I'm interested in experimenting with eye (head) tracking to customize the vertical viewpoint for a small number of users: I assume each user is in a distinct and non-overlapping horizontal range, tracked by e.g. Kinect. I would like to render the quilt in real time, adjusting the projection matrix for each quilt element according to the height of the "current" viewer. Is the current renderer modifiable to handle this? I'm mostly interested in WebXR to start with; the code for the WebXR polyfill renderer looks like it could be customized, albeit calculating the projection matrix will be a bit more complex and probably more costly I experimented with something similar for multiple WebGL displays (single person view in those days) using Xbox Kinect: https://www.youtube.com/watch?v=fKX84boIm0g . At about 4sec mark the view/camera "ducks" down to view under the car. I used Robert Kooima's off axis projection method: https://www.semanticscholar.org/paper/Generalized-Perspective-Projection-Kooima/14d1b312aba825bcce17edd67e3fdc139f1a76a2 Discord post: https://discord.com/channels/573562186597662720/573562186597662722/1243569039138099291