World Anchor from SpatialTapGesture ?? At 19:56 in the video, it's mentioned that we can use a SpatialTapGesture to "identify a position in the world" to make a world anchor. Which API calls are utilized to make this happen? World anchors are created with 4x4 matrices, and a SpatialTapGestures doesn't seem to generate one of those. Any ideas?

In ARKit for iPad, I could 1) build a mesh on top of the real world and 2) request a people occlusion map for use with my application so people couls move behind or in fromt of virtual content via compositing. However, in VisionOS, there is no ARFrame image to pass to the function that would generate the occlusion data. Is it possible to do people occlusion in visionOS? If so, how it is done—through a data provider, or is it automatic when passthrough is enabled? If it’s not possible, is this something that might have a solution in future updates as the platform develops? Being able to combine virtual content and the real world with people being able to interact with the content convincingly is a really important aspect to AR, so it would make sense for this to be possible.

I would like to know whether it will be possible to access high-resolution textures coming from ARKit scene reconstruction or by having access to camera frames. In session 10091 it appears that ARFrame (with its camera data) is no longer available on visionOS. The use cases I have in mind are along the lines of: Having a paper card with a QR code on a physical table and use pixel data to recognize the code and place a corresponding virtual object on top Having physical board game components recognized and used as inputs: for example you control white chess pieces physically while your opponent's black pieces are projected virtually onto your table Having a user draw a crude map on physical paper and being able to use this as an image to be processed/recognized These examples all have in common that the physical objects serve directly as inputs to the application without having to manipulate a virtual representation. In an ideal privacy-preserving way it would be possible to ask ARKit to provide texture information a specially-defined volume in physical space or at least a given recognized surface (e.g. a table or a wall).