I think many people would like their fallback avatars to look like their main, only more optimized. Typically Cryptovoxels avatars are highly expressive with custom wearables. The issue with them is that it was very hard to tell who was who when we all became the same sprite. They enabled low end machines to be able to render 100s of avatars which was useful during large events. ![]() ![]() The avatars looked like plain Cryptovoxels avatars. During a large gathering at the 2020 virtual burning man event I noticed something interesting, some of the avatars from a distance were replaced with billboarded images that changed into full 3D when I walked up to them.Ĭryptovoxels also implemented a similar sysem in 2020. It's a bit jarring to have different realities where you might appear completely different to the person you're talking to.Īltspace is another highly popular social VR platform. See the announcement on ().įallback avatars are the backup avatars you pick per slot for people on low powered systems or Oculus Quest. Recently VRChat pushed an update to utilize the fallback avatars more for performance settings. How VRChat would look on a low end machine. It sucks a lot of the charm from the experience. If you're on an underpowered system you probably have seen many of these grey robots. Here's an awesome video that overviews the vrchat fallback avatar system. Their solution to the performance problem is by implementing a rating system and replacing avatars that exceed a user's settings with a grey robot or 'fallback avatar'. VRChat is the world's biggest social VR platform with over 1M custom avatars uploaded to it. The likeness of a users avatar from any angle will be visually preserved while optimizing down to 2 triangles and 1 texture material. I propose we build a system which takes any VRM file, animates it with a walk cycle (to start with), and generates a sprite sheet out of it. Even some of the most popular avatar builders like () and () aren't all that optimized enough. However, one of the bottlenecks for UGC platforms is that people are not good at optimizing their own avatars. This time the further away an avatar is from another player, the less information they have to send - including details about the avatar itself. ![]() The concept utilizes popular gamedev optimization techniques typically used for environments. I believe I may have a scaling solution for virtual worlds to be able to support the number of custom avatars in a single world an order of magnitude higher than what's possible today. The bottle neck is usually that people are not the best at optimizing their own avatars. :arrow_right: [**Part 1: :arrow_right: [**Part 3: NFT 3D part 1: virtual worlds struggle to scale beyond 60-100 concurrent users per server instance, especially those that support UGC and custom avatars. Description: Technical concepts to scale the number of simultaneous users per virtual worlds by an order of magnitude.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |