Watch the frozen pair
29 frames · 30 fps · square pixels · silent
Both players show the entire frame; the small dot is intentional. The responsive page scales the view, not the downloaded files. Posters are the actual full first frame and the full-frame editable SVG. Browser seeking and shared controls are inspection aids, not frame-lock or native-resolution quality evidence. Each native player works without JavaScript.
Official source film ↗ · Source interval [445/30, 474/30) s = [14.833333…, 15.8) s. Frame 474, the first separate leftward movement, is excluded.
OBSERVED OPERATION
A centered color run.
Gray, yellow, blue, orange, lavender, pink, then green. The radius grows early and settles; there is no repeating pulse, positional travel or inferred success event in this excerpt.
The chosen interval begins at the cut to the isolated dot. Desktop and person shots are not needed to reconstruct it.
| State | Local | Source |
|---|---|---|
| Gray | 0–3 | 445–448 |
| Yellow | 4–7 | 449–452 |
| Blue | 8–11 | 453–456 |
| Orange | 12–15 | 457–460 |
| Lavender | 16–19 | 461–464 |
| Pink | 20–23 | 465–468 |
| Green | 24–28 | 469–473 |
Use the editable build
Portable setup instructions · Per-frame Resolve controls · First full-frame SVG
Original excerpt ↓ · Remake ↓ · Technical checks · File identities
Frozen pre-review snapshots: neither ZIP, movie nor source-code file has been revised for publication. The archives' embedded “pending” / “not approved” review fields predate the September 17, 20:23:27 UTC verdict. The dated review records the subsequent accepted preference; it does not rewrite that history. The Resolve recipe remains untested in native Resolve.
REPORTED CRITIC SCOPE
A narrow preference for the edge and fill.
Hypatia's reported reason was the remake's cleaner circular outline and flatter fills. Size, center, color sequence and timing were essentially tied. The largest reported gap was the original's flattened or notched lower-right contour and tinted rim during pink, local frames 20–23. Their frame-start times are 0.666667–0.766667 s; the pink state ends exclusively at 0.8 s.
The critic inspected all 29 frames in each clip with full-frame context and unscaled native-detail crops, then rechecked frames 19–24 together. The reference was the actual official distributed 1920 × 1080 rendition, not an upscaled 480p proxy and not an uncompressed production master.
With T3 unavailable, the critic used local headless Chromium at 1×. Both clips completed with 29 reported total frames, zero dropped frames and zero corrupted frames. Distinct callback counts were 27 for A and 26 for B; callback gaps are not confirmed playback drops. Visual judgment came from decoded progression, not continuous human observation.
This is the supplied independent agent verdict, not a new editorial comparison by the publication builder, official OpenAI approval or a universal playback guarantee. Earlier authoring tests separately recorded two candidate drops on an uncaptured first attempt and zero on a buffered replay; the readiness change did not establish a cause. The earlier equal-480p preference supplies no additional HD credit.
Source, clock and rights
The official AV1 source was acquired September 17, 2026 at 19:46:35 UTC, format 399, 1920 × 1080 at 30/1 fps. The original excerpt is a near-lossless H.264 re-encode of that rendition, not an AV1-identical stream. Both paired downloads retain their frozen bytes.
- Full-source SHA-256
7d8042783446d894f98e1a5c977143cca03664ad304348585ae579893d9a7d9e- Source PTS, time base 1/15360
- 227840 to 242176 inclusive; 242688 exclusive out; 512 ticks per frame.
- Excerpt / remake PTS, same time base
- 0 to 14336 inclusive; 14848 exclusive out. Duration 29/30 s; no audio.
The original remains comparative research material. Copyright and mark rights remain with their owners; access to the film or this study does not grant permission to reuse footage, branding or demo content. The remake contains no source screenshot, traced contour, source wording, logo, font, voice or music. Its measured values are not official design tokens. Dependency licenses still apply.
Frozen source provenance · Research and reuse boundary · Publication status and scope