Crowd Rendering
Last updated
Was this helpful?
Last updated
Was this helpful?
The animated crowd is our solution to handle rendering massive scales of players in an experience in a performant way. Instead of using a full "Unreal Actor" per player in the world, we replace distant actors with a highly optimized "animated crowd member", and swap out players between real actors (LoD 0) and the crowd as they move closer or further away.
Crowd members do not have a real skeletal mesh in the world like a full actor would. This means you cannot add things to the Skeletal Mesh such as collisions, attached UI, sound emitters etc. This means if your project is configured to use 30 full actors and 10k crowd members, things like collision checks, raycasts etc. would only be possible on the closest 30 players. This should be considered whenever designing gameplay features that rely on players having a full actor representation.
The legacy is for rendering large numbers of characters that use Unreal Skeletal Meshes, rather than MML. It can still be used for rendering non-MML avatars and gameplay attachment that use Unreal meshes.
See the linked document for further information - the remainder of this documentation refers to the Crowd Rendering system used for MML avatars.
The Crowd Rendering system supports rendering thousands of unique MML player avatars at once. It uses automatic level of details to render arbitrary numbers of player avatars with a fixed triangle and texture budget, changing the level of detail of each avatar to best use the available budget.
For further details on how to use and configure the system, see these sections: