XR General Grab Transformer not working

I was trying to add one- and two-handed interactions to a MR scene.
The behaviour is strange:

  • with a single hand, objects can be moved and rotated.
  • as soon as a second hand is interacting with the same object, the object doesn’t rotate (freezed position) and doesn’t scale.

My assumption is that the “attachment points” where the transformer attaches to are both in the center of the object and so nothing works.

Is there a way to get this to work without having to reimplement the logic?
For reference, this is already broken in the “Manipulation_XRI” scene from the VisionOS samples.

I have the same question.
I added XRGeneralGrabTransformer.cs to GameObjects in Manipulation_XRI.unity in PolySpatial 1.2.3, but it didn’t work.

However, I found that XRGeneralGrabTransformer is working in a sample visionOS/VolumeDemo.unity in XR Interaction Toolkit 3.03.


The structure of the hierarchies are different. I guess some components needs to be attached on some GameObjects to make it available. I didn’t find the solution yet. Please take a look at the visionOS/VolumeDemo.unity in XR Interaction Toolkit 3.03, anyway.

Indeed the visionOS sample in xri 3.0.3 has the best practices for using xri with shared space input on visionOS.

Note that far interaction with two hands is not in an ideal state because scaling and rotation is implemented using attach transforms anchored to the target objects and vision os doesn’t put the interaction position exactly where you’re looking, which means the offset between them is extremely minimal.

I have some planned improvements to the general grab transformer to better support manipulation using only the relative motion between the pinch position.