# Review — C-final 4 fatal · 2 major · 11 minor · 2 cosmetic · 16 measured but NOT judged engines: {"measure": {"cuts": 12, "shots": 9, "findings": 29}, "say": {"scored": 18, "worst": [["hairdresser", 0.13], ["instantly", 0.57], ["spraying", 0.71], ["brushing", 0.71], ["protection", 0.71]]}, "caption": {"events": 33, "ocr_calls": 60, "coverage": 0.852, "errors": 0, "sync": {"median_lag_s": -0.09, "events": 31, "p90_abs": 0.19}}, "overlay": {"boxes": 9, "undeclared": 1, "expected": 3}, "fit": {"devices": 2, "pacing": {"devices": 2, "per_minute": 2.3, "coverage": 0.126}, "major": 0}, "pair": {"body_overlays": 1, "not_hers": 0}, "click": {"clicks": 0, "discontinuities": 0, "seams_hit": 0, "elsewhere": 0}} | At | Kind | Severity | What | |---|---|---|---| | 33.6s | continuity | fatal | the shot from 33.6s to 37.16s is dressed differently from the rest of the ad: torso colour correlation 0.375 against the other shots, measured over 14 frames. The donor's own least consistent shot measures 0.639 and the take Kalp rejected for its top measured 0.374. | | 33.6s | identity | fatal | the face in the shot from 33.6s to 37.16s does not match the rest of the ad: SFace cosine 0.517 against the other shots (every labelled control sits above 0.82) | | 37.16s | continuity | fatal | the shot from 37.16s to 41.76s is dressed differently from the rest of the ad: torso colour correlation 0.4 against the other shots, measured over 11 frames. The donor's own least consistent shot measures 0.639 and the take Kalp rejected for its top measured 0.374. | | 46.88s | identity | fatal | the face in the shot from 46.88s to 51.72s does not match the rest of the ad: SFace cosine 0.712 against the other shots (every labelled control sits above 0.82) | | 14.24s | editing | major | the picture changes inside the spoken word "are" (14.10-14.62 s). The track runs on through it, so the line is whole and the EDIT lands mid-word, which reads as clumsy rather than broken. | | 20.12s | editing | major | the picture changes inside the spoken word "hair." (20.04-20.34 s). The track runs on through it, so the line is whole and the EDIT lands mid-word, which reads as clumsy rather than broken. | | 0.0s | overlay | minor | STICKER 775b9394_C-hook-keyed.png: on screen 4.0 s; the swipes hold an overlay at most 3.0 s (median 1.2 s) | | 18.24s | editing | minor | the cut lands 20 ms after the word "causing" ends, so the line has no tail before the shot changes | | 19.72s | editing | minor | the cut lands 120 ms after the word "grabby," ends, so the line has no tail before the shot changes | | 20.12s | editing | minor | the cut lands 80 ms after the word "knotted" ends, so the line has no tail before the shot changes | | 23.04s | editing | minor | the cut lands 0 ms after the word "just" ends, so the line has no tail before the shot changes | | 24.76s | editing | minor | the cut lands 100 ms after the word "Then" ends, so the line has no tail before the shot changes | | 24.76s | editing | minor | the cut lands 20 ms after the word "you" ends, so the line has no tail before the shot changes | | 33.6s | editing | minor | the cut lands 20 ms after the word "let" ends, so the line has no tail before the shot changes | | 33.6s | editing | minor | the cut lands 80 ms after the word "me" ends, so the line has no tail before the shot changes | | 41.76s | editing | minor | the cut lands 20 ms after the word "through." ends, so the line has no tail before the shot changes | | 46.88s | editing | minor | the cut lands 20 ms after the word "strengthening." ends, so the line has no tail before the shot changes | | 0s | lipsync | cosmetic | sync in window: MFCC best lag +0 ms (peak R2 0.18, at zero 0.18); loudness correlation -0.026 at zero lag, best 0.101 at -240 ms, 3 frames loud-with-closed-mouth, 95 open-over-silence, NOT judged (the donor reads 0.087 / 26 / 1); LipLeak 0.674 | | 37.16s | editing | cosmetic | the cut sits in a silent pause that the transcript spans as the word "and" (36.94-37.40 s): a cut in a gap, not in a word | | 23.04s | overlay | unknown | PERSISTENT GRAPHIC not on the edit track: 65x46 px at (320,211), 1 shot(s) from 23.04s to 24.76s, edge persistence 