The area and speed of optic flow determine sickness intensity

What it is

Texture sliding in the same direction across a large part of the field is optic flow. The larger the retinal area it covers, and the faster it slides — acceleration especially — the stronger vection and the heavier the sickness. A crisp central patch often matters less than the ring of ground and wall washing the periphery. Narrowing the field, adding a nose-guard, swapping ground texture for still objects, are the same move: cut the area or the speed of the flow. They are not an essay on why two senses argue.

The spicy part of speed is change. Steady straight wash of the ground is milder than a sudden accel or a snap turn.

Why it happens

Peripheral vision is more motion-sensitive; a large coherent flow is read as “I am moving”. Area is the solid angle of the coherent region, not how many triangles the scene has: a small central animation barely sickens; a full ring of floor plus skirting washing together is what lifts people. Angular and linear speed convert through viewing distance; near-ground texture at the same forward speed yields higher retinal speed, so looking down at a virtual floor is often worse than looking at the far field.

Acceleration turns “already washing” into “the wash rate is changing”. If the vestibular organs do not supply a matching linear or angular accel, the mismatch is stretched in time. Decel, reverse, and in-place turns flip flow direction in a short window; area unchanged, intensity still jumps a step.

A vignette works because it cuts that peripheral area, not because a darker picture is prettier. Cut too hard and sickness falls while navigation and presence are cut with it.

Studying it

On one path, vary the flow: mask the periphery (dynamic FOV restriction), change ground-texture density, change speed and the acceleration profile, hold path geometry fixed.

Independent variables: solid angle of coherent flow, retinal speed, acceleration, whether texture sits in the periphery. Dependent variables: vection onset and strength, in-session discomfort ratings, how many people finish the path.

Use gaze or head pose to check whether people are looking at the ground — “the scene contains flow” is not “flow entered peripheral retina”.

Where it stops holding

Still viewing of object-only motion (no self-motion component) sickens much less; the area rule is discounted. A very dark, textureless space has no washable flow, and sickness takes other paths (pose latency, for instance). Sitting in a real vehicle whose acceleration matches the flow, large area is often accepted as true. High-contrast lab grids overestimate sickness from everyday low-texture scenes. Monocular, narrow-field devices have already cut area; a further restrictor has little margin.

Applying it

• Treat peripheral flow during high speed, acceleration and turns as the sickness knob: shrink the periphery dynamically, slow ground texture — do not start with the plot.
• Do not carpet the lower field with near, high-contrast, repeating ground texture. If spatial feel is needed, use sparse landmarks, not dense stripes.
• Open a restrictor only in the seconds when flow is strongest, not as a black ring for the whole session.
• How to check: one forward path, change only ground-texture density and peripheral visibility, rate in-session. Area or speed down and ratings down: this is working. Ratings flat: inspect the acceleration profile before stacking another vignette.

Related

• Same group: N1.13.2 Self-initiated motion causes less sickness than imposed motion · N1.13.3 Standing is more prone to imbalance and sickness than sitting · N1.13.4 Discomfort must be sampled continuously in-experience with comparable scales
• Nearby: N1.06 Cybersickness · N4.07 Continuous Locomotion and Discomfort
• Search terms: optic flow · vection · FOV restriction