AUDIOVISUAL THREE-PASS REVIEW — TWO FICTIONAL WORKED EXAMPLES Status: local publication companion; not deployed, uploaded, queued, posted, promoted, or publicly verified Author disclosure: Original work by Alfred Purpose: Exercise the audiovisual three-pass review method with synthetic failure cases. These records are method tests, not reports about real videos, accounts, people, customers, destinations, or publication. EVIDENCE BOUNDARY Both examples use invented candidate identities and complete synthetic SHA-256 values. No real media, copied interface, third-party asset, account surface, personal information, upload, audience result, customer activity, payment, or revenue result is represented. The examples test whether picture-only, audio-only, combined, caption, and correction evidence remain separate. Allowed lane results: PASS / FAIL / BLOCKED / NOT TESTED / SUPERSEDED Allowed overall decisions: HOLD / REVISE / LOCAL PLAYBACK REVIEWED / SUPERSEDED A clean lane cannot repair a failed lane. A combined pass cannot be copied into an isolated lane. A changed candidate cannot inherit playback evidence attached to earlier bytes. ====================================================================== EXAMPLE 1 — CLEAN ISOLATED LANES, CONTRADICTORY COMBINED CLAIM ====================================================================== EXAMPLE STATUS Scenario type: explicitly fictional method test Candidate state: synthetic local master Destination rendition: NOT TESTED Upload state: no upload represented Public state: not public Expected terminal decision: REVISE A. REVIEW ENVELOPE Review ID: FICTIONAL-AV3P-EXAMPLE-01 Review revision: 1 Reviewer role: synthetic method reviewer Reviewer disclosure: Alfred Intended presentation class: vertical explainer Intended aspect ratio: 9:16 Realistic review display size: 360 x 640 logical pixels Normal playback speed: 1.0x Audio output path: local stereo headphones Caption presentation: authored captions burned into synthetic presentation Decision this review may support: local playback decision for one exact master Decision this review may not support: rights, accessibility, destination, upload, publication, audience, customer, payment, or revenue status Required passes: [x] Complete picture-only pass at normal speed [x] Complete audio-only pass at normal speed [x] Complete combined pass at normal speed [x] Failure-boundary inspection B. EXACT SYNTHETIC CANDIDATE IDENTITY Stable candidate ID: SYNTH-VID-A17 Master filename: synthetic-periodic-checks-r1.mp4 Master byte size: 4,812,340 synthetic bytes Digest algorithm: SHA-256 Complete master digest: 3c90ff59a137f24288d6a7dd9e129a4456759dd943998274f3a95c6ee49ab071 Duration: 00:18.000 Width x height: 1080 x 1920 Frame rate: 30 fps Container: MP4 Video codec: H.264 Audio codec: AAC Audio channel layout: stereo Approved script identity: SYNTH-SCRIPT-A17-r1 Caption file identity: SYNTH-CAPTION-A17-r1 Render specification identity: SYNTH-RENDER-A17-r1 Rights/provenance record identity: SYNTH-RIGHTS-A17-r1 Claim-review record identity: SYNTH-CLAIM-A17-r1 Disclosure text: Original work by Alfred. Identity checks: [x] Digest was stipulated as the complete identity of the opened synthetic file [x] Probe description names the same synthetic file [x] Script, caption, render, rights, claim, and disclosure identities agree Identity result: PASS Evidence location: this fictional trace, section B Finding: none C. OBSERVATION SETUP Player and version: synthetic local player 1.0 Rendition class: local master Display orientation: portrait Display scaling: fit to 360 x 640 logical pixels Audio output device class: stereo headphones Captions: authored and visible only during combined playback Overlays or masks intentionally shown: none Unexpected controls or overlays: none represented Mute confirmation method: player mute indicator plus zero output at device Audio-only attention-removal method: picture fully covered by an opaque local review window while player transport remained reachable Setup limitation: no destination processing, generated captions, interface controls, recompression, alternate device, or assistive technology is represented Setup result: PASS Evidence location: this fictional trace, section C Finding: none D. COMPLETE PICTURE-ONLY PASS Pass range: 00:00.000–00:18.000 Reached first frame: YES Reached final frame: YES Normal speed maintained: YES Mute independently confirmed: YES Observed sequence: - 00:00.000–00:03.200: title reads “Periodic review checks.” - 00:03.200–00:08.600: diagram shows a clock and a bounded batch. - 00:08.600–00:13.800: qualifier reads “sampled on a declared schedule.” - 00:13.800–00:18.000: takeaway reads “State the coverage boundary.” - AI-assistant disclosure remains visible in the lower authored region. Picture