# Review — C-final 17 fatal · 62 major · 39 minor · 6 cosmetic · 41 measured but NOT judged engines: {"measure": {"cuts": 43, "shots": 29, "findings": 80}, "speech": {"findings": 11, "script": true}, "say": {"scored": 29, "worst": [["brightening", 0.17], ["something", 0.4], ["treating", 0.55], ["assuming", 0.55], ["aggressive", 0.56]]}, "caption": {"events": 87, "ocr_calls": 303, "coverage": 0.994, "errors": 0, "sync": {"median_lag_s": -0.34, "events": 77, "p90_abs": 0.8440000000000001}}, "overlay": {"boxes": 38, "undeclared": 6, "expected": 13}, "fit": {"devices": 2, "pacing": {"devices": 2, "per_minute": 1.3, "coverage": 0.057}, "major": 1}, "pair": {"body_overlays": 1, "not_hers": 1}, "click": {"clicks": 0, "discontinuities": 0, "seams_hit": 0, "elsewhere": 0}, "sweep": {"findings": 57, "fps": 5, "window_s": 11.0, "verified": 45, "confirmed": 34, "refuted": 6}} | At | Kind | Severity | What | |---|---|---|---| | 36.42s | speech | fatal | SHE SAYS "(nothing)" WHERE THE SCRIPT SAYS "da da da da". A viewer hears the wrong line; fix by redubbing that line in her own cloned voice (a-mispronounced-brand-is-redubbed-not-regenerated) | | 82.62s | speech | fatal | SHE SAYS "war tank's hot" WHERE THE SCRIPT SAYS "wore a tank top". A viewer hears the wrong line; fix by redubbing that line in her own cloned voice (a-mispronounced-brand-is-redubbed-not-regenerated) | | 6.467s | editing | fatal | the word "it." (6.40-6.62 s) is cut in half: the picture changes AND the track steps -10.6 dB inside it | | 50.3s | editing | fatal | the word "than" (50.14-50.38 s) is cut in half: the picture changes AND the track steps -10.3 dB inside it | | 56.3s | editing | fatal | the word "The" (56.18-56.78 s) is cut in half: the picture changes AND the track steps -15.2 dB inside it | | 64.933s | editing | fatal | the word "Put" (64.88-65.10 s) is cut in half: the picture changes AND the track steps +14.6 dB inside it | | 84.5s | editing | fatal | the word "I" (84.40-84.86 s) is cut in half: the picture changes AND the track steps -19.2 dB inside it | | 12.07s | continuity | fatal | the shot from 12.07s to 16.07s is dressed differently from the rest of the ad: torso colour correlation 0.229 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. | | 16.07s | continuity | fatal | the shot from 16.07s to 19.47s is dressed differently from the rest of the ad: torso colour correlation 0.167 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. | | 19.47s | continuity | fatal | the shot from 19.47s to 21.57s is dressed differently from the rest of the ad: torso colour correlation 0.255 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. | | 21.57s | continuity | fatal | the shot from 21.57s to 23.2s is dressed differently from the rest of the ad: torso colour correlation 0.331 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. | | 23.87s | continuity | fatal | the shot from 23.87s to 27.17s is dressed differently from the rest of the ad: torso colour correlation 0.176 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. | | 27.17s | continuity | fatal | the shot from 27.17s to 28.63s is dressed differently from the rest of the ad: torso colour correlation 0.398 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. | | 48.37s | continuity | fatal | the shot from 48.37s to 50.3s is dressed differently from the rest of the ad: torso colour correlation 0.196 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. | | 64.93s | continuity | fatal | the shot from 64.93s to 66.77s is dressed differently from the rest of the ad: torso colour correlation 0.5 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. | | 74.77s | continuity | fatal | the shot from 74.77s to 76.53s is dressed differently from the rest of the ad: torso colour correlation 0.257 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. | | 82.33s | continuity | fatal | the shot from 82.33s to 84.5s is dressed differently from the rest of the ad: torso colour correlation 0.386 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. | | 28.16s | speech | major | SHE SAYS "my" WHERE THE SCRIPT SAYS "mine". A viewer hears the wrong line; fix by redubbing that line in her own cloned voice (a-mispronounced-brand-is-redubbed-not-regenerated) | | 28.8s | speech | major | SHE SAYS "originality" WHERE THE SCRIPT SAYS "ordinary". A viewer hears the wrong line; fix by redubbing that line in her own cloned voice (a-mispronounced-brand-is-redubbed-not-regenerated) | | 43.66s | speech | major | SHE SAYS "dough" WHERE THE SCRIPT SAYS "doe". A viewer hears the wrong line; fix by redubbing that line in her own cloned voice (a-mispronounced-brand-is-redubbed-not-regenerated) | | 45.32s | speech | major | SHE SAYS "dough" WHERE THE SCRIPT SAYS "doe". A viewer hears the wrong line; fix by redubbing that line in her own cloned voice (a-mispronounced-brand-is-redubbed-not-regenerated) | | 56.78s | speech | major | SHE SAYS "dough" WHERE THE SCRIPT SAYS "doe". A viewer hears the wrong line; fix by redubbing that line in her own cloned voice (a-mispronounced-brand-is-redubbed-not-regenerated) | | 77.38s | speech | major | SHE SAYS "nocibly" WHERE THE SCRIPT SAYS "noticeably". A viewer hears the wrong line; fix by redubbing that line in her own cloned voice (a-mispronounced-brand-is-redubbed-not-regenerated) | | 80.56s | speech | major | SHE SAYS "shave" WHERE THE SCRIPT SAYS "shaved". A viewer hears the wrong line; fix by redubbing that line in her own cloned voice (a-mispronounced-brand-is-redubbed-not-regenerated) | | 23.26s | deadair | major | DEAD AIR: she stops talking for 0.74 s after "weakness." (23.26-24.00 s) while the picture keeps running; the picture cuts at 23.87 s inside the gap. A viewer sees her move with nothing said; trim the tail or fill the line | | 62.04s | deadair | major | DEAD AIR: she stops talking for 0.80 s after "whatever." (62.04-62.84 s) while the picture keeps running. A viewer sees her move with nothing said; trim the tail or fill the line | | 4.4s | editing | major | SWEEP (confirmed on the frames): The before/after underarm comparison images appear instantly with no transition, floating unnaturally clipped above the speaker's head | | 21.1s | editing | major | SWEEP (confirmed on the frames): Clip cuts off abruptly mid-speech, mid-gesture before the current spoken sentence is completed | | 40.1s | editing | major | SWEEP (confirmed on the frames): Discontinuous jump cut: the product bottle and holding hand teleport position, finger arrangement on the bottle changes instantly between frames with no natural motion transition. The gesture resets abruptly. | | 56.2s | editing | major | SWEEP (confirmed on the frames): Hard unmotivated jump cut: hand positions, head angle, background shelf objects, and which hand holds the deodorant all change instantly with no smooth transition | | 68.3s | editing | major | SWEEP (confirmed on the frames): Abrupt jump cut: camera framing jumps instantly, the woman shifts from upright to leaned forward with no smooth motion transition | | 70.0s | editing | major | SWEEP (confirmed on the frames): Abrupt hard cut at clip start, the shot begins mid-gesture and mid-speech with no lead-in, previous audio is clipped off cleanly | | 80.4s | editing | major | SWEEP (confirmed on the frames): Unmotivated jump cut: camera framing, distance and subject position change instantly with no transition, the woman's pose and arm position jump discontinuously | | 80.5s | editing | major | SWEEP (confirmed on the frames): Abrupt jump cut: the woman's head position, arm angle, and camera framing change instantly with no natural motion between shots | | 82.4s | editing | major | SWEEP (confirmed on the frames): The blue deodorant product vanishes completely from the woman's hand instantly with no movement to put it away | | 7.0s | editing | major | SWEEP (confirmed on the frames): The floating underarm comparison images vanish instantly without any transition, exactly aligned to the speech switching topic | | 9.667s | editing | major | the picture changes inside the spoken word "spent" (9.56-9.70 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. | | 12.0s | editing | major | SWEEP (confirmed on the frames): Abrupt jump cut: framing, head position, and angle of the woman shifts instantly with no natural movement or transition between shots | | 23.2s | editing | major | the picture changes inside the spoken word "weakness." (22.98-23.26 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. | | 28.633s | editing | major | the picture changes inside the spoken word "The" (28.52-28.80 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. | | 32.8s | editing | major | the picture changes inside the spoken word "every" (32.66-33.14 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. | | 42.2s | editing | major | SWEEP (confirmed on the frames): Hard jump zoom: the camera framing snaps instantly closer to the woman and product with no smooth pan/zoom motion, creating an unnatural cut. | | 48.367s | editing | major | the picture changes inside the spoken word "don't" (48.16-48.42 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. | | 72.167s | editing | major | the picture changes inside the spoken word "darkness" (71.94-72.22 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. | | 87.5s | editing | major | the picture changes inside the spoken word "not." (87.40-87.62 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. | | 88.067s | editing | major | the picture changes inside the spoken word "10" (88.04-88.46 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. | | 92.4s | editing | major | SWEEP (confirmed on the frames): On-screen caption