pressure of the solar wind? And I'm going to say that's a question for the designers. We'll have to work that out. Uh, come up with a back story. Yeah, that's one of those things that luckily artists don't have to worry about. We just make the pretty pictures. Yeah, what is quite interesting I don't know. No, after you. Um, what is quite interesting about the uh like the gas cloud itself is that it does drive gameplay parameters. Uh so, for
00:19:20
instance, say the pressure, we have um atmospheres systems within the gas cloud density. Uh so, whenever you go inside the volume of one, um it will drive like a pressure value in the atmosphere, and that drives other systems. So, what we can do is that we actually drive like the particle effects in there. So, you see more different types of particle effects um and like the intensity increases um when you go like deeper in and the pressure gets bigger. so it's quite interesting how these different systems end up being tied together, and that's something that like happens quite a lot across Star Citizen Citizen, where we tie a lot of these parameters uh together. Uh like with folks who saw the the vertical slice for Squadron 42 saw how the coil had some adverse effects on uh
00:20:11
the ship systems. Yeah, with the with the communications being garbled and stuff like that. So, it's uh there are certainly gameplay elements to this work, but again, those the which gameplay elements uh come into play when, where, and how are the realm of uh the system designers as opposed to the VFX artist. Uh Zandor has an interesting question. Uh do you find Houdini Pyro VFX or do you find Houdini Pyro effects better in your opinion than non- node-based pipelines uh in like a program like FumeFX? Um it's kind of just So, we we were talking about this earlier. Um I'm biased towards Houdini's Pyro because um I'm more of a Houdini kind of person. I've used um the Pyro solver and the the
00:20:59
whole Pyro tech um a bit more robustly, and it is capable of really powerful things that the one of the perks of Houdini being node-based is that even the solver, like all the calculation that happens to calculate how the smoke sim should operate is all totally exposed. So, I'm able to go into essentially the the code. Not not quite code, but the node that builds up this entire solver. I can manually tweak the parameters. In Fume, I can't quite, but it it is kind of like you pick your tools for it. I know Fume does produce results a lot faster and a lot easier than Pyro. Pyro can be a bit You know, you've got to work at it to be like clay. But I mean
00:21:47
Fume's a perfectly valid I think Jake Gainey, our um assistant creative director, I think is his new title, is it? his new title? Is that his new title? He was He was assistant art director the last time I talked to him about Oh, I think that was it, yeah. Right, yeah. Yeah. Um he he uses um Fume quite a bit. He doesn't really use um Pyro. And some of the stuff that he produces just with Fume is crazy. Yeah, he's got to like when I talked earlier about different artists bringing different methods um and different like and we like learn every time that we do this. Uh like Jake Gainey is one of these people who's done that. Like he like he takes the sim ends from FumeFX, brings them into Houdini, and then has them as like separate no chains. So, at
00:22:36
the moment we're just going down like one linear path, but he would have like quite a few different aspects in and then build up and layer up the simulation and like the VDB grids um and it's a it's a really versatile method um and that's why like Fume's great uh but like Fume with Houdini is like really like super powerful and it does everything that we need it to do. And so, that station does a little bit to tell us to show us the scale that we're working at here. I'm just trying to find a nice area. I think where it was was probably best. Location scouting. But yeah, we've got another artist um, Jake Dunlop. Um, he uses um, particle sims um, as his base. So we all have different methods. Yeah, that's one of the
00:23:24
It's one of the neat things about it we we see this quite often through different departments in in the in CG. It's it's a a push towards using the tools that our artists are most effective with. And it's like, you know, if you're if you're more effective with this, then use that. If you're more effective than that, then use that. And we we we have an entire tools team that works to build all of the connectors and the and the translators that are necessary to allow that to happen. That's cool. And they both give quite They will they will give results like good results, but different results as well. So if you want more of this kind of look, we're going to have to use um, the the the a a fume simulation. If we want another particular, we have to use um, a particle simulation. And then we have to start maybe thinking about tools that would help aid the process. So if a artist who uses
00:24:13
fume can't quite get the result that he wants with the particle um, that the particle workflow produces. We start this where again, Ollie and I start to come in like, what tools do they need to make this easier so any artist can come and use this so it's not exclusive to this one artist. Aha, got it into game. yeah, so what's really great like with this process of being actually like being able to copy get that that volume into game is that we can just really quickly fly around it. And I made this pretty big now, so it's going to take a while to fly around, but it could see like a good like idea of what you're looking at. So, yeah. It's It's interesting. I think
00:25:05
it's it's still really rough, and I think a lot of what we're going to make this session is going to be quite rough. We did a couple of dry runs earlier, and we were going over time, so we had to like speed it up a little. Um So, we'll point out the things where we would spend more time on um and like make it nicer, more presentable. Cuz we can see here that we're not getting very good noise at the moment on this section, and that's something we can improve further down the line in Houdini. Um So, what we want to do now is add a child VDB in here. So, this is going to be a local VDB which is going to sit around the station. I think like Jake, what do you think? Do I Do you think I should make this bigger?
