Narrowing the field of view can reduce discomfort
Aliases: FOV restriction · tunneling · dynamic vignetting
What it is
Push the stick and a dark ring closes around the view, as if looking out from a tunnel. Dynamic FOV restriction narrows the effective field of view while virtual motion is happening and opens it when motion stops. It is a mitigation: it cuts the area of optic flow on the peripheral retina and thereby lowers discomfort. It does not explain why sickness arises. It changes how much of the flow is seen.
A statically narrow headset FOV is a hardware limit. This leaf is the ring of black or blur that content adds during motion.
Why it happens
The peripheral retina is more sensitive to large-area flow. Darkening, blurring, or cropping the edge removes the annulus that drives self-motion illusion, while the central task region still shows the heading. The mitigation roughly couples to how hard the motion is: no restriction at rest, more restriction at higher speed, so the tunnel does not become a resident narrow FOV.
The restriction has to lock to the motion. If the ring lags, flow has already hit the periphery; if the ring lingers after motion stops, people read a broken display and keep paying the spatial cost of a narrow view. Soft edges are less often read as a shattered image than a hard black circle, but a soft edge that spreads too far fails to cut the periphery.
Studying it
The classic comparison is the Fernandes and Feiner style of speed-coupled dynamic FOV: same continuous locomotion, no restriction versus a fixed narrow FOV versus a tunnel coupled to linear or angular speed. Take sickness with SSQ or in-session FMS, and also presence or “did you notice the black edge” — otherwise how much immersion was spent to buy comfort is unknown.
Manipulate restriction amplitude, whether coupling is to linear speed or yaw rate, and solid black versus blur. Record FMS peaks, task time (the tunnel hides peripheral targets), and whether people turn the option off. A fixed narrow FOV and a dynamic restriction are different interventions: the former charges the spatial cost for the whole session.
Where it stops holding
When real walking already matches virtual translation, there is little fake flow to cut and the tunnel is pointless occlusion. On an already-narrow consumer FOV, another ring eats the task region; lower sickness is paid for with wall hits and misses. People who react strongly to black borders or confinement will exit because of the tunnel itself. Restricting only translation and not turning leaks the turning peak. On optical see-through, painting black reveals the real environment; blocking flow with a tunnel is a video-see-through move.
Applying it
- Couple the tunnel to virtual linear and angular speed. Fully open at rest. Do not ship it as a resident narrow FOV.
- Use a translucent or blurred edge, not a sudden hard black circle. Provide an amplitude control; default slightly weak.
- If task targets can appear at the view edge, pull them inward during the tunnel, or bring them into centre before restricting.
- How to check: run the same continuous-locomotion segment with the tunnel on and off. FMS peaks should drop, and a pointing or counting task should confirm peripheral targets are not systematically missed. If misses break the experience before sickness does, reduce amplitude rather than increase it.