Class ARSessionOrigin
An ARSessionOrigin
is the parent for an AR setup. It contains a Camera
and
any GameObject
s created from detected features, such as planes or point clouds.
Inherited Members
Namespace: UnityEngine.XR.ARFoundation
Assembly: solution.dll
Syntax
[DisallowMultipleComponent]
[HelpURL("https://docs.unity3d.com/Packages/com.unity.xr.arfoundation@4.2/api/UnityEngine.XR.ARFoundation.ARSessionOrigin.html")]
public class ARSessionOrigin : MonoBehaviour
Remarks
Session space vs. Unity space
Because an AR device will be used to drive the Camera
's position and rotation,
you cannot directly place the Camera
at an arbitrary position in the Unity scene.
Instead, you should position the ARSessionOrigin
. This makes the Camera
(and any detected features) relative to that as a result.
It is important to keep the Camera
and detected features in the same space relative to
each other (otherwise, detected features like planes won't appear in the correct place relative
to the Camera
). The space relative to the AR device's starting position is called
"session space" or "device space". For example, when the AR session begins, the device might
report its position as (0, 0, 0). Detected features, such as planes, will be reported relative
to this starting position. The purpose of the ARSessionOrigin
is to convert the session space
to Unity world space.
To facilitate this, the ARSessionOrigin
creates a new GameObject
called "Trackables"
as a sibling of its Camera
. This should be the parent GameObject
for all
detected features.
At runtime, a typical scene graph might look like this:
- AR Session Origin
- Camera
- Trackables
- Detected plane 1
- Detected plane 2
- Point cloud
- etc...
You can access the "trackables" GameObject
with trackablesParent.
Note that the localPosition
and localRotation
of detected trackables
remain in real-world meters relative to the AR device's starting position and rotation.
Scale
If you want to scale the content rendered by the ARSessionOrigin
you should apply
the scale to the ARSessionOrigin
's transform. This is preferrable to scaling
the content directly, which can have undesirable side effects. Physics and NavMeshes,
for example, do not perform well when scaled very small.
Properties
Name | Description |
---|---|
camera | The |
trackablesParent | The parent |
Methods
Name | Description |
---|---|
MakeContentAppearAt(Transform, Quaternion) | Makes |
MakeContentAppearAt(Transform, Vector3) | Makes |
MakeContentAppearAt(Transform, Vector3, Quaternion) | Makes |
Events
Name | Description |
---|---|
trackablesParentTransformChanged | Invoked during Application.onBeforeRender whenever the trackablesParent transform changes. |