00:25:51
I think you should embed They've got to embed that uh station into something else. Into into into the VDB so you can actually show out of that detail. Quite thin. Like maybe I should put it on here. Let's put it on the like cloud back. Yeah, let's do that. Like the Star Trek Voyager intro. Yeah, let's just whack them in there. See what happens. That's cool, yeah. No, maybe a a gas a gas mining station, who knows? Okay, cool. So, let's figure out where this guy sits. Uh uh That's one of the challenges is that uh when you go location scouting in engine
00:26:39
and then you try to do it in Houdini, you just get lost very easily because they look very different. Your Your sense of perspective is changed. I think I should rotate this. I did Let's take that. That's good. Uh okay. So, we think it is there. Um you think it's that one? Yeah. Yeah, that looks right. Yeah, you've got that hole. Yeah, that looks right. I'm going to guess where it is. Right. So, let's pop the sphere down there. So, what um what he's trying to do now because if um the way that our parent-child tech works is that the when you put in another VDB, a child
00:27:30
VDB, into a parent VDB, what it actually does is it chops out that chunk out of the parent, and what you see is the child VDB. And some of the things that can happen is um scenes you'll be able to see like a slight slight separation between them. So, the method that Ollie's using is um we actually take we cut the chunk out of the parent, and then we do our process to to up-res it or add more detail to it. So, it doesn't So, it still looks connected when you're really far away, but when you're um approaching it with the QT, for example, you can start to see the detail start to fade in. Which is the the the entire process of having a parent-child relationship is to bring back some of the detail that point
00:28:18
of interest that's lost because your parent VDB is just so large. Yeah, this is a process we had to develop for working on the coil because we have the coil, which is like this outstanding location, but it's huge. Um and we've got to get right inside of it. And there's we want to be get spaces in there which are really tight and also expansive. Um and one of the issues with working with these VDBs is that we have this memory limit. Um at the moment, I've got a like quite a reasonable memory limit. Uh, yeah, it's quite good. But, if we go too high, what we'll do is just run out of like graphics processing power. Um, and there's a trick that we can see the voxels quite clearly here. Um, that one of these across is going to be just
00:29:07
really big if it's just a huge thing. Um, so like at the moment, this is really kind of small compared to uh, the coil. We have We have a station in there which is probably about this size. Um, and this the station probably just sits sort of like here. Um, so what we have to do is have this child um, there to like make the play space realistic for us. Um, and we also use this tech for um, the coil tunnels and the interiors as well because we need to have those at high res. like the jump point that we showed off at CitizenCon, um, the the jump point location would the the entirety of that area would sit in maybe a 100 by 100 voxel cubed area. You wouldn't be able to get any detail at
00:29:59
all if we just worked at that scale. So, that that's where the whole parent-child came in. We could put in a more high res um, VDB in that point of interest location to make it a bit more pretty to look at rather than just a couple of boxes. That's what you would have seen. Just a couple of boxes with different densities. Do you happen to remember what Oh. Oh. I think it's going to come back. I'm going to cross my fingers and hope it do do. Um, but it's okay. We're at the point where if it doesn't come back, we're we'll be okay. We got the saved one. Let's Can we alt tab to the other engine and look around there Well, it's Well, it's Well, it's resurrecting itself. Uh we had some questions in chat about uh how
00:30:47
performative this is this Is this Is this something that would chunk the frame rate as it is right now or Implementation has been No, this is like something we have to keep an eye on because like rendering 3D volumetrics in games, not many people do it and the reason why people don't do it is because it know honks your graphics card. Um so, like this is something which we're just always looking at. Um and this is probably related to rendering and how many voxels we see on the screen as well as the lighting. Like the lighting cost is a huge thing. Um that's why like a lot of our lights are quite static in here. At the moment, we have like this um like this little animated light that we can see, but