checklist result: - essential text is legible at the declared display size; - no blank, duplicated, stale, or out-of-order frame is represented; - the qualifier and disclosure survive the frame; - no personal information, account data, contact detail, secret, credential, private path, copied interface, or third-party media appears; and - the final frame remains visible for the complete declared hold. Picture-only result: PASS Evidence location: this fictional trace, section D Finding summary: none Claim boundary: picture alone communicates periodic, sampled coverage E. COMPLETE AUDIO-ONLY PASS Pass range: first sound at 00:00.280 through final sound at 00:17.620 Reached first sound: YES Reached final sound: YES Normal speed maintained: YES Picture hidden: YES Audio source: audio encoded in the identified synthetic master Approved synthetic narration heard: “An automated monitor can check every review as it arrives. Group the results on a declared schedule, then state what the monitor covered.” Audio checklist result: - first syllable and final consonant are complete; - the narration matches SYNTH-SCRIPT-A17-r1; - no click, repeated word, unexpected level shift, or clipped edit is represented; - no personal information, account data, contact detail, secret, credential, private path, or third-party audio is heard; and - the sentence is understandable without seeing the picture. Audio-only result: PASS Evidence location: this fictional trace, section E Finding summary: none Claim boundary: audio alone claims every review is checked as it arrives F. COMPLETE COMBINED PASS Pass range: 00:00.000–00:18.000 Reached complete ending: YES Normal speed maintained: YES Caption implementation reviewed: SYNTH-CAPTION-A17-r1 Timestamped finding: Finding ID: FIND-SYNTH-A17-01 Candidate digest: 3c90ff59a137f24288d6a7dd9e129a4456759dd943998274f3a95c6ee49ab071 Pass: COMBINED Range: 00:03.200–00:08.600 Observed: the picture and authored caption say “Periodic review checks” and “sampled on a declared schedule,” while the simultaneous narration says the monitor checks “every review as it arrives.” Picture and audio are individually clear, but their scopes contradict one another. Expected: picture, narration, and captions must express the same approved coverage boundary without dropping the periodic qualifier. Severity: RELEASE_HOLD Affected surfaces: combined meaning, claim review, script, captions, promotion Proposed correction: revise narration and authored captions to state periodic sampled coverage, then create a new master identity and rerun affected passes Combined result: FAIL Evidence location: FIND-SYNTH-A17-01 Finding summary: clean isolated presentation lanes produce a contradictory combined claim G. CAPTION DEPENDENCY Caption source: AUTHORED Caption identity: SYNTH-CAPTION-A17-r1 Destination-generated captions: NOT TESTED Words compared with approved speech: YES Timing compared with exact master: YES Line breaks and wrapping reviewed: YES Clipping and contrast reviewed: YES Collision with authored text reviewed: YES Caption result: FAIL Evidence location: FIND-SYNTH-A17-01 Finding summary: captions preserve the periodic visual qualifier but conflict with the broader spoken claim Accessibility conformance: NOT TESTED H. FAILURE-BOUNDARY INSPECTION | ID | Range | Boundary | Expected | Observed | Result | |----|-----------------------|----------------|--------------------------|----------------------------------|--------| | 01 | 00:00.000 | First frame | Complete title | Complete title | PASS | | 02 | 00:00.280 | First sound | Complete first syllable | Complete first syllable | PASS | | 03 | 00:03.200–00:08.600 | Claim overlap | Same coverage scope | Periodic picture, every in audio | FAIL | | 04 | 00:08.600 | Caption change | Qualifier stays readable | Qualifier stays readable | PASS | | 05 | 00:17.620–00:18.000 | Ending | Sound and frame complete | Sound and frame complete | PASS | Diagnostic method: normal-speed replay, followed by focused replay of the claim overlap Diagnostic limitation: this explains the local contradiction only; it does not represent a destination rendition or audience interpretation Boundary-inspection result: FAIL I. DECISION Picture-only result: PASS Audio-only result: PASS Combined result: FAIL Boundary-inspection result: FAIL Caption result: FAIL Identity result: PASS Open findings: FIND-SYNTH-A17-01 Overall decision: REVISE Why one attentive watch is insufficient: The two isolated lanes are technically clear, but their coverage claims differ. Only the combined comparison exposes the contradiction at the exact overlap. The two passing lanes cannot be averaged into approval. Not established: - rights or license clearance; - accessibility conformance; - platform policy or destination processing; - upload acceptance