# Published R02 comparison record

This publication copy carries the frozen pre-publication findings. It omits the anonymous crosswalk and private binding records; no finding is rejudged.

# Results after authorized unblinding

Date: September 15, 2026. This is a mapping of already sealed static findings,
not a new review. The parent verified the source and label seals before
authorizing unblinding. Criteria, findings, confidence, and evidence limits
remain unchanged. These sanitized results remain local; nothing was published.

## Mapped findings

| Variant | Single biggest remaining evidenced visual gap | Review disposition |
|---|---|---|
| Current R02 — default | none established | Moving review outstanding; no quality approval. |
| Previous R01 — default | C3: two held preview illustrations omit the source's between-cut picture-state progression. | Frozen static gap retained; moving review outstanding. |
| Current R02 — adaptation | none established | Moving review outstanding; no quality approval. |
| Previous R01 — adaptation | C3: two held preview illustrations omit the source's between-cut picture-state progression. | Frozen static gap retained; moving review outstanding. |
| Research source reference | none established | Reference-only bytes; no production or quality approval. |
| Research source reference | none established | Reference-only bytes; no production or quality approval. |

“None established” means no material contradiction was established within the
frozen static coverage. It means **moving-review-outstanding**, not a blind win,
overall rank, quality judgment, verified repair, or equivalence at playback
speed. No new gap is invented after mapping. Default and adaptation remain
separate variants; there is no overall winner.

For the two Previous R01 labels, the frozen C3 finding concerns picture-state
progression within the wide and close segments. Both have the late preview cut,
working trim/readout changes, and a nonempty preview. The finding is not a
missing cut, blank picture, or whole-interface freeze. Intentional still
illustration was already considered as an alternative explanation. No new
causal conclusion about the changes between R01 and R02 is made.

## Actual movie and source hashes

SHA-256 values identify the inspected movie bytes, not an overall quality level.

| Variant or reference | Actual movie SHA-256 |
|---|---|
| Current R02 — default | `2b41e87db6670eaaad1c3b9316f1a16147ced37063e5baa1b5b1578eb9073aab` |
| Previous R01 — default | `934ae575b22340c454a8c9af9cf38a69655e4fe56482c3edd76eb23c5e2adbf0` |
| Current R02 — adaptation | `2b4d58161aa8304b96c8610d72d1b14ff9e16a4c1dbacdb10544c6c8de140b4b` |
| Previous R01 — adaptation | `d2bfab28489115aff32b3d6afe9782a18cce37ef919b8df0af97d5a2bc026a3d` |
| Source reference | `61061b6a1a6a73bd4f22e4ee03118f4683a07bed02a210790f491765b0937263` |
| Source reference | `61061b6a1a6a73bd4f22e4ee03118f4683a07bed02a210790f491765b0937263` |

The frozen source excerpt SHA-256 is
`61061b6a1a6a73bd4f22e4ee03118f4683a07bed02a210790f491765b0937263`.
The two source-reference movies equal those bytes. Their identity was already
disclosed during the anonymous stage because packet hash verification exposed
it. Their images were nevertheless inspected independently by supplied path.
Research-source bytes are not approved as original presentation assets.

The frozen source seal SHA-256 is
`ecc34b9b063493b528edc356519d1cbccbdb25b640153d60e0e358d17e748964`.
The frozen label seal SHA-256 is
`6632627397572d20bb210877b8186f791e5b25a244abbac951ce541234f6a081`.
The anonymous packet SHA-256 is
`ae781d657b04886e17be4368bbc78ba08ef68b4601a91066fc335e89dc822649`.
Authorization to map these labels is not authorization to publish or approve
the movies. Private binding material is excluded from these results.

## Timing basis

The frozen actual-file checks establish, for each of the six movies: one silent
H.264 video stream, 1920 × 1080, 180 decoded frames, frame rate `60/1`, time base
`1/15360`. Local frame `n` has PTS `256*n` and duration 256 ticks. The first PTS
is 0, the last is 45824, and the exclusive end is 46080. Thus frame time is
exactly `n/60` seconds and the local interval is `[0,3)` seconds.

The frozen phase observations include the boundary/readout plateau `[104,131)`
(27 frames, `9/20 s`), the first close preview at frame 150 / PTS 38400
(`5/2 s`), and the final boundary/readout state `[171,180)` (nine frames,
`3/20 s`). UI marker strings do not replace that actual movie clock. P/B markers
in the four variants are explicitly ordinal; the source UI's dot-separated
suffix convention was not independently established.

These timing statements are carried forward from the sealed review. This
unblinding stage performed no new decoding, playback, or movie inspection.

## Coverage basis and secondary limitations

All coverage below was completed before the label findings were sealed:

- Source-first basis: all 180 chronological thumbnails, nine individual
  native-raster files, 180 native text crops, 56 native edge/cursor crops, and
  six native preview crops.
- Anonymous basis, each label: all 180 chronological thumbnails; native-raster
  files at 0, 104, 150, 179; 12 native preview crops; 27 native text crops.
- Anonymous total: 1080 chronological thumbnails, 24 native-raster files,
  72 native preview crops, and 162 native text crops, in 78 opened images.
- Preview-crop frames per label: 0, 60, 104, 123, 124, 129, 149, 150, 151, 171,
  172, 179. Text-crop frames: 0, 58, 59, 60, 64, 70, 75, 78, 81, 86, 90, 94,
  99, 104, 130, 131, 134, 137, 140, 145, 150, 153, 157, 160, 165, 171, 179.

This was chronological thumbnail review plus selected native frames/regions,
not moving playback or native-pixel-complete review. Viewer scaling may occur.
Decoded-frame hashes, clock probes, and pixel differences do not establish
perceived motion, fluency, easing, or frame pacing under playback.

The frozen C3 caveat remains: distinct picture states do not establish
autonomous playback, trim causality, or visible change on every held-endpoint
frame. Small pixel differences may be compression differences; the source
samples around 104..129 do not independently establish geometric progression.
The retained gaps rely on broader visible between-cut state comparisons.

Actual application behavior, clicks, pointer hotspots, original SVG/code
provenance, and full-film quality remain unverified. These are limits of the
evidence, not newly assigned defects. General learned familiarity could not be
audited. Incidental filename/storage-label exposure and source-hash identity
were disclosed before unblinding; no earlier critic's conclusions were used.

**Moving-quality credit: 0. Native-Resolve credit: 0. Source-quality approval:
0. Full-film credit: 0.** No quality approval, automatic approval, publication,
new rendering, implementation change, or new visual review occurs here.