00:31:35
calculating that um all the time or calculating a full light with shaders is very hard. So, we have to make sure that we're doing like a performance analysis on it all the time. Um when we're building it, we're building it in like very uh just like logical ways instead of things which are going to be um just like really bad for the user for your playing experience. So, um yeah. That's why we um the gas cloud's actually static and not like an animated gas cloud. Right. Cuz it has to every time you load the gas cloud, it's got to start calculating the lights. It's got to process the VDB file itself. And it can take some time to do. So, why it's not an animated because the VDB
00:32:23
file format is actually a static format. Otherwise, you'd need to render X amount of files to to play over each individual frame. shading that These uh do the particle simulations are they generally run on GP or CP to sim to simulate? Um a bit of both, actually. I think we're uh starting to go towards the GP it is just better. Um but but before we were primarily CP, but now uh we've got a lot of stuff on the GP, so we're taking a bit of stress off the CP. Just means we can add more particles. Right. And I know a lot of that and a lot of the prep work for Vulcan will will help with that as well. I think a few of our artists can
00:33:12
actually finally crank up those particle counts and a lot of cool noise. One in particular. No, I picked the wrong area. I'd say what where do we get that cube from? Where's the cube from? Did I save the right one? Good guys, try. Oh, good guys, try. That's the right one. I believe in you. I think you picked the I think you picked the wrong location. I would not pick the wrong location. That would be very embarrassing. that's it. That's the wrong location. That's the wrong location. I did it. I knew I would do this. Okay. Okay. This is no We can fix this. Don't worry, guys. Don't worry. But this is the one of the one of the issues is that there's a disconnect
00:34:00
between what you see in engine and what you see in Houdini. So especially something as big as the coil, it's quite difficult to find. Or if you are That's why it's like important for us to like get this stuff in early and quickly we don't want to like work in Houdini and then go to game and then be like, "Why is it wrong?" Uh like we did. Uh for those asking questions about Vulcan since I mentioned it, uh I point to you to uh calling all devs from a month or half ago, maybe maybe 2 months ago. We had our director of graphics engineering Ali Brown on and he talked to at length about Vulcan and what you can expect from its implementation in Star Citizen. So, he already said it far better than I could. So, you should
00:34:47
check that out. It's up on our YouTube channel right now. Cool. So, what we see now is this is the child VDB. if we if I just turn down the density quite a lot, you can see it cuts out the parent one when it renders it. So, what we do to do this up res is that we take a copy of the high res high res version and then paste it back in to the child. And then we just sort of like make that more like gradually more detailed and we put more detail around that location because like you want to fly around something that um That's good. Well, like just like instead of having like the large blocks you bits, you want to have like them the finer stuff. So, So, I'll just go and start doing that, I guess.
00:35:36
what I'm going to do is probably going to hijack what I did earlier and use that as a base for making this child VDB. Um although I should do the um copying of the the versions. Uh what is the what is the tool I'm you looking for, Jack? Volume merge. Volume merge. I made it as well and I can't even find it. I was going to say something. I just wanted to see how long it took you to figure it out. You didn't even notice the new one. Uh-oh. So, the the cubes keys just trying to
00:36:29
find the the actual locations hopefully going to make something nice for us right. I for one think it's going great. Oh thanks. that's what we want. Doesn't look like much but it's well never never does when you take a slice. I'll double check. Where's the parent gone? Where is the parent gone? Just going to check that. Get a job from the parent. Careless, very careless. No.
