Check occluded demonstrations in iCouldPlay: separate visible states from hidden-action inference

When a hand covers the work area, a spoken claim of completion does not mean viewers saw every action. Watch a lawful tabletop demonstration in iCouldPlay, separating visible states, occlusion intervals and spoken explanation. This concerns blocked visibility within a continuing shot, not jump cuts. No seeing through hands, automatic step recognition or completion verification is assumed. The original paper sample is an observation format, not equipment-operation instruction.

Define a visible paper-sample task

Use only a blue paper square and a white circle drawn on a sheet, with fixed camera, square initially left and finally right of the circle. Plan a hand covering circle and square region mid-sequence while other desk references remain visible. Add no blades, circuits or real equipment buttons. Make lawful footage or a three-panel plan, labelling unfilmed work untested. Record filename, version, duration, view and circle identity; matching colour alone does not prove one object across shots. Check evidence scope, not honesty or presumed deliberate concealment.

Record states at three stages without filling in actions

The paper timeline places the blue square left at0s, hand covering the target at1–2s and square visible right when the hand is away at3s. Record endpoint positions and “region occluded; square position unseen” during1–2s, not “crossing the centre.” The plan specifies no complete intervening states at0–1 or2–3s either. Add actual continuous visibility segment by segment if your footage provides it. Times are original suggestions, not frame measurements derived from a progress bar. Without a made sample, endpoint states remain planned, not observed in iCouldPlay.

Find occlusion onset and end in normal playback

Play from a fully visible target region and note hand entry, covered objects and renewed visibility. Watch coverage, not only the square: losing the circle also affects reference continuity. Revisit boundaries with available pause controls; if timing is coarse, use an approximate second plus a content anchor, not invented milliseconds. Distinguish a visible corner from fully hidden areas under partial occlusion; its movement does not expose all edges. Log freezing, defocus or captions as different visibility problems rather than calling everything hand occlusion or blaming an engine first.

Separate endpoint change from a fully seen process

For this paper sample, say the square changes from left to right across visible endpoint states, not that it necessarily travelled straight through centre, stayed on the table or moved only once. Hidden routes and pauses remain unknown. Conversely, unseen movement does not prove no movement: unobserved is not absent. Record a spoken “push left to right” as narration with its time and separately state any visual support. The aim is not rejecting narration, but avoiding expansion into details supposedly witnessed. Keep state and process conclusions distinct.

Use additional views only with established correspondence

If desired, refilm the paper plan with the hand away from the target and save a clear version. That is a new demonstration, not automatic truth about the earlier event. A simultaneous side view needs confirmation of same event, timing and sample identity before adding its visible evidence. A clear action from another take cannot establish what necessarily happened under the earlier hand. Without supplemental footage, request explanation or demonstration rather than searching for a nonexistent see-through control. Two views can both be obstructed; preserve each scope instead of combining blur into certainty.

Describe both visibility and its limits

Write a neutral note: square left of circle; hand covers circle and square; square right after hand withdraws. Do not substitute “hand passes square through centre” without further evidence. W3C guidance on necessary visual information and objective description helps note quality, not a claim that iCouldPlay generates audio description or this note has accessibility certification. Label source narration, your description and inference separately. Clear sound does not ensure visible action; a caption saying completed is textual information. Keep quotations brief and lawful; paper-example wording is not a source-film quotation.

Check display obstruction without rewriting the event

Temporarily hide optional captions or menus through available controls and revisit the same anchor, then restore settings. A hand or text burned into footage will remain; magnification cannot recover details never recorded behind it. A display change may crop hand-side references, so log modes and change one setting at a time. After buffering, replay from before occlusion rather than claiming skipped seconds were observed. Screenshots record display state and may later be scaled/compressed; interpolated blur must not become a certain gesture or button state.

Reopen and save a bounded completion record

Reopen the same version and play from an unobstructed start through the final state, confirming columns, covered objects and anchors. Distinguish visible completion state from fully verified process. Leave the latter unknown when evidence is insufficient rather than forcing yes/no. Preserve source and first notes, appending reasons for corrections instead of renaming old observations to hide omissions. Another reader should know what you saw, heard and could not see, not receive an authenticity verdict. A new tutorial needs new target/reference definitions; this paper example proves neither universal reliability nor unreliability of occluded steps.

Related tutorials and sources

iCouldPlay tutorial jump cuts: separate visible states from omitted operations; iCouldPlay recorded pointers: separate targeting, click feedback and resulting state.

Sources checked on 11 November 2026: the official iCouldPlay App Store listing and Apple Lookup show version 1.1.1, requiring iOS / iPadOS 15.0 or later. This workflow uses ordinary playback and manual records through controls actually present. External editing and sample creation are not newly promised iCouldPlay features. Confirm capabilities against installed version, engine and source. Examples and plans are not step-by-step device tests or guarantees about every device.

Method reference: W3C WAI’s description guidance covers visual information necessary for understanding and recommends objective description without interpretation. The blue-square/white-circle sequence is an original observation plan, not demonstration-authenticity detection or an accessibility compliance assessment.