Vergence distance and accommodation distance do not match

What it is

Look at a virtual cup forty centimetres away: the eyes rotate inward to the cup’s distance, while the lenses are still required to focus at the headset’s fixed optical plane — often one to two metres. The two distances have been pulled apart. That is vergence–accommodation conflict (VAC): vergence distance is driven by binocular disparity, accommodation distance is locked by the display’s focal length. Two coupled ocular responses in the real world take separate paths in a headset.

In natural viewing, near objects pull vergence and accommodation nearer together; far objects send both farther. The conflict is not “stereo is broken.” Stereo is still driving vergence; the focus channel cannot follow.

Why it happens

Vergence is the eyes toeing in or out so the fixation point sits on an object. Accommodation is the crystalline lens changing optical power so the retinal image is sharp. Development couples them into a reflex: vergence demand pulls accommodation, accommodation demand pulls vergence. A stereoscopic headset paints the image on a screen of fixed dioptric power, so the distance of a sharp image does not change whether the virtual object is labelled at thirty centimetres or ten metres. Disparity still rotates the eyes by object distance, so vergence goes to the object and accommodation stays nailed to the screen.

The gap is clearer in dioptres than in centimetres: forty centimetres is about 2.5 D, two metres about 0.5 D, a conflict of roughly two dioptres. The nearer the object, the larger the vergence demand, the larger the gap from the fixed screen. Varifocal, multifocal, and light-field headsets try to let accommodation follow the object — they are unlocking this freeze, not raising resolution.

Studying it

Use a display with an adjustable focal plane, or dual focal planes, and manipulate vergence distance and accommodation distance independently. Measure subjective clarity, binocular single vision, and whether the two responses still pull on each other. The Shibata–Banks–Hoffman line typically takes the acceptable conflict range as the dependent measure, not “does the stereo look good.”

Independent variables: vergence distance, accommodation distance (or their dioptric difference), fixation duration. Dependent variables: clarity reports, onset of diplopia, accommodative response (autorefraction or wavefront), vergence response (eye tracking).

On a real headset, measure that machine’s optical screen distance before talking about object distance. Treating the object’s world coordinate as the accommodation distance, without measuring the screen, computes the conflict wrong.

Where it stops holding

Monocular viewing has no vergence demand, so half the conflict path is closed; monocular sharpness cannot be used to dismiss a binocular conflict. Older adults whose accommodative range has already fallen show a different conflict pattern on a far screen: they already park accommodation farther out, and near vergence demand is even harder for accommodation to follow. Video see-through that also focuses the real world onto the same fixed plane puts real near objects in conflict too — it is not virtual objects only. Optical screen distance varies with pupil position and lens tolerance across people; the spec’s “two metres” is not two metres in every eye. Conflict is a geometric and optical fact. Whether it is reported as discomfort is a further layer.

Applying it

• Write down the optical screen distance of the current headset (from optics or platform docs, not guessed from world coordinates). The dioptric gap between an object’s vergence distance and that screen is the conflict magnitude.
• For objects that must be fixated for a long time, pull vergence distance toward the screen distance, not toward “within arm’s reach.” What the hand must touch and what the eye must stare at need not be the same object.
• Do not try to “fix” near blur by raising texture resolution — if the blur is accommodation failing to match, pixels will not help.
• How to check: mark the screen-distance plane in the engine, put the main fixation object on that plane and then near, one minute each. If the near minute takes effort to keep clear, the conflict is already large enough for an operator to feel. Record the gap in dioptres, not as “the near one looks more 3D.”

Related

• Same group: N1.04.2 The conflict causes eye strain and discomfort · N1.04.3 Interface distance must fall in a comfort zone
• Nearby: N1.10 Binocular Disparity and Vergence–Accommodation Conflict · N3.02 Interface Distance
• Search terms: vergence-accommodation conflict · focal distance · stereoscopic HMD