Technique / SelectionContext

Keep the selected scope through a reframe

Connect an edit request to the object it addresses.

Run the original

6 s · 30 fps · silent

Original neutral assets · 6 seconds · phase values are chosen defaults, not source timings.

npm run render -- SelectionContext out/SelectionContext.mp4 --codec=h264 --muted

Download the complete pinned project · Inputs, phase meanings & full source code

Official source / inspected interval

Work with docs, sheets, and slides in ChatGPT

Watch the official OpenAI upload ↗ · Published 2026-07-23T07:17:31-07:00

Inspected 23–27.5 s. Consecutive subset: 23–24 s, 24 frames. Dates come from the official watch-page publication field.

Table already selected before the close view at source 23.023000 seconds — 23.023000 s · Table already selected before the close view

Attached action in close framing at source 23.398375 seconds — 23.398375 s · Attached action in close framing

Complete request visible in sample at source 26.276250 seconds — 26.276250 s · Complete request visible in sample

Wider context returns at source 27.110417 seconds — 27.110417 s · Wider context returns

Source: OpenAI. Bounded excerpt and annotated stills for the analysis below; not included in the project download.

What changes across the edit

Shot boundary: A move into the selected table is followed later by a pullback; this is not a proven instantaneous crop cut.
Framing: Table close-up becomes readable while the presenter inset stays screen-anchored. The wider conversation returns near the end.
Layer ownership: Selection under table text, attached action/input, document and conversation; separate presenter inset.
Motion: The contextual action becomes a request field, text grows, then the camera pulls back.
Typography: Table glyphs remain above selection fill; entered request stays attached to selected scope.
Action order: Selected table already present → close view → request → wider context.
Pacing & readable holds: Request is complete in the 26.276 s sample; pullback appears at 26.693 s. Selection onset predates the interval.

Recreation versus observation

Matches

Selection, attached request and restored context remain related.

Differs

Original table rows and simplified request. The source already has selection; ours introduces it. No presenter inset is included in the presentation study.

Measured

Consecutive 23–24 s frames; full inspected interval 23–27.5 s.

Inferred

The selection identifies scope, but this interval does not prove the requested modification completed.

Chosen defaults

All Remotion phase frames, cubic Bézier (.16,1,.3,1), geometry, colors, font choice and reading holds are chosen defaults. None is a recovered OpenAI specification.

{
  "title": "Make room for the idea.",
  "prompt": "Arrange a small exhibition.",
  "result": "Three studies, ready to compare.",
  "label": "EXHIBITION / 03",
  "accent": "#40685b",
  "action": 30,
  "change": 60,
  "settle": 120,
  "hold": 60,
  "focusX": 0.66,
  "focusY": 0.64,
  "zoom": 1.55
}

Increase hold to extend only the final quiet state. Change the phase frames to change choreography; preserve their ordering. Content limits and the exact meaning of each phase are in the handoff.

Exact original geometry, layer ownership and asset contracts · Schedule against an actual recording