Handling the real environment and overlays at once costs extra attention; they compete for bandwidth

What it is

Watching the kerb for a step and reading the next line of navigation are not “seeing together.” They compete for one pool of attention. Dual-task attention makes environment processing and overlay reading two concurrent loads; when bandwidth runs out, the slower one — usually the surprise in the environment — is what drops.

The fight is for attention, not only for pixels covering each other. Translucency, small type, and edge placement still cost a cognitive cut.

Why it happens

Central attention can stably push few reportable tasks at once. Walking, avoiding, reading social signals already occupy one; an overlay that must be read, compared, decided, becomes a second. Switching has a time cost and a miss cost: detection-response-task (DRT) style paradigms see misses and slowing on an oddball signal once the secondary load rises. The more the overlay is something to process (sentences, numbers, tappable controls), the harder it grabs; the more it is ignorable texture, the lighter the grab — and the easier it is truly ignored.

People also set the priority wrong: the overlay is new and moving, the environment is old and background. Attention is sucked onto the layer; the step, the approaching person, the fork tine pass in the “background.” It is not that the layer is too large. It has been treated as the primary task.

Studying it

A dual task: primary is walking a real path or operating a real object, secondary is reading/tapping the overlay — or the reverse. Add a DRT or an oddball environment event (a sudden obstacle, a flash). NASA-TLX can support load; it cannot replace misses.

Load manipulations: overlay information rate, whether a decision is required, walking speed. Observations: missed environment events, overlay reading errors, gait change, rated load, gaze-switch rate between layer and world.

Sitting and reading an overlay cannot see this fight. There has to be a real environment task that must be handled.

Where it stops holding

Stationary, unchanging environment, the task is reading the overlay (a manual at an empty desk): the environment stream is almost idle and competition is weak. Walking automated on flat ground can free bandwidth; stairs, crowds, traffic break that automation at once. Auditory overlays and visual overlays do not share a peripheral channel, but they still share central attention; “make it voice” is not free. Attention deficit, fatigue, and age make dual-task collapse earlier; bandwidth from young lab subjects will not transfer. After density has already been cut against the environment and show/hide already follows the task, competition can still happen on “this one necessary current item” — reduction and show/hide lighten the load, they do not cancel the dual-task structure.

Applying it

• Design overlays as glanceable state, not as a paragraph that must be finished before acting. When a decision is required, pause or slow the demand to move through the real world.
• Keep the environment-event channel easier to capture than the overlay: people and steps get larger, more peripheral cues. Do not make them fight a menu for the same fovea.
• While walking, carrying, or talking to a real person, drop overlay decision density by default; leave a heading or one word for the next step.
• How to check: with overlays on, let an unannounced real obstacle enter the path (a cushion, a carton). If people hit it, or yield later than a no-overlay baseline, attention has already been taken by the layer. Asking only “tired?” is not enough.

Related

• Same group：N5.12.1 Overlays should show and hide with task relevance, not stay resident · N5.12.2 Information bound to a specific object is easier to associate correctly than information floating in view · N5.12.4 When the environment is already information-rich, overlays must reduce, not add · N5.12.5 Density control aims at task completion, not at showing as much available information as possible
• Nearby：N5.07 Overlay Density · N5.13 Safety Risks from Occluding Reality
• Search terms：dual-task attention · detection response task · inattentional blindness