November 19th, 2024 Livestream Q&A: Why not smaller Lights?

November 19th, 2024 Livestream

Q&A: Why not smaller Lights?

00:00 Intro
00:51 Unreal Engine MegaLights

https://youtu.be/S_Qdy6INTno

This question may have been asked previously at least 5 other times, as recently as November 2024 and as early as June 2021.
This question has 2 related videos
Suggested Wiki Reference Code <ref name="yt-S_Qdy6INTno">[https://youtu.be/S_Qdy6INTno YouTube - November 19th, 2024 Livestream - Q&A: Why not smaller Lights?]</ref>

Topics

Transcript

why not smaller lights current lights are huge and hard to use in size medium and smaller spaces so I talked about that last week, but tldr essentially is that we I think we chose this size of lights because lights are quite expensive in satisfactory, and the main thing really is like the volumes, like the amount of volumes so what we're fearing is that if you have smaller lights that means that people are going to place more lights in, the scene and that means that it will be less way less performant than just having one big light so we want to discourage the scourge placing too many lights because you're going to run into frame issues like pretty soon regardless of how like good your computer is and it's it's very hard to optimize away that it's just there's no getting way of getting around it I actually want to mention one thing though, in the latest version of Unreal Engine they they've put up they've introduced an experimental mode of something called Mega lights Mega light, which is a, and it's even more of a rabbit hole because it's actually based on a paper that, my professor wrote when I was studying computer Graphics at schmer, so it's based on like a lot of those, that research that was done here in ch anyway regardless so that's story for another time but essentially Unreal Engine is is is has introduced something called megga lights which is essentially a way to be able to render many light sources and being able to render Shadows for many light sources because Shadows is used typically the one of the most expensive parts of of of light rendering lighting, and it's also one of the most important ones, so in 5.5 they introduced Mega lights which makes it possible to have many lights in the scene, without having too much of a big of an impact in terms of performance, because it's a it's a very different way of of handling many light sources, in 5.5 I believe it's not production ready so if even if we try it out we probably won't be able to use it until 56 which is I mean it might also be 57 when they actually, have it before before I before I go too far maybe I just want to confirm that actually is not production ready yeah it's it's an experimental feature right now so so much like Lumen was also experimental feature for a long time, in like 53 or two I think or wait I'm thinking of something else regardless it could be beneficial for satisfactory, and and and if if it is it might make sense to use Mega lights for, lighting but we haven't done any testing ourselves so we have no idea the thing about these like implementations is that you know they have trade-offs and, it's never certain that they'll work in our game because we might need to make changes to our game to to accommodate for, certain techniques like nanite was something that a lot of people always like why don't you use more nanite nanite actually requires a lot on our end, to be able to set up a properly and you know implementing n out on everything is not worth it, like in both in terms of like, us working on the game and like the time it takes to implement something that effectively might not change the gaml too much for you guys, but also like performance- wise it might not actually make too much of a difference it might just like look nicer if anything it might be worse performance, and that's actually something we saw with nanite for a long time was actually that a lot we thought that we were going to get better performance but when we actually implemented it and tried it out we actually got worse performance, but it looked better so we made that tradeoff at that point where like it's not that much worse and it does like way better so we thought it was worth doing it on some parts of the game but not all, who needs performance yeah right Andrew who does, anyway long story short, mega like because I've gotten this question so many times and I didn't get it specifically here but I just want to address it that like Mega lights might be interesting for us in future right now it's not production ready so we won't be able to to implement for satisfactory we also are like two versions behind on a real engine because I think we're on 53 and now we're real 55 so we've got some catching up to do and that always comes with complications and we don't know if during 1.0 if it's like safe to do these engine upgrades because we've in the past have a lot had a lot of issues when upgrading engine versions, so like it's up for debate if we actually dare to upgrade the engine at this point because we don't want the quality of the game to go down, like in early AIS is a different story because we're still actively developing game but so so yeah hopefully that made sense good and then oh that was the next question I didn't even see that say are you Clairvoyant holy okay cool I just answer the question is one stone nail it yeah