Space got larger; the readable angular range did not

What it is

A room offers four walls and a ceiling, as if more could be posted than on one monitor. What a person can actually read at once is still a small solid angle around the fovea. Angular bandwidth did not grow because the world did. What grew is the area that must be turned into that small angle; not the simultaneous readable capacity. Ten panels around a room are not more desktop windows. They are the same narrow band wired to ten stations that must be foveated in turn.

Why it happens

Reading and fine discrimination spend the high-resolution central retina. That patch’s visual angle is the same order in a headset as at a monitor, and often narrower and mushier at the edges. Spatial layout changes the bearing of content relative to that patch; it does not change the patch. “The space is large” is then misread as “more can be shown at once”: shown, yes; read, no. To read a second panel the small angle must be aimed at it, and the previous panel leaves the band. Desktop windows at least share a screen that needs no head turn; spatial panels do not share by default. Information architecture budgeted in square metres treats turns as free; each foveation spends time and a vestibular bill. Angular bandwidth is instantaneous capacity; room area is inventory. Inventory can be huge; the counter still serves one plate at a time.

Studying it

Hold readable visual angle fixed and compare one large board with the same information split into several boards around a room, on a task that cannot finish until all of it is read.

Independent variables: panel count and layout (one / same-depth row / around the room), angular size of each, walking allowed or not. Dependent variables: total time, integrated head motion, missed panels, a rating of “more information or just more scatter”.

Useful-field-of-view and visual search transfer directly: space only changes the geometry of the search set from a screen to a room. “How many boards fit in the room” is not a capacity metric.

Where it stops holding

Content that need only be detected as present can genuinely multiply with the room — bearing cues spend the periphery, not angular bandwidth. A trained control-room scan can time-share several boards into the fovea; instantaneous capacity is unchanged, multiplexing is skilled. Users of ultrawide desktops or many monitors already scan with the head, so the drop into space is smaller, but headset edge optics are worse and desktop saccade radius must not be reused. Low vision compresses angular bandwidth further; a larger room does not rescue it.

Applying it

• Budget information by “what can be finished in one foveation”, not by wall area.
• Put content that must be compared inside one foveation’s angle. Do not split it to opposite walls and reassemble from memory.
• Use extra room positions for inventory and presence, not for spreading the layer currently being read.
• How to check: on a task that must read everything, count foveations. If the count tracks the number of panels, capacity did not grow with space — it was amortised across many aims.

Related

• Same group: N3.14.2 Stacking panels in depth does not increase effective capacity · N3.14.3 Spatial layout has no automatic typesetting; overlap is constrained only by rules · N3.14.4 The extra placement freedom is also a tidying cost
• Nearby: N3.09 Interface Distance and Comfort Viewing Distance · N3.07 Content at the Edge of the Field of View
• Search terms: angular bandwidth · useful field of view · information capacity