This task will document the unending cycle of trying to find out how we can create an Avatar that is not simply two images, but an actual 3D model that can be rigged and animated.
Prior work to this includes:
- three.js is already included on master
- I have a branch where I've loaded a model in the space where the avatar currently goes (https://marisa.cloud/source/tometo/history/temp%252Fthree/):
- In my local checkout, I've updated the aforementioned branch to work with the refactors since then. I'll probably push this to a separate, *actual* WIP branch.
What remains the most important question is not how to get a model loaded onto the website, but how to "map" images onto that model. Those images would be the mouth, and potentially other parts of the face (but most importantly the mouth). I currently have no idea how to do this, and I would welcome anyone who knows their way around using 3D modelling software (I'm using Blender locally) to help me figure this out.
Once that's figured out, I would create some abstractions around this image-changing functionality. It's likely that I'll make a simple model that will serve as a stand-in for now, and temporarily remove the two-avatar-images functionality because it wouldn't serve any purpose then.
We'll see how this goes!