A mismatch in degrees of freedom produces dislocation when moving
Aliases: DoF mismatch · viewpoint decoupling · locomotion dislocation · pose contract failure
What it is
Content built for six degrees of freedom, headset that accepts rotation only: lean toward the table edge and the table retreats with the head, never gets closer. The other way around, a panorama pinned to a sphere played on a headset that tracks position: take a step in the room and the whole sphere comes along, the floor sliding out from under the feet. Both are degrees-of-freedom mismatch. The dislocation shows up at the first movement. It does not show up while the wearer looks around from a standstill.
What is wrong is not “the picture is soft” or “the turn is late.” It is the contract between viewpoint and bodily translation: the body says I have moved, and the world either refuses to change or carries the entire environment along as if it were part of the helmet.
Why it happens
When a person moves in a real room, vestibular and proprioceptive signals report translation, and the retina reports parallax consistent with that translation. A matched system keeps the two aligned. A mismatch means one side is lying. Content expects position, the device does not supply it: the retina misses the near-field expansion that should have arrived, the vestibule reports forward motion — the world is a photograph glued to the face, walking feels like marching in place, and object positions relative to the body no longer line up with the hands. The device supplies position, the content has none to give: the camera advances with the real head, the sphere or skybox follows the origin, and the whole “room” becomes an attachment of the head, the floor’s place relative to the feet rewritten every frame.
Dislocation is a spatial-relation error, not a synonym for nausea. People can point and say “the table should be thirty centimetres in front of me, but it is coming with me,” before anyone feels sick. Movement is the probe: sitting still, the contract is never called, and mismatch can hide.
Studying it
Use one short translation as the probe. Cross the conditions: 3DoF content × 6DoF device, 6DoF content × 3DoF device, against two matched controls. The action is “take one step toward the target” or “lean toward the target,” not a head turn.
Independent variables: content degrees of freedom, device degrees of freedom, translation amplitude. Dependent variables: error in the target’s position relative to the body (the hand points at “where it is now” versus where it should be), occurrence of phrases such as “glued to my head” / “I cannot get closer,” whether the person spontaneously steps back after the move.
A post-trial sketch of “the target and me” is often more direct than a discomfort scale: mismatch groups draw the target at a fixed offset from the head, not as an object in the room.
Where it stops holding
When translation is near zero (seated viewing, head on a rest), mismatch can run the whole session unnoticed; that task cannot be used to claim a pairing “has no dislocation.” Teleport locomotion never passes through continuous translation, so the probe is not fired and the problem becomes heading after landing, which is a different issue. Helmet interfaces that are meant to stick to the head (a reticle, a status strip fixed in the view) are not a mismatch; their contract is to travel with the head. Dislocation is a world-claimed object travelling with the head. A few centimetres of unplanned sway, under matched 6DoF, is ordinary parallax and should not be tagged as a fault.
Applying it
- Align shipping with device capability: do not open a 6DoF scene as the default on a 3DoF headset; do not present a spherical panorama as a room one can walk into on a 6DoF headset. Branch on a capability check before launch, rather than letting the first step reveal that the contract is reversed.
- If rotation-only content must play on a position-tracking device, turn the position channel off for that segment, or make the content an explicit head-following viewer. Do not leave the channel half-open and let people take a step.
- Call translation in the first teaching beat (a lean or a sidestep) as a contract test. On failure, stop there; do not send someone through a whole level already dislocated.
- How to check: take one step in a matched pairing and in a mismatched pairing, and have the person point to the target’s place relative to the chest. If the pointed location travels with the head, or stays in the same patch of the view no matter how they move, that is dislocation, not “still getting used to the controls.”