Inside Star Citizen: Preparing Prototypes

Transcript — every row opens the video at that moment

1. 00:00:00

   foreign to another episode of inside Star Citizen our weekly look at the behind the scenes development of what my mother calls that video game Jared's making whenever she explains it to her friends it's quarter three and if you've been following this show in any of its incarnations for a while now you know that the road to Citizen Khan is a very interesting one for our weekly video content as we skillfully avoid all the Hot Topics being saved for the big now two-day event and use that challenge as an opportunity to explore aspects of our development that might otherwise never get a chance to shine it's an opportunity this quarter to look at how rastar is being used to remaster existing ugfs in Alpha 320. uh to explore racing a little bit and a new ship that readers of jump Point magazine

2. 00:00:49

   in our social media channels probably already know is on its way an update to Claudius and how it's being used to create even more immersive environments heading your way and then a spin around the whole bfx department to see what they're up to Yes Virginia it's true there will be fire we can do all this and more as we navigate our journey to the big two days in October but up this week and my reason for sitting here at my desk dumping just a little bit of scl into our ISC is to talk to Torsten and Jacob from the eupu feature team about the prototyping process how they test design and programming implementations at the earliest stages and how it relates to munching an important component of Salvage and my nightly routine just 20 minutes before I go to sleep person Jacob how you doing people doing

3. 00:01:41

   good I am fine thank you all right so we are here talking about uh prototyping process um before we get into it uh give me the top level just the just the the like the LinkedIn version answer what is a prototype and why do we do it a prototype basically can have several facets in game development and for us particular it is always about gameplay and in this particular case it is about the technical the technical problems that we might face and we might have to solve so for that we always use prototypes and uh play testing those prototypes is a crucial part in game development where it is mainly used to prove theories whether theory is either gameplay is fun

4. 00:02:30

   does it technically work and this is basically what what the the purpose of a prototype is anything you want to add Jacob from from more of a Dev perspective uh we tend to take what the designer's vision is for uh gameplay and then we've got to figure out how can we achieve that in the game and just as often that's not crystal clear from the start so we have to try a few things and see what technical hurdles might be ahead of us that we're going to have to consider for the full development so what we're going to do on this week's show the first of our quarter three shows where things are a little bit different we are going to look at a feature at the absolutely earliest stage this is earlier than we've ever shown any feature ever the Prototype phase uh

5. 00:03:19

   what we're about to show you is not going to look good it's not going to look pretty uh it's going to be weird Poppy and buggy so we're just warning you ahead before you take the footage and put it in your YouTube videos uh what is the Prototype that we're gonna look at right now Jacob we're going to be looking at the prototype for uh monitoring the next tier of Salvage the idea being we need some way to take apart a larger ship and break it into little pieces so that you can then pick it up in your Salvage trips gain some material from that and you'll be able to sell that later but we're going to be looking at just the breaking a ship apart phase which is what we're prototyping here right so I got it queued up here let's take a look at this here so obviously this is a vulture approaching a Gladius yep

6. 00:04:07

   um so with the Prototype we've added in a new monitoring mode sub mode to salvage so all of that whole scraping uis disappeared and we're going down to just what we need for this prototype which is the ability to break a partnership which at the moment is going to look exactly the same as if you blew it up with your guns but in future that would look rather different what's happening right here I see the bar filling up on the left okay with the ship yeah so this is the placeholder for the actual gameplay that would happen here the design will have some actual gameplay for what you have to do as a player to break the trip apart but we put in that little placeholder which is just essentially a timer a progress bar based on how big the target is and that is actually on on purpose because

7. 00:04:55

   the question that we are asking in this prototype are more technical that means that we don't have we shouldn't focus on the gameplay aspect so we should keep the gameplay as minimal or as simple as possible and then so that we can fully focus on the technical bits of of the Prototype okay so we saw it we saw it broken into some you know what look like normal debris but now we got to break it into smaller munchable steps we'll go back to the video here so breaking it into normal debris is stuffy that you've seen before but when we have this bigger part it doesn't come apart into more pieces naturally so in this prototype we are uh just deleting the thing and replacing it with loads of tiny little placeholder pieces to act as the monster Department's debris uh what you're also

