Very few depth layers remain usable

What it is

Space looks as if panels could stack forever along z; the eyes and attention can barely keep two or three layers alive. A depth layer budget says that interface content which must be read and operated at once can be assigned only a few depth slots. Adding another layer is not adding a page — it makes the existing layers fight for vergence and attention. This is not the same claim as “this one panel sits too near or too far”. Even when every layer is at an acceptable distance, count itself can crowd the system out.

Why it happens

Each layer is a vergence distance. Changing layer means reconverging and re-accommodating (on a headset whose focal plane is often locked at one optical depth, accommodation cannot follow). Attention switches more slowly in depth than a saccade on one plane, so misses and wrong-layer selections rise. Once layers overlap along the line of sight, order, which layer was hit, and which type is clearer all become extra discriminations. Desktop tabs and stacked windows share one focal plane; in space each layer is a fresh ocular settlement. The right response to “add another panel” is often to place it beside at the same depth, or to time-sequence it — not to open another z slot. HUD work that compresses information into two layers (far scene + near controls) is not taste; the budget is already spent.

Studying it

Hold information load constant and compare layouts with 1 / 2 / 3 / 4 depth slots, on a task that must read and click across layers.

Independent variables: number of concurrently active depth layers, spacing between layers, whether layers overlap along the line of sight. Dependent variables: time to switch layer, wrong-layer clicks, reading accuracy, rated ocular fatigue (sampled with a short in-session scale, not only at the end).

Vergence or an eye tracker’s z estimate shows whether people actually change layer or only cope with the rear layer in periphery. Four layers that look present while only one is read do not count as budget used.

Where it stops holding

Layers used only as presence (a glowing anchor in the far scene) and never read barely spend the budget and can be numerous. Games that stack depth for space, with no reading task, are a different use. Monocular or non-stereo users lose part of the vergence-switching cost, but attention switching remains, so the gain in layer count is limited. On optical see-through the real world already occupies one “environment depth”, so the virtual budget is tighter — a slot is spent before the UI starts.

Applying it

• Keep simultaneously read and operated UI in at most two layers. A third layer holds unread presence cues only.
• Prefer placing new content beside an existing layer, or sequencing it in time. Do not default to “nudge it back and open another plate”.
• When layers overlap, land operations on one explicit layer (front as modal, or hide the rear while operating). Do not make people guess which layer they hit.
• How to check: list the depth slots actually read during the task. Regular traffic across more than two or three, or wrong-layer hits, means the budget is blown — merge or switch, rather than retuning each layer’s metres.

Related

• Same group: N3.09.1 Angular size is the design variable, not physical size · N3.09.3 Letting people set distance is more reliable than one universal preset · N3.09.4 Comfort viewing distance narrows toward mid-range as the session lengthens
• Nearby: N3.02 Interface Distance · N1.04 Vergence–Accommodation Conflict · N3.14 Information Density and Stacking in Space
• Search terms: depth layer budget · focal switching · multi-plane HUD