Elin Höhler spent the weekend pulling together a new demo in SideQuest's Banter showing a full speech to 3D generation pipeline making objects in a multiplayer world.
Höhler walked me through the pipeline on Monday, showing the three step generation process – first processing speech to text, then an image from the text, and finally outputting a 3D model directly into the shared multiplayer world space for resizing and placement.
"It's kind of perfect for a zany platform like ours," explained SideQuest founder Shane Harris. "I think our users are gonna enjoy this, decorating their home spaces, using this for making props, I know some of our users have already been using it to generate clothing items that they then put on their custom avatars."
Our video embedded below goes into slightly more depth explaining the toolset, as recorded on my Quest 3, as Höhler considers newer models in the future producing sharper and better optimized assets, but "playing around with Trellis is like getting your feet wet with what's coming."
The creator example is available if you search "world building" on Banter, available now on Steam and Quest. We've seen other generation tools in platforms like Snap's Spectacles as well as in single player experiences, but this is the first we've tried that has been this successful in a multiplayer VR session.
