Shader Not Working on Unity Polyspatial and Vision Pro

I have created a shader in Unity 2022.3 that makes objects disappear as the camera gets closer. It works perfectly in Unity normal scene, but when use the shader on unity for vision pro configured not work, and more when tested in the Vision Pro Simulator, the objects disappear entirely across the whole scene. Why does this shader not work with the Vision Pro camera? Other shaders, like moving water, function correctly. Here’s my Shader Graph setup:
. Any suggestions?
.

If it’s possible either to submit the shader graph in a bug report and let us know the incident number (IN-#####), or to post the .shadergraph file, or just to post a screenshot in which we can read the node names, that would help us debug. Off the top of my head, I can think of at least one possible issue: it looks like you might be using Alpha Clipping. If you are, you should be aware that visionOS treats alpha clipping as all-or-nothing when enabled (that is, as if the object were opaque); if a fragment has an alpha lower than the threshold, it will be invisible. If it’s higher, it will be opaque. I’d recommend trying it with alpha clipping disabled.

“Definitely, thank you very much. It would be incredibly helpful and valuable for me if you could accomplish what’s shown in the attached file and video. I believe I’m not using alpha clip correctly, but it’s not working for me. Here is the original file with the shader, along with a photo and the video. Thank you again!”
video:
https://youtu.be/zShxd4vxNRQ

here the unitypackage full acces:
GIVE_shader_400PPP.zip - Google Drive

Thanks! I didn’t realize that you meant for the shader to be opaque, so ignore what I said about alpha clipping.

The issue here seems to be with the tangent space position. If you just delete the node on the lower left (Position with Space: Tangent), it seems to work as expected in visionOS. By definition, the tangent space position of a fragment is (0, 0, 0). However, it seems that when we convert to MaterialX (and generate an ND_position_vector3 node with space = “tangent”), visionOS is generating some other kind of position for the node. I’ll make a note to fix this by just generating a constant (0, 0, 0) for tangent space position.