00:37:18
I'm finding all all the bugs. I wonder if that is Oh yeah. Oh have you reached the limit? I might have done. There's 25 million. Is that 25 million? No no what you sliced out was 25 million. Yeah yeah this is where Houdini comes in useful is that it's very data driven and we can kind of guess how it's going to perform in engine just by looking at the data and the values that Houdini is analyzing. What are you trying to figure out?
00:38:06
Why is it so large? Your transform's huge. It's four times the area. How do you mean? So Well look in the view port the cubes massive in relation. Yeah, and then it's interesting how it's then you're cutting out. Oh, I see. Ah, so it's kind of like doing the inverse of it. It's filling and then subtracting a a hollow cube on the inside. Silly me, I got the inputs wrong way around. There we go. There we go. That's better. Now we're cooking.
00:38:58
Yeah. So you can't really see a change. Uh you can kind kind of. You can see the the the seam. So it it has copied and pasted it across. So that what's the the space station is now like a higher res child VDB. So this now allows Ollie to add a bit more detail that you wouldn't be able to on the parent to the points of interest location. And modeling in the scale is quite tricky cuz you don't really have a uh a point of reference on on scale. What do you use? How many addresses across? I mean you already got the space station. It's about as big as you can get at the moment. I do I do use that reference quite a
00:39:47
lot. I like um but that is it's like 40 20 m across, 40 m wingspan. so that you have to make sure that like that's like a a nice benchmark to use. Uh I could use a person, but like this is Star Citizen. We've got We've got Gladius assets to use. 20 Gladius across, four Cutlasses in height. did I Yeah, that's not the right So, I got to just like start that workflow again um to add some more stuff on top here. Should use the sphere one. That was quite good. Oh, yeah, I should use the sphere one. How are we doing for time? We could probably do it. 20 minutes left. We faster. Yeah, maybe better to start again.
00:40:39
So, welcome to sausages and live. Yeah, there's not There's no one way to do things. There's 20 ways to do the one thing. For For everything. That's why they call it game development, not game construction. You can't You're not sitting there with a set of blueprints. You know, everything's already been figured out for you. All the engineering's been done. You just You just have to follow instructions. Yeah, that's not how this works. Yeah, so he's trying to, you know, create that abstract shape again to what he did with the with the grid before to give a bit of complexity so you won't be able to see into the sphere right now. It's just a We'll create a hollow a wall a a a filled sphere.
00:41:26
Words. A filled sphere. I was just going to let you go. I wanted to see where you were going. You stopped me. No. No. We fast. I know. I In theory, I'm supposed to be here to save you when that happens, but sometimes I just I just like to like watch the train wreck. I'm just like Where is he going? Like he's thinking fast. Oh, good. It is. Okay, so I got to do that referencing again. Yes, to Oh, don't forget your rotations. Yeah, and like doing this is really fun.
00:42:17
I did a couple of tests over the last couple of days. Make sure like what we're doing might be actually interesting for everyone. and it was just super light. It was a blast. Um, like just working with this tech it's just it's just great fun. Um, and it's really just it's just great to like use it. And you can just make so much lovely stuff with it. Um, so that's big one. My my favorite part is working out the different ways you can you can create it. This yeah, you know, there's the the pirate simulation, there's a particle simulation. You can build it by hand. I think at one point I tried simulating what life cycle of um of stars. That went well. That was Yeah, we should explore that
00:43:05
again, though. I think I still think it's a valid method or I don't know if it's a valid method. Um, what was So, what we were hoping to do was that um, an artist would create a base shape or just populate an area of um of particles. And you would place the rest of your bodies in there. Like um, suns and stars. what we were hoping to do is simulate over time if the star gets dense enough um, it would collapse on itself and explode and spew more particles. And then the gravity of the star would start pulling the particles towards it. And as it dies and explodes, it would push it away. Can you tell that Jacob was watching a lot of like physics videos at that point.