or public availability; - audience response, inquiry, customer, payment, or revenue result. ====================================================================== EXAMPLE 2 — A VISUAL CORRECTION SUPERSEDES EARLIER PLAYBACK EVIDENCE ====================================================================== EXAMPLE STATUS Scenario type: explicitly fictional method test Initial candidate state: synthetic local master reviewed Revised candidate state: synthetic local master with changed bytes Destination rendition: NOT TESTED Upload state: no upload represented Public state: not public Expected terminal decision: HOLD A. INITIAL CANDIDATE AND REVIEW Review ID: FICTIONAL-AV3P-EXAMPLE-02-r1 Stable candidate ID: SYNTH-VID-B04-r1 Master filename: synthetic-final-hold-r1.mp4 Digest algorithm: SHA-256 Complete master digest: 7dc9e3dc59fa7aac671eb03b0962e021e184f489e15f686d574be42cdf005d91 Duration: 00:14.000 Width x height: 1080 x 1920 Frame rate: 30 fps Container: MP4 Video codec: H.264 Audio codec: AAC Audio channel layout: stereo Approved script identity: SYNTH-SCRIPT-B04-r1 Caption file identity: SYNTH-CAPTION-B04-r1 Render specification identity: SYNTH-RENDER-B04-r1 Rights/provenance record identity: SYNTH-RIGHTS-B04-r1 Claim-review record identity: SYNTH-CLAIM-B04-r1 Disclosure text: Original work by Alfred. Declared local setup: - portrait playback at 360 x 640 logical pixels; - 1.0x speed; - device-confirmed mute for picture-only review; - fully hidden picture for audio-only review; - authored captions visible for combined review; and - complete first and final frames and sounds reachable. Initial review results: - picture-only pass: PASS, 00:00.000–00:14.000; - audio-only pass: PASS, first sound through final sound; - combined pass: PASS, 00:00.000–00:14.000; - authored caption review: PASS; - boundary inspection: PASS; and - decision: LOCAL PLAYBACK REVIEWED for the complete r1 digest only. Narrow initial supported statement: “The identified r1 local master received complete picture-only, audio-only, and combined normal-speed playback passes, plus focused checks at the recorded failure boundaries.” That statement applies only to: 7dc9e3dc59fa7aac671eb03b0962e021e184f489e15f686d574be42cdf005d91 B. CORRECTION REQUEST Finding ID: FIND-SYNTH-B04-02 Source: separate synthetic claim review after playback Observed in r1: final card says “Reviewed locally” for 00:13.200–00:14.000 Expected: final card must say “Local playback reviewed” and remain visible for at least 1.200 seconds so the scope is explicit and readable Severity: RELEASE_HOLD Proposed correction: replace the final-card text and extend the final scene Affected evidence predicted before rebuild: - picture-only playback; - combined playback; - authored caption timing if the timeline changes; - failure-boundary inspection; - render identity; - claim identity; and - exact candidate identity. Audio-only evidence may remain historically informative only if the encoded audio bytes and timing are independently shown unchanged. It cannot silently inherit a pass merely because the narration source filename is unchanged. C. REVISED SYNTHETIC CANDIDATE Stable candidate ID: SYNTH-VID-B04-r2 Master filename: synthetic-final-hold-r2.mp4 Digest algorithm: SHA-256 Complete master digest: 9e738983c81be4754847c891fdcd5b6c4307f29af7d660a980643d34755a944b Duration: 00:14.400 Width x height: 1080 x 1920 Frame rate: 30 fps Container: MP4 Video codec: H.264 Audio codec: AAC Audio channel layout: stereo Approved script identity: SYNTH-SCRIPT-B04-r1 Caption file identity: SYNTH-CAPTION-B04-r2 Render specification identity: SYNTH-RENDER-B04-r2 Rights/provenance record identity: SYNTH-RIGHTS-B04-r1 Claim-review record identity: SYNTH-CLAIM-B04-r2 Disclosure text: Original work by Alfred. Change record: - final-card text changed from “Reviewed locally” to “Local playback reviewed”; - final scene extended by 0.400 seconds; - authored final caption cue extended by 0.400 seconds; - encoded master bytes and complete digest changed; - picture timing, combined timing, caption timing, and final boundary changed. D. INVALIDATION RECORD Previous candidate digest: 7dc9e3dc59fa7aac671eb03b0962e021e184f489e15f686d574be42cdf005d91 New candidate digest: 9e738983c81be4754847c891fdcd5b6c4307f29af7d660a980643d34755a944b Historical r1 evidence states after correction: - r1 picture-only pass: SUPERSEDED for r2; - r1 audio-only pass: SUPERSEDED pending exact r2 audio review; - r1 combined pass: SUPERSEDED for r2; - r1 caption review: SUPERSEDED for r2; - r1 boundary inspection: SUPERSEDED for r2; - r1 identity result: remains historical evidence about r1 only; and - r1 decision: remains historical but cannot name or approve r2. Required r2 reruns: [x] identity and probe agreement [x] complete picture-only playback [x] complete audio-only playback from the encoded r2 master [x] complete combined playback [x] authored caption review [x] final-frame, final-caption, and final-sound boundary inspection E. CONVENIENT BUT INVALID SHORTCUT Shortcut attempt: A reviewer opens a still image of the corrected r2 final frame, confirms that “Local playback reviewed” is legible, and copies every r1 playback result to r2. Why it fails: - a still does not establish the 1.200-second hold at normal speed; - a still does not show whether the timeline introduced a blank or repeated frame; - a still does not show whether the final caption and sound finish coherently; - the encoded r2 audio has not been listened to from first sound to final sound; - the complete r2 digest differs from r1; and - no r1 lane observed r2 bytes. Shortcut result: FAIL Overall r2 decision after shortcut: HOLD F. PARTIAL R2 EVIDENCE Identity result: PASS Evidence: synthetic r2 digest, probe, script, caption, render, rights, claim, and disclosure identities agree Picture-only result: PASS Evidence: complete r2 picture played 00:00.000–00:14.400 at 1.0x with mute independently confirmed; corrected final text remained readable through the complete final frame Audio-only result: NOT TESTED Evidence: blank; the r2 master audio has not been listened to independently Combined result: NOT TESTED Evidence: blank; no complete r2 combined playback has occurred Caption result: NOT TESTED Evidence: source cue timing changed, but the authored r2 presentation has not been reviewed through the complete ending Boundary-inspection result: NOT TESTED Evidence: the corrected still does not replace normal-speed final-boundary review G. DECISION Picture-only result: PASS Audio-only result: NOT TESTED Combined result: NOT TESTED Boundary-inspection result: NOT TESTED Caption result: NOT TESTED Identity result: PASS Superseded evidence: all r1 playback and caption lanes as applied to r2 Open finding: FIND-SYNTH-B04-02 remains open until complete r2 reruns finish Overall decision: HOLD Narrow statement currently supportable: “The identified r2 local master passed a complete picture-only normal-speed playback in the declared setup. Audio-only, combined, caption, and final-boundary review remain not tested.” Statement that is not supportable: “The revised video passed three-pass review.” Why the hold matters: The correction improved a real scope label, but it also created new bytes and a new timeline. Preserving r1 as historical evidence is honest; assigning r1's passes to r2 would claim observations that never occurred. Not established: - complete r2 audiovisual review; - caption or accessibility conformance; - rights or policy clearance; - deterministic reproduction; - destination processing or generated captions; - upload acceptance, publication, or logged-out availability; - audience response, inquiry, customer, payment, or revenue result. ====================================================================== METHOD OUTCOMES ====================================================================== Example 1 expected outcome: REVISE Example 1 actual outcome: REVISE Reason: individually clear picture and audio lanes contradict each other when combined; passing presentation quality cannot repair inconsistent claim scope. Example 2 expected outcome: HOLD Example 2 actual outcome: HOLD Reason: a correction created a new candidate identity and timeline; r1 evidence is historical and superseded for r2, while required r2 lanes remain untested. The two examples test different controls: 1. Independent lanes must be compared without being collapsed into one score. 2. Combined meaning can fail even when picture-only and audio-only each pass. 3. Caption evidence is separate from clean speech. 4. A complete digest binds a result to exact bytes. 5. A correction creates a new candidate and invalidates dependent evidence. 6. A still frame cannot replace complete normal-speed playback. 7. A source narration file cannot stand in for audio encoded in the master. 8. Blank evidence remains NOT TESTED. 9. Local evidence does not transfer to an unobserved destination rendition. 10. No playback result implies upload, publication, audience, customer, payment, or revenue evidence. PROVENANCE NOTE This companion, its candidates, digests, reviews, findings, timestamps, and results are original fictional method examples created by Alfred, an AI assistant. They contain no real person, customer, account, contact detail, private operational material, destination result, upload, public post, audience measure, inquiry, payment, or revenue result. No third-party media or copied interface is included. The examples exercise a local review method only and do not establish legal, rights, policy, accessibility, publication, or performance outcomes.