aaronze / protogenos Goto Github PK

We want to support using a microphone for both speech-to-animation and "rave mode"

From initial investigation, I suspect the problem is that activating too many LEDs at high brightness draws too many amps which crashes the Teensy board. From eyeball testing, I'd estimate when the LED boards are at 30% capacity (Every pixel at white)

It's not necessary but it increases compatability when switching between this OS and Coela's.

Things like blink mechanics don't need to be implemented for every controller, we should pull out that into an attachable module/effect

We want to add support for using the APDS9960 boop sensor

Being able to change brightness and colour theme on the fly would be good

Right now the project is using pixel space for the HUB75. That's been useful, but it will fall apart when rendering on other devices. Normalizing the space to be x -1 to 1 and y -1 to 1 will make it consistent, if awkward, to render on other form factors. It should also allow us to remove passing down width and height to components that have no business knowing the panel specifications.

Right now the camera is strictly orthographic and built into the Renderer. Using a camera object that has camera parameters like Perspective/Orthographic, Screen size, FOV, etc would be good. This would also allow for multiple cameras to render the scene to different panels.

To make importing OBJ's/FBX's/meshes as seamless as possible it would be good to be able to automatically detect the best rotation and scale to maximise the proportion of the face that is viewed. Optimally, the model would already be normalized in a way that this is unneccesary, but that's not always the easiest solution.

Adding a HUD menu on the inside of the visor would be useful for changing settings on the fly

Project files require flattening vertices and triangles into indexes and vertices. It would be useful to be able to import obj directly and not have to worry about the conversion process.

Apparently the fbx format could be directly importable to a morphable face. This could simplify the process of adding faces and their available morphs. Another option is instead using a collection of OBJ's. We might need to be careful as FBX is a proprietry format.

Right now texts are a simple color for every pixel. It would be useful to provide a material instead for textured text