iCouldPlay repeating actions: define cycles with the same reference state

Four appearances of a starting pose do not necessarily show four complete cycles. Play a self-made repetition clip in iCouldPlay, define a recognisable boundary and count intervals between boundaries. No A–B loop or automatic counting feature is assumed; times and counts are external notes.

Prepare a directionally clear action

Film a lightweight card with red and white faces: turn it white, move it outside a drawn box, return it and restore red, repeating several times. Use paper rather than exercise or hazardous machinery. Keep box and card in frame and preserve the original sequence. Exact five-second periods are unnecessary; clear reference states matter first.

Define a boundary by position, face and phase

Define the start as red face up, fully inside the box, before the next flip. “Inside the box” alone also fits a white-face return state. End at the same resting red state, not at a white-face-inside state mistakenly treated as the next cycle’s start. Observe one full sequence to confirm recurrence; if ambiguous, add a small corner mark instead of changing the action path to fit an existing note.

List reference-state occurrences by playback

Play the original from before the action and note approximate file time and an index for each recurrence. An initial resting red state can be boundary one, noting whether action follows. A clip ending halfway back contains an incomplete return, not an automatic endpoint. Watching the file twice does not double the actions contained in it.

Check three intervals from four boundaries

For a paper example, boundaries at 2, 7, 12 and 17 seconds define three intervals: 2–7, 7–12 and 12–17, nominally five seconds each. Four points are not four complete intervals. Write endpoint pairs; adjacent intervals can share a boundary without duplicating an action. Actual footage need not match these values, and whole-second display is not precision timing.

Decide how resting pauses belong to a cycle

Specify whether the boundary is return-to-rest or the next flip onset when resting pauses occur. Either convention changes interval allocation, so use one throughout rather than subtracting pauses selectively. For sequence study, describe phases without computing speed. Learning pauses may be intentional; unequal durations alone do not prove an action error or playback stall.

Distinguish edited replay from new action

Identical background and hand details may suggest a replay but similarity alone does not establish it. Consult replay labels, course notes or your recording log and mark insufficient evidence. Three edited presentations need not be three independent attempts and cannot establish consistency. Ordinary playback cannot recover omitted recording intervals or original event times.

Revisit both sides of each boundary

Play from before each candidate time to confirm the return has ended and the next flip has not begun. A still alone may not identify phase. Revisit the most ambiguous boundary and note face, box position and whether the hand still pushes. Start earlier and play continuously when seeking is coarse; no frame-step button is required. If no resting phase exists, use a clear directional crossing and apply the revised definition throughout.

Handle occlusion and incomplete ends

Mark palm-obscured card faces as uncertain rather than inferring a return from rhythm. Treat a white-face opening as partial action until the first full boundary; handle the end likewise. Report complete cycles and partial opening/ending segments separately. A clearer re-recording is a new file; do not backfill its observations as if visible in the original.

Save interval pairs and a replay route

Each cycle should have endpoints, a state definition and completeness status; recompute the four-point/three-interval example. Record version and conditions, reopen before a confirmed boundary and find it again. Restore viewing position and preserve original counts. This is a manual index of visible content, not evidence of real-world frequency, training effects, independent takes or frame-level device testing.

Related tutorials and sources

Visual information, Course time index.

Checked on November 7, 2026 against the official iCouldPlay App Store listing and Apple public lookup: version 1.1.1, requiring iOS / iPadOS 15.0 or later. The public 1.1.1 release notes list Kodi, OpenList / AList and S3-compatible storage, with improvements to source switching and captions after seeking. The tasks use this public scope and existing playback capabilities. Check controls in your installed version. Self-made samples, manual records and external file preparation are workflows, not newly promised automatic features. No physical-device or private-server test is claimed.