00:43:56
I think watched a lot of uh, Neil deGrasse Tyson with science videos. And like it's I think like that's a great method like actually taking from that life as well. Like what we're doing wrong here I would actually say is that we're we're making just creating from randomness and we don't have a lot of like reference in the background. And like when people like really start doing this they do go get that reference. And it's just you need that when you're like doing art creation stuff. It's it's really Yeah, when you look at we do concepts. That's why we do concepts. It's I think it's quite tricky when you just pull up like the Eagle Nebula or the Horsehead Nebula you say I'm going to model that. I want that in my game. You're kind of
00:44:44
saying how do I do it? Yeah, and that's already like abstracts as well because they there's they're like artist interpretations a lot of time as well. So it's all It's it's hard to like I think we're losing your audio there. Oh, sorry. You you hear me okay? Yeah, yeah. Yeah, that that a lot of those are artist interpretations of um like what the render what the the data is. So it's kind of it's kind of interesting to try to like find like the truth in that as well. So this is weird. This is going to be So I'm just copying this on detail Yeah. Let's let's put this in now and then see how it looks.
00:45:34
and then Let's see where we've lost our station. Oh, that's got a nice little pocket down here. This could work. Uh yeah, like one of the things that is really important when we're actually looking at um in the locations is actually bedding it into the scene cuz right now we've just got blank space around it. And what we do really is like this parent that we see this big one, and the child. The parent is huge in comparison, and that does a lot of like the background space gaping. We do a lot of fading in there. Um and we mix that with like uh planes of um we use have this like shader called VFX transparent, and we use that to fill the large areas behind it. Um and then we start putting the VDV on
00:46:23
that. So, like you really bed in from just like a very soft thing to this like very wholesome volume, and then we have the fine detail with the child. Um and then when we go down further, um we get we'll start seeing more VFX uh showing up. Uh the ones that are driven from the the gas cloud itself, uh as well as like the local effects to like help uh space scape the scene. And those are really important cuz they they especially like flying through coil tunnels, it they give you like a really important like sense of scale and like distance around uh the player. Um and and that's something that it's quite easy to lose like where you are within one of these um gas cloud forms. So, yeah, let's drop a light in. Yeah, when when modeling, I think one of
00:47:12
the bigger challenges is to actually try to make it fit you're trying to fade it into the universe. So, towards the edges, you want maybe like a a simulation or you know, identity fading from nothing to to something so it just doesn't look like you've just put a massive ball of gas right in the middle of nowhere and just like yes, this is acceptable. Yeah, this is this is how this is how it is. grounding in because I like this scene that we have at the moment, it doesn't make sense. Like, what is this thing doing here? Um and that's like what we have to do. We have to like build the history of like a system of um a location. Um and like what any environment artist will tell you like that is such an important thing to do when you're making uh like video games. You tell story through the
00:48:02
arts and situation. Yeah, like if you put like an asteroid field like in the middle of the all the sphere and the on your parent, you could maybe say, "Oh, maybe this was a a planet and exploded and it was just mined and then collapsed on itself and it released all of these the gas and dust and this is how it came to form. Kind of thing just give it a bit of a history to it. Yeah. Now, can these VDBs be used for things other than gas? Like can we use these for asteroid fields or Yeah, crystal mountain. I think we're looking at So, like um I mean not not phenomenal but I mean it's as as much as you can think of well, how what other things can you use a nebula for, but the the data in itself that the VDB offers such as
00:48:50
density can be used to help populate a level. So, instead of having an artist hand placing asteroids in a vast areas large as a parent is, they could put use the VDB's density value to start spawning asteroids in a massive vast area and it would you know, you could have asteroids only in very thick areas. So, you'd have asteroids in the more dense parts and then not so many maybe a couple in in the the low density areas. Yeah, we have the mean Oh, yeah. We have the Sorry. We have the asteroid gas cloud density which does that like I didn't want to show it to today cuz a lot of the defaults are squashed and stuff, but it's it we saw an image
00:49:39