8. 00:05:44

   seeing is a prototype for how you will get those pieces into your cargo which is a little suction feels it grabs those little pieces drags them towards the grinder in the mouth of the vulture and from there they're converted into cargo there's a little pieces that we saw are obviously not arted and everything I want to remind people again we've literally never shown anything in a prototype Facebook Facebook so forgive me for reiterating this two or three more times before we're done here this is designer art just small pieces to represent the kind of the mass and the size of of stuff that we might be using it's even worse it's colder art yes indeed this is this is the programmer equivalent of a white box test uh where I'm not even qualified to make white boxes so I just rip off other assets

9. 00:06:33

   that I found in the engine and just slap them together so that's what this is and obviously without being a prototype there's no VFX there's no explosions there's no beams there's none of the things that would mask this transition that would normally happen you know so that you don't see the the pieces just uh just pop in you know like like they're currently doing yes um all right so that was done on your local machine uh just you know you on your workstation doing an a local version of of the the universe here uh the next phase as I understand it is to test it in pu conditions and we're going to go ahead and start this video because this video is going to go for a little bit tell me what we're doing here yeah so I'll tell you what we're doing and then we can discuss later why we're doing that exactly um what we're doing right here is uh I

10. 00:07:21

    want to test a worst case performance conditions so I'm chucking in some console commands to load up or stream in uh several different Landing zones at the same time uh because otherwise me is just one player on the server I wouldn't really be loading that much I wouldn't be stressing the server so I'm gonna stress it quite a lot by loading up as many Landing zones as I can think of off the top of my head basically until I see that server FPS count go appreciably lower [Music] um and then only once I've done that am I going to step into the actual gameplay test I also noticed you've upgraded from a Gladius 2 and 890 jump yes because when I think worst case for breaking a

11. 00:08:08

    big thing into small things I want the biggest possible thing which is going to require lots and lots of small things so yeah 890 jump is a good candidate for that okay so what are you typing here oh yeah uh I'm yeah just typing in some comments for the uh other recipients of this video because this was uh an internal video for you know reviewing stuff do you have to use console commands to leave notes for people or otherwise uh but it's the most convenient so all right so we've got so we've got the server simulating pu conditions all the all the landing zones are are loaded at once instead of you know streaming in and out like they would normally would so it's actually worse than you would see on a normal pu in date and we're going to see that

12. 00:08:56

    immediately that uh this breaking the trip apart uh that we saw work so nicely in the editor well I've really annoyed the server here because all of these pieces uh yeah popping in in a not so great way and I make a bit of a sarcastic comment about it in the uh in the console um but uh yeah it's important to create these conditions you know to test these things not just in ideal conditions not just the ideal Universe of your own computer or even the Pu when it's like first loads and nobody's in and everything's great no indeed what I'm testing here is worse than we would hope it ever to be we don't want it to ever be this bad but we gotta test it because might be I'd say now beams we're going shooting out starting to break apart

13. 00:09:44

    yeah there'd be extra gameplay here that'd be fair thanks whatever but in the editor version we saw pretty much as soon as the the big piece disappeared all those tiny little bits turned up well it's not going to go that way in in this situation we're going to be waiting quite some time for all those little pieces to turn up gotcha and so while we're waiting for this uh pieces to to turn up here um let's talk about what we learned from uh what we've learned from these tests so far um obviously there's some form and stuff but there's also some gameplay implications from what we've seen uh touristy you want to talk us through it well I can talk about some parts that we learned from right so it's so the best thing that you can learn from a prototype is actually everything that goes bad because this is the stuff that you can then iterate on and make the the

14. 00:10:32

    gameplay or the the technical stuff even better so um from the gameplay perspective so I will just talk about the gameplay stuff and Jacob can talk about what we learned from technical perspective there is from the gameplay we learned that what about big ships so how for example now you saw the vulture trying to attach to the 890 jump that is a bit of a wonky situation and it doesn't feel right so what what about vultures actually munching bigger ships so how can we make that an interesting gameplay then you saw in the first video what happens when uh the the ship transitions into the smaller bits so the masking is is a huge topic that we have to solve so that it

15. 00:11:20

    feels right and you don't notice that one element pops away and another element Pops in so we definitely have to mask that then uh what you might have seen in the first part is the conversion of materials I don't know if you paid attention to the filler station bar that basically fills the SEO crates on the back side of the wheelchair but that feel drastically faster than it would normally do with Hull scraping that means that it would be more painful for you to munch because you have to stand up from your seat more often go down to the filler station remove the boxes so this is definitely an issue that we then have to solve because we really don't want to have this gameplay to be

