[ PROOF SURFACE / ENCODING ]

ZPE-Video 3D Surface

This surface may show video-lane mechanics, but the dossier remains the source authority.

imported augmented route

ZPE-Video imported 3D console

imported-static ZPE-Video imported 3D console

visual panel

zpe-video-receipt-machine

Physical video-frame receipt machine with film rails and black calibration hardware.
Video receipt fixture generated for the video proof surface. ZPE-Video

interactive module

Frame Mechanics Stack

Video-lane mechanics are presented as review reference, with dossier authority left intact.

source 68 mechanic
mode
frame
frame
F07
weight
068

related surfaces

Imported showcase ZPE-Video imported 3D console Product dossier ZPE-Video Source PRIVATE License SAL-7.0 Portfolio encoding

source metadata

Branch And Path

branch
origin/codex/proof-console-zpe-video-3d-2026-04-29
ref
origin/codex/proof-console-zpe-video-3d-2026-04-29
path
product-pages-uplift/lanes/ZPE-Video-3D.html
authority
docs/WEDGE.md
commit
07187909
route
/proof/zpe-video-3d/

what it shows

Shows the ZPE-Video 3D review reference for video-lane mechanics.

  • Bit-exact cross-writer output under default settings — pinned by tests/test_receipt.py::test_cross_writer_independent_implementation_matches, which hand-rolls an independent from-spec encoder and asserts byte-identical output.
  • Per-frame CRC32 integrity on every decode; any corruption raises ReceiptCorrupted — pinned by tests/test_receipt.py::test_crc_mismatch_raises.
  • Deterministic hash: receipt_hash(blob) = SHA-256 of the encoded bytes, stable across conforming writers — pinned by tests/test_receipt.py::test_receipt_hash_is_stable.
  • Round-trip fidelity: decode_receipt(encode_receipt(r)) recovers the same frame content up to canonical sort order — pinned by tests/test_receipt.py::test_round_trip_small.

what it does not claim

This surface may show video-lane mechanics, but the dossier remains the source authority.

  • We do not claim universal video-codec superiority. Falsified at Phase 08 against AV1, VVC, and learned baselines on both detector families; record preserved.
  • We do not claim the archive-query-on-box-state surface is a ZPE-specific wedge. Falsified at Phase 09.4.1.1.2 against a 60-line raw-struct+zlib baseline; raw-struct+zlib strictly dominates ZPE on both storage and query latency for that workload.
  • We do not claim the ROI/foveated-sidecar produces a packet-specific lift. Killed at Phase 09.4.1.1.2.1-A when a spatially-uninformed mean-importance control lane produced the identical +7.93% matched-bitrate mAP@50 lift.
  • We do not claim Parquet cannot be configured to match. The wedge is specifically about default settings; Parquet tuned with enforced encoding and sorting can close the gap.

proof anchors

Surface Anchors

  1. lane.what_we_prove
  2. lane.verification
  3. lane.proof_anchors

dossier proof files

Authority Files

  1. docs/WEDGE.md :: VERIFIED
  2. docs/WIRE_FORMAT.md :: VERIFIED
  3. docs/TRANSPARENCY_JOURNEY.md :: VERIFIED
  4. docs/QUICKSTART.md :: VERIFIED
  5. docs/ARCHITECTURE.md :: VERIFIED
  6. docs/STATUS.md :: VERIFIED

dossier metrics

Declared Values

CROSS_WRITER_HASH_STABLE
TRUE
vs default-Parquet FALSE
STORAGE_VS_DEFAULT_PARQUET
0.18×
vs default-Parquet 1.00×
CACHE_READ_VS_DEFAULT_PARQUET
0.067×
vs default-Parquet 1.00×
RECEIPT_MANIFEST_BINDING
TRUE
C2PA-style hash reference; receipt bytes unchanged