of it yesterday and it was really phenomenal. And I I did play around with it. I just put square VDB square and put some rocks in the other day and just like flew through it and it was really it was really fun. I just like dodging around all these asteroids. I mean it could be really cool when it ties into gameplay. So, if we did have things like EM distortion on your on your ship, you know, and you're mining and all the asteroids are in the dense parts of the dense parts is really um uh really dangerous, you know, you could actually lose your ship or you just get hurt just by mining by getting too close, being too adventurous. So, we could have So, we started adding some detail in here. I've just lightened it up cuz it's still
00:50:27
quite thick here. One of the things that we like spend a lot more time on is making sure that we're not getting like too much noise down the voxels. Uh as well as like making sure that we're bedding this parent into the child better. Like I'm really spitballing with this design, but you know, kind of interesting like having this thing And that's kind of fun. Should we fly through it? Sure. See how that looks. And uh just to answer one of the questions I saw that someone say with
00:51:15
simulation time, it's close to real time but in Houdini speaking uh because we don't we're not looking for Pyro solvers or simulation to actually give us the detail. We just want it to give us the density for us to work with. We actually add the detail after simulation. So, with with Fume, so when an artist does a simulation in Fume, they will take it to Houdini and then Houdini can start processing with that cloud detail node to add um the detail that that solar wouldn't be able to ever give. Yeah, and it's like a like it's like we mentioned earlier, it's a a tool in your in your arsenal, right? Like Um all these things and like every different artist has sort of like come
00:52:02
at it and sort of approach approach it from a different uh method and it every every time they do that we learn this new thing. Um and that's like it's really it's really fun. It's really exciting. Um So I hope This makes it more difficult when teaching new artists how to make a nebula. It's like which method do you want to learn? Funny enough, the Jake Jake or Jake's method. Cuz Jake Dainey and Jake Dunlop and then me Jake that and and Ollie that all code VDBs. I I I if if you're called Ollie and you work with um volumetric software, please apply. We we definitely Um how many do we have now?
00:52:51
Three four? Um I don't know how many Four Ollies. Not enough Ollies, I can tell you that right off the bat. We don't have as many as we as we do Chrises. You ever have to go through the address book and Outlook and look for a Chris? It's like oh my gosh, there's so many So I realized that we were actually exporting quite a low res uh VDB so I bumped res up to a bit. Yeah, we'll see when this comes through. Yeah, it's quite nice how we can just export it. Um what happens in the background is that we take this VDB file and then we convert it to our own like special SIG voxel file. Um and that is like a lot more uh
00:53:43
more optimized running engine and we like compress a lot of the values down because we don't need all the values that are in the VDB like the VDB file format is a great file format it's used in like films, games and science but like uh I like it's fantastic but like we have like a yeah cuz we have like 32 bit 64 bit values there and we don't need that for this so we like trim things out and we have to make it better to read so that we can see this stepping here that's coming from that copy paste what I would normally do is actually take that high res version and then add some more detail in cuz what we can see if we look at the volume slice
00:54:31
is that like disconnect in in the voxel sizes so if you can see here we have every polygon here is one of the voxels of the the high res child and then this parent is what we see with the blue squares at the moment and this is actually a really minor example of it but like in other cases we'd have like one blue square which would be like half of this so that's really like where this parent child workflow comes from example yeah it's a good example for the the jump point that that's what you would have seen if we didn't have a a child VDB in there with a much higher resolution around the point of interest Yeah cuz all you see is like this sort of blurry fog
00:55:20
and what we do is that we layer some of the PFX on there to get that like really in smoke like quite close to your canopy detail but between that and just that that big voxel it's there's such a difference so that we need some intermediate stages to do that. Okay, what should I do to others? Yeah, about 3 minutes. Okay, let's do this. You want to take another flight around in the Yeah, sure. That sounds good. But what we could do if um if Ollie handed this his um project file to another artist, they could change the sphere to maybe a a cylinder and run it through the exact same network. You would just get an entirely different results just