16. 00:12:08

    more tedious than it's has to be so we have to find a solution how we can make it more like a yeah much nicer for for the player to actually remove those crates there then uh the navigation around the pieces is also something that we notice isn't that fun so we will probably utilize the tractor beams there because we are anyway working on them but maybe that makes it much easier for players instead of flying around and getting the pieces between the fork of the wheelchair and then having it disintegrate but instead you pulling the pieces in with the tractor beam another Fork of the wheelchair is very limiting in terms of what pieces will fit so this is also something we learned

17. 00:12:56

    and a big thing is uh what which already is standing out is how can we make it fun and unique for the reclaimer so we just we're looking at the Prototype at the vulture and the big question that our team was asking in the end of that prototype was like okay how can we facilitate the claw and the unique neatness visuals or the uniqueness in general of the reclaimer to actually facilitate that gameplay and have it stand out and be meaningful and fun all right so a lot of a lot of gameplay a lot of gameplay implications from just that little prototype uh Jacob what would you what was your takeaways uh from the from the tech perspective I'm looking at things like the performance and also uh just

18. 00:13:45

    um what am I going to have to account for to achieve the designer's vision here I know that we're going to want things like the time that it takes to destroy something to be somewhat proportional for its size and mass but from the initial prototype I could already see that uh a bigger piece taking longer it already takes you know a few seconds with the Gladius if we're going to scale that up and scale up the time as well proportionally to a to a big Target we're going to have to think about what equation are we going to use for scaling that so that it doesn't take you ages uh to like sitting there for minutes to to break something apart I got to think about how I'm going to hook in the visual effects and the audio whatever what dependencies am I going to need to

19. 00:14:35

    make well get involved speaking of you know judging time uh The Entity system is taking a little bit of time to spawn all the spawn and debris we've been talking uh this is still going at the moment obviously there's going to have to be some optimization work there with the entity system this is a kind of what we've got with with PES is that uh it's not quite as quick to create entities as before we had PS um and you we have all sorts of great stuff that comes with PES so this is just a small limitation that we have to learn to work around uh but it tells us that gameplay ideas that we had before PS might need some adaptation in order to work efficiently uh in the the post PES world

20. 00:15:25

    so yeah so speaking of which let's go back to the video I think we're about to get the entity spawns now and he's saying this yeah any moment now yeah there we go there we go and then play a bit of a flicker as they come in but they're finally here so yeah this this tells us well in this worst case scenario uh this performance is obviously unacceptable uh we're going to have to do something about this so we are thinking about uh how can we change our approach how can we optimize what we've got discussions in various different directions but this is a key takeaway from this prototype because if it had gone differently if if we tried this and the piece is turned up in just a few seconds or so then we'd have known that this approach on the whole was probably acceptable and with a

21. 00:16:14

    little bit of optimization it would work but that's not what happened the stuff took minutes to arrive and that means we have to do some more serious rethinking about how we're going to approach this the vision and the the basically the acceptance criteria or directors Define they they they are still untouched right so those are still to be fulfilled and even the overall Vision that we as the team defined for how that gameplay should play out those those will be like those are already set in stone because we all agreed on those but how the actual gameplay then plays and feels this is still affected by uh by the Prototype a prototype like this is really perfect for for stepping in early and making sure that every every

22. 00:17:04

    tick box on our acceptance criterias has met and uh we can make a better gameplay out of it I I think we all have a a a real uh desire and responsibility towards you know testing these things as often and as early and under as much duress as possible uh to as we work to stabilize the Pu and make it a more enjoyable experience going forward so I torson uh Jacob thank you thank you for hanging out with us today thank you very much bye bye thank you

23. 00:17:57

    so what do we learned this week well we learned that prototyping doesn't really look like much of anything but it's an essential step towards proving out design and performance before all the artists and other devs go in to do their thing that it's just as important to prove out good ideas and good paths to take as it is to encourage developers to seek out another way towards success and that it's important to continue refining not just the features and content being made but the very way that we test those things earlier and earlier and under a scale of duress for inside Star Citizen I'm Jared Huckabee thank you for letting us share the process of game development with you and we'll see you all here next week