When the eye pursues a moving object, smear is paint on the retina

What it is

When the eye tracks a bird in flight, the bird is steady on the retina and the background slides. In a headset the “bird” is a pixel that lights in frames: the eye follows continuously, the pixel stays on at last frame’s location, and that point is painted into a short line on the retina. That is retinal smear (hold-type blur) under pursuit. It is not liquid-crystal lag, and it is not the tick-tick of a refresh that is too low. It is the geometry of “the eye is moving and the lamp is still on”.

Saccades are brief, so the paint window is short. Smooth pursuit fills the window with the whole persistence interval, which is why smear is most obvious when the eye is following.

Why it happens

Smooth pursuit holds the target’s retinal image nearly still. A hold-type pixel does not move relative to the visor during that interval, while the eye keeps rotating, so the pixel’s image sweeps an arc of “persistence × pursuit angular speed” on the retina. A high-contrast edge becomes a ribbon; ribbon length scales with on-time and angular speed, not one-to-one with how many times a second the panel refreshes. Refresh can be high; if each frame is still fully lit, the paint remains.

The analogue is camera motion blur: shutter is pixel on-time, relative speed is the eye’s pursuit. The eye has no electronic shutter, so the only levers are shorter illumination or less retinal slip of the object. Slow LCD response stacks a second, true pixel trail. The two blurs look like one and have to be separated: photons still arriving, versus the crystal not yet there.

Stare at a fixed point and turn the head, and the same paint lands on world-locked edges — the “pursued target” is then a wall that was supposed to stay nailed in space.

Studying it

Have people pursue a high-contrast thin line of known angular speed (eye tracking to confirm pursuit, not catch-up saccades), vary persistence, and measure reported ribbon width, or estimate retinal sweep from afterimages or high-speed film. Control: the same motion under jump-to-fixate instructions (no pursuit).

Independent variables: pursuit angular speed, illumination duration, contrast, pursuing the target versus holding gaze while the background slides. Dependent variables: reported smear length, pursuit gain on the eye tracker, classification as “blur” versus “ghost”.

Without an eye tracker, at least use two instructions — “follow with your eyes” versus “stare at screen centre” — or saccades and pursuit stay mixed.

Where it stops holding

If the target jumps and people catch it with saccades, the paint window is short and the complaint shifts from a ribbon to a misaligned next frame. Low-contrast, already-soft content masks the ribbon. When one eye’s optics are already diffuse, extra smear is hard to name on its own. Lab high-contrast hairlines overestimate noticeable smear on everyday UI blocks. Footage already shot with a cinematic shutter carries one layer of motion blur; the display paints a second, and the two fuse so blame is hard to assign.

Applying it

• Take high-contrast hairlines, text, and world-locked frames off paths the eye must pursue for a long time, or drop their contrast against the background.
• Keep moving objects’ on-screen speed low, or let people designate with a gaze point rather than chase a flying trajectory with the eyes.
• Do not file every “it smears” as panel response. Split a pursuit ribbon from a crystal trail that is there even when still.
• How to check: have the eyes follow a vertical line across the field at constant speed, then watch the same line pass while gaze is held at centre. A ribbon only under pursuit is retinal smear. Drag under both instructions, then look at pixel response.

Related

• Same group: N1.12.1 Refresh rate jointly determines flicker visibility and motion continuity · N1.12.2 Low persistence trades shorter illumination for less smear · N1.12.4 Raising refresh rate lowers available brightness
• Nearby: N1.05 Motion-to-Photon Latency · N1.11 Render Latency and Motion-to-Photon Time
• Search terms: retinal smear · smooth pursuit · hold-type blur