0.901, pixel std inside 59.6 vs frame motion 16.21, reads "工 1."; crop /root/r-604ea15dad/out/_overlay/overlay-320x211-23.3s.png | | 31.06s | skin-tone | unknown | BODY OVERLAY skin a*b* (137.7, 142.3) vs the presenter's cheek (142.9, 143.9): distance 5.4 (a*-5.2, b*-1.6), reported only until an accepted pair is measured | | 0s | cadence | unknown | CADENCE, for the eye only: the first 10 s carry 0 visible change(s), one every 10.0 s; the general lean in short-form is one every 1-2 s. The whole ad runs one every 4.0 s. A yapper or an older audience wants longer holds, so this is case by case and never a defect. | | 0.0s | skin-tone | unknown | skin NOT JUDGED: the shot from 0.0s to 12.04s sits a*+1.3 b*+3.8 from the other shots (distance 4.8). The donor reaches 7.1 at its own cuts, so this scale cannot separate a light change from a skin change. A human decides; positive a* and b* together read as warmer and more yellow. | | 0s | caption | unknown | CAPTION SYNC, reported only: median -0.09 s from the first spoken word to the caption appearing over 31 events (90th percentile 0.19 s); the correct cut reads +0.00 / 0.19 and unshifted controls read -0.24 to -0.29, so no verdict is drawn | | 6.8s | pronunciation | unknown | REDUB SHORTLIST: "hairdresser" is heard by an acoustic model with no vocabulary as "jus o" (overlap 0.13). Whisper reports the right word because its language model supplies it. The donor control slurs its own words to 0.38, so this is a ranking for a human, not a verdict. | | 14.24s | audio-seam | unknown | the track steps -23.4 dB here, and one side of the cut is silence, so this is a speaker starting or stopping rather than a seam in the sound | | 14.24s | skin-tone | unknown | skin NOT JUDGED: the shot from 14.24s to 18.24s sits a*+1.6 b*+3.8 from the other shots (distance 4.5). The donor reaches 7.1 at its own cuts, so this scale cannot separate a light change from a skin change. A human decides; positive a* and b* together read as warmer and more yellow. | | 18.24s | continuity | unknown | the shot from 18.24s to 19.72s is NOT JUDGED: a face was found in only 4 of the frames sampled from it, which is too little to profile. Its numbers are top_hist 0.013, SFace 0.082, skin distance 14.3. | | 20.12s | skin-tone | unknown | skin NOT JUDGED: the shot from 20.12s to 23.04s sits a*+1.6 b*+3.7 from the other shots (distance 4.7). The donor reaches 7.1 at its own cuts, so this scale cannot separate a light change from a skin change. A human decides; positive a* and b* together read as warmer and more yellow. | | 33.08s | audio-seam | unknown | the track steps -19.0 dB here, and one side of the cut is silence, so this is a speaker starting or stopping rather than a seam in the sound | | 33.6s | skin-tone | unknown | skin NOT JUDGED: the shot from 33.6s to 37.16s sits a*-2.8 b*-4.6 from the other shots (distance 6.8). The donor reaches 7.1 at its own cuts, so this scale cannot separate a light change from a skin change. A human decides; positive a* and b* together read as warmer and more yellow. | | 37.16s | skin-tone | unknown | skin NOT JUDGED: the shot from 37.16s to 41.76s sits a*-4.0 b*-7.1 from the other shots (distance 10.7). The donor reaches 7.1 at its own cuts, so this scale cannot separate a light change from a skin change. A human decides; positive a* and b* together read as warmer and more yellow. | | 38.18s | pronunciation | unknown | REDUB SHORTLIST: "instantly" is heard by an acoustic model with no vocabulary as "insol" (overlap 0.57). Whisper reports the right word because its language model supplies it. The donor control slurs its own words to 0.38, so this is a ranking for a human, not a verdict. | | 46.88s | audio-seam | unknown | the track steps -10.6 dB here, and one side of the cut is silence, so this is a speaker starting or stopping rather than a seam in the sound | | 46.88s | skin-tone | unknown | skin NOT JUDGED: the shot from 46.88s to 51.72s sits a*-5.5 b*-4.7 from the other shots (distance 8.3). The donor reaches 7.1 at its own cuts, so this scale cannot separate a light change from a skin change. A human decides; positive a* and b* together read as warmer and more yellow. |