I’m assuming math.noise() counts as external/random input?
math.noise is fine, as long as the parameters aren’t coming from smth like math.random() without calling math.randomseed() math.noise is deterministic.
Yes this feature was amazing what ever happened to it are we still getting it in the future?
I’d imagine this was more a proof of concept / prototypical implementation. It’s nontrivial after all, especially considering performance on low power devices e.g. mobile
Would be cool to see but I can understand if we wouldn’t get it anytime soon
Both Future Is Bright phase 3 and procedural skies are something we still very much want to do but after prioritization of other rendering projects we realized these are unlikely to fit into 2019, which is why they aren’t on the roadmap.
When we inevitably get to the faster Lua VM, will we get some details on that? I’m curious about what exactly is being done to it to make it faster and if any practical changes to the language are occurring.
@zeuxcg You had a thread a while back about getfenv and setfenv and it mentioned a performance hit with using it. Is that something that’s related to improvements to the Lua VM, or was that totally separate?
Really can’t wait for custom materials, i think it would be amazing, because you can insert any image file and make it into a material like unity and unreal engine.
I see a lot of good things on here but I can’t wait for those custom materials
The plan for this year is to not have any changes to the language and preserve syntax and behavior of existing code as much as humanely possible. We will have optional syntax for type annotations as part of “Typed Lua”, but currently aren’t planning any other syntax changes.
Having said this, it’s possible that we’ve missed something obscure so we’ll see how it goes
And yeah this thread was related to the performance work. It currently looks like we’ve found a way to work around the issues with fenv and preserve behavior of code that uses it, but as mentioned in that thread we’ll be advising against using these functions since they won’t be compatible with the typing effort and will cause new interpreter to hit some slow paths.
Aww, that’s a bit of a shame. Specular highlights was something I was really looking forward to this year
I hope this (Terrain Vegetation) will have the ability to be disabled, for certain maps that might use the grass terrain with a different color for something other than grass, or for maps like mine where I am making my own custom vegetation, with its own meshes and rules of placement. I think just a global default vegetation might clash with the custom style of some maps.
There’s been a bit of miscommunication internally wrt rendering roadmap, we’ll have better specular highlights from the sun as “phase 2.5” and this hasn’t been accurately reflected in the roadmap - just not the full phase 3, other things like terrain vegetation and custom materials took priority.
The goal is to have as much ability for developers to configure this as possible by adjusting colors/density/variation and providing custom vegetation meshes/textures. Of course this will be optional as well. We haven’t implemented this yet so details are up in the air, for example it’s possible that we will provide support for custom meshes but the polycount limits on the meshes will be very strict.
I would say that’s a pretty good call. I can’t speak for others, but I would MUCH prefer custom materials and terrain vegetation over full phase 3, especially if we’ll still get specular highlights in some form even without it.
Sounds great! As long as moving forward the team is aware of how important options and more importantly the ability to opt out is. But I am really excited, about this!
Hey Zeuxcg - could you elaborate on what the ‘Custom Materials’ item includes, or link to a previous comment on it? If this is what I think it is - user-generated custom textures as materials with custom spec and normal maps (created and authored by us devs), this is incredibly exciting to me.
That’s what you think but I believe this item is supposed to be in Q4, not Q2. We’ll fix the rendering roadmap up soon.
Future is Bright Phase 3 indeed will slip into 2020 according to current calculations
That’s absolutely fantastic, and a great step towards making Roblox feature equivalent to more traditional game engines. Thanks for replying!