00:56:07
by changing the shape or changing the amount of polygons you had on the sphere itself. Oh, we could do that. We could just completely trash this. Well, get us a Let's Let's get a screenshot for the uh Get the money shot. for the for the YouTube thumbnail and for social media before we before we mess it up. No, you trust me to make nice work. I trust you to make nice work. I just you know, it it's It's the non-destructive workflow. It's just But um because it's non-destructive, we actually um allows designers to create uh things like the coil and white box the coil with um VDBs instead of just geometry, they can give Houdini the piece of geometry
00:56:55
Houdini and that's it. They would just hit save and then they get a sort of like a Art would actually create this network before before this. So, it's what Art says, "This is a nice a nice starting point. This is what we were imagining for it." So, when you put your designer geometry in here, this is what it could look like and And of course, it goes with that It should go with that same We're going to say anyway, this is uh less than an hour's work. of what we've done. So, so, so, so what what what what we're showcasing here what we always showcase on Game Dev is not is not even a close approximation to what can actually be delivered with the with the tools at our disposal and uh the work of many folks on the team going through and optimizing and cleaning up Yeah, let me show you some of the other
00:57:44
examples I did quickly. This is the one we did earlier today. Um so this you can see this big parent here and then we waxed this uh little station in down here. Um and this same workflow we have this side with a small local detail. Uh you couldn't do it in the big one. Um but like this is another like really challenging one as well. It did We did them all of them. Um and there was also this one which kind of interesting. This one has some quite nice details. Um my my local trials aren't very exciting and this one's out of place. What Why are you over there? Um So this this whole time we've been working in a level that actually had multiple gas clouds and stations going. I mean just in case like things went
00:58:33
wrong. I had backups, guys. I mean they went like partially wrong a couple of ways, but like we didn't have to resort to plan B. Uh and then let's see other guys. Really it's saying 3 hours you made three different nebulas. In three different locations. Um and this one this one actually takes the pig head as an example. Uh I don't know if you can like make out what I did to him. Poor guy. Like yeah. I see it. Yeah, you see the eyes as the light looks like it's the eye of it. Yeah. And then I think the um ears they say it's And then I just stuck another station in. Uh it's good fun. Uh and it's just it's just really interesting. Oh, we can actually see the particles here. I don't know if this is
00:59:21
going to come through on the stream. Uh but if you if I like move through you can just see uh particle effects. I can see them. Um so when we like fly through, we'll be able to see those. Um and that just adds so much more deeps. Ooh. Oh yeah, there's a lot of turbulence on as well. So Oh yeah, you can go and walk. Yeah, you can use density to drive forces as well. Makes it a dangerous landing zone. It's quite interesting, isn't it? Vape station. Welcome to Vape station. Jeez.
01:00:13
You really piled it on on this one. Just every citizen with a with a vape. Just walking through the station. It's the requirement. All right. All right, we're we're at time. Let's go ahead and wrap this up before it degrades any more. you can go ahead and stop the screen share and then uh if you can, if you if it will just There you go. Oh, now you're in Now you're in the big screen. Hold on. Now you're Now you're in the big screen. Now I'm Now I'm covering you. Hold on. What if I just There's no You moved places, so sorry. So you're You're just You're just missing Jacob. Sorry, Jacob. Uh so yeah. So Uh thank you so much, guys, for taking the time to be here uh on the show with us to I got to get get get Jacob in there
01:00:59
somewhere. I'll just I'll just get rid of me. There we go. Hold on, we'll do We'll do this. Hi, everybody. So hi, everybody. Thank you for taking the time to be here on the show with us this week. Uh remember uh three Alpha 3.9 is live uh on on the servers right now, so check that out if you haven't already. And um uh Oliver, Jacob, thank you so much for taking time out of your out of your out of your Friday at the end of the week to be here and take us through your process. It's actually been pretty nice to like share this with us today. This is this is my best angle ever. You switched places so I had to move. So yeah, so for Stars in Life, that's Oliver. That's that's Jacob. I'm Jared. We'll see you next week everybody.

Transcript from YouTube’s automatic captions — punctuation and Star Citizen jargon are approximate. The video itself is hosted by CIG on the official channel.