Welcome to some more Star Citizen with some summaries of 2 of the CitizenCon Panels – Breathing Life into Physics & Crafting Space. They go over Physics and Cloth Simulation, Signed Distance Fields & gas cloud tech. I am going to highlight the more interesting and useful information.
The presentation talked about the addition of physics simulation to assets and characters.
Starting with Environments – Wind, Wind Entities, Simulated vs Canned Animations.
They prefer to use simulated physics rather than pre-canned animations.
This means that assets will always act appropriately when exposed to environments or other entities, wind, players, npcs, explosions.
They have wind on planets, this can be based on the planet or biome you are in.
They had to remove the old LUA systems and rewrite all of this for a 3d world.
They now have a wind area entity, a sphere of wind that can be completely customized to have localized wind.
Previously wind may have been represented with a pre-baked animation. But now cloth in environments will move based on how the wind is blowing and to anything else dynamic that happens to it, like a player steps on it OR there is an explosion OR thruster near it.
Pre-Baked animations would obvious not react and just repeat.
This extended cloth entity allows for efficient simulation, easy editing, self collisions, volumetric & deterministic simulation, gravity, friction
They have done a lot of work to avoid what they refer to as “poly soup” where the cloth over time might bunch together in an odd way OR be otherwise glitchy.
The system also allow for object, ships and characters to directly interact with cloth, they could hide under it for example, or have a tarp over a ship.
The way the cloth can fold or collapse onto itself is fantastic.
It’s a deterministic sim regardless of framerate or time step size and the cloth will always be able to “recover” from any simulation without becoming a “poly soup” mess.
You may have cloth assets blowing around areas, or if you start a ships engines they may blow over a character or another ship.
We are going to see tons of this in the environment in the future!
They are unifying they way they do character cloth with the environments, so it will work in the same way, receiving dynamic info and simulating appropriately, moving away from the previous pendulum sim.
This new vertex cloth approach allows for much more dynamic simulation and higher fidelity, the cloth will react to character animations MUCH better. They are able to define what parts of the cloth are rigidly attached to the character or armor and which are simulated.
This will also allow for many layers of cloth interacting and they want to give as much cloth on characters as possible as much dynamic simulation as possible without clipping!
Future things – deformables will react to an explosion force but then become rigid, so the hull of a ship having a hole blown into it.
They are getting various performance improvements to the tech and better density.
Signed distance fields will be used in the future too (which are 3d textures that store distance data.), allowing for better conforming of shields to the hull of a ship for example, this will also allow shields to distort and shape around damaged areas and parts of a ship that have been destroyed.
Cloth will eventually be able to be torn, ripper and holes shot through it.
They are adding more soft body collisions and entities too, this is basically things that are squishy… like pillows.
The “Crafting Space” Presentation focused on Gas Cloud tech and these areas of interest in space.
They can have huge clouds and nebula in space.
The issue is that these are massive and explorable, therefore they may be empty
Asteroid fields around planets and moons are great for them, they allow for dense pockets of content that is interesting.
This year they are looking at volumetric areas within clouds and large areas, creating corridors, signposting and pulling you toward interesting spaces.
They can make these corridors and interesting spaces in modular sections, maybe 40km long at a time. These can be pieced together for various missions and vistas. Visual effects and lighting should mean that you don’t recognise any reused parts in the future.
They use chunks to tell the game how to render gas & areas.
There is a hierarchy of chunks, they all sit in a large chunk, then Chunks that are further away can be broken up into larger chunks.
Areas that are proximate to the player are in smaller chunks.
The chunks that are not needed or being viewed are just culled.
This means that the whole gas cloud does not need to be rendered, nor does what is visible have to be rendered at the highest detail even if it’s super far away.
They then efficiently simulate the lighting through the cloud.
Then they can defuse various colors into the clouds
They want these areas to feel dangerous and for you to not get too lost. There is no geometry, walls or ceilings tho.
They use a load of VFX help sign post direction, and help define dangerous areas.
So you may see the movement of gas and objects.
EM interference on your systems, HUD and MFDs or any other screen they want.
Denser areas of gas clouds will tend to give more interference and be more dangerous.
They use signed distance fields, to work out how close you are and apply the effects appropriately.
They still have story / triggered events too, like an asteroid getting struck by lightning and breaking apart.
They use a lot of lightning in gas clouds, they use spline guided particle emitters for efficiency!
Contrails and Turbulence are also affected in gas clouds based on density too.
Everything needs a bit more polish, they are getting the VFX procedurally generated in the future, they want to get everything as efficient resource wise as possible.
Gas Giants may be used using the same tech, as could planetary rings and other areas of interest. Next year they will be looking at even larger scale areas