Stereo depth has a comfort zone, and it narrows up close

What the 2011 study actually measured

On 21 July 2011, the Journal of Vision published The zone of comfort: Predicting visual discomfort with stereo displays by Shibata, Kim, Hoffman and Banks, reporting how far in front of or behind the physical screen stereoscopic content can be placed before viewers judge it uncomfortable. Rather than testing a single distance, the study modelled comfort as a function of viewing distance itself, since a stereo display places content by mismatching the angle each eye must converge on against the fixed distance the eye must focus at, a mismatch the eye tolerates differently depending on how far away the whole scene is.

What creates the discomfort, and why it depends on distance

The mechanism is vergence-accommodation conflict: eyes normally converge and focus together, but a stereo display forces them apart, since focus stays locked to the screen while convergence is pulled toward wherever the stereo pair places an object. An earlier paper by Hoffman, Girshick, Akeley and Banks established this mechanism directly, finding that when a display's vergence and focal distances differed by around 0.67 dioptres across test distances of 31 to 54 centimetres, viewers reported measurably more symptoms. The 2011 study's contribution was showing that the tolerable range of this conflict is not a fixed number of dioptres or centimetres; it changes shape with viewing distance. For a simulated ten-metre cinema screen, the paper reports a comfortable range extending 'from 1.6 m to infinity'; for a forty-centimetre handheld or desktop display, the range narrowed to only 'from 0.28 to 0.44 m'.

What this bounds, and what it does not

This is laboratory data about a specific, measurable form of discomfort, not a universal verdict on 3D content. The paper's regression-based zone boundaries describe averages across its tested viewers, and sensitivity to vergence-accommodation conflict is known to vary between individuals, which the studies do not claim to eliminate. Neither paper opened here tested a head-mounted display directly, so translating these numbers to a headset, where focal distance is typically fixed by the optics regardless of the simulated scene, is an extrapolation the source material does not itself perform.

Questions to bring to a demo

• At the intended viewing distance, how far in front of and behind the screen plane does the content place objects?
• Was the comfort range calibrated for this specific screen size and seating distance, or borrowed from a different format?
• For a headset, is focal distance fixed regardless of simulated depth, and by how much does that diverge from vergence?

Depth budgets for stereo content are not a single rule of thumb; the comfortable range measured in 2011 stretches or shrinks with viewing distance, and the two studies together explain why: it is the mismatch between where the eyes converge and where they must focus that the eye ultimately has to tolerate.