text changes abruptly, and the new caption is placed directly over the product bottle, obscuring the deodorant label exactly when the product is being presented to the camera. | | 15.9s | lipsync | major | SWEEP (confirmed on the frames): Lip sync failure: the woman's mouth is held static open while spoken audio continues for the line, mouth movement does not match words | | 51.5s | action | major | SWEEP (confirmed on the frames): The deodorant stick never actually touches the underarm skin, it hovers noticeably away while the woman acts as if she is applying product | | 48.7s | physics | major | SWEEP (confirmed on the frames): When removing the cap, the solid deodorant stick incorrectly detaches partially and sticks to the cap for one frame, then snaps back onto the bottle base. This is physically impossible for this type of product. | | 49.3s | physics | major | SWEEP (confirmed on the frames): The loose cap clips through and hovers incorrectly over the deodorant stick top; the two solid objects do not collide or interact naturally when moved near each other. | | 92.9s | physics | major | SWEEP (confirmed on the frames): When the deodorant is moved close to the camera, the physical edges of the bottle warp slightly and the surface texture blurs in an unnatural non-real manner that real objects do not exhibit. | | 3.9s | anatomy | major | SWEEP (confirmed on the frames): Hands pop into frame abruptly; when first raised, multiple fingers are distorted, merged together with wrong geometry on both hands | | 6.1s | anatomy | major | SWEEP (confirmed on the frames): While gesturing to count items, the right hand's thumb and index finger are merged into one distorted shape, with warped nail geometry | | 10.0s | anatomy | major | SWEEP (confirmed on the frames): When hands open forward, multiple fingers have unnaturally stretched length, merged knuckles, and misaligned white nail polish that does not follow real finger shape | | 60.1s | anatomy | major | SWEEP (confirmed on the frames): Anatomy error: the gesturing right hand has distorted, warping fingers with unnatural spacing and position as it moves | | 64.9s | anatomy | major | SWEEP (confirmed on the frames): Hand anatomy error: when opening the deodorant, both hands have misformed, fused-looking fingers and an unnatural, physically impossible grip position | | 72.1s | anatomy | major | SWEEP (confirmed on the frames): Left hand holding the cap has misformed, extra and fused fingers; 6 distinct finger tips are visible incorrectly | | 84.7s | anatomy | major | SWEEP (confirmed on the frames): The woman's gesturing left hand has incorrectly formed, fused extra fingers that do not match human anatomy | | 88.6s | anatomy | major | SWEEP (confirmed on the frames): The hand holding the deodorant has an incorrect number of mis-shaped overlapping fingers | | 90.4s | anatomy | major | SWEEP (confirmed on the frames): Fingers on the woman's holding hand mutate unnaturally between frames: finger count, position, and shape shift rapidly while she grips the bottle, with merged/distorted fingers appearing briefly. | | 4.34s | overlay | major | BODY OVERLAY (before_after) IS NOT HERS: no source_frame inside this ad names the presenter frame it was made from. Fix: make the pair from a frame of this ad and name it in the plan | | 3.633s | audio-seam | major | the track steps +9.3 dB at this cut with speech running on both sides of it, which is heard as a jump in level | | 4.34s | overlay | major | OVERLAY pair-C.png: outside the platform safe zone on the top (y 0.03 < 0.15, under the TikTok and Shorts header). Fix: move or resize it in the plan and re-render | | 6.467s | audio-seam | major | the track steps -10.6 dB at this cut with speech running on both sides of it, which is heard as a jump in level | | 16.067s | audio-seam | major | the track steps +22.6 dB at this cut with speech running on both sides of it, which is heard as a jump in level | | 19.467s | audio-seam | major | the track steps +36.1 dB at this cut with speech running on both sides of it, which is heard as a jump in level | | 40.133s | audio-seam | major | the track steps -35.6 dB at this cut with speech running on both sides of it, which is heard as a jump in level | | 50.3s | audio-seam | major | the track steps -10.3 dB at this cut with speech running on both sides of it, which is heard as a jump in level | | 56.3s | audio-seam | major | the track steps -15.2 dB at this cut with speech running on both sides of it, which is heard as a jump in level | | 64.933s | audio-seam | major | the track steps +14.6 dB at this cut with speech running on both sides of it, which is heard as a jump in level | | 68.367s | audio-seam | major | the track steps +13.0 dB at this cut with speech running on both sides of it, which is heard as a jump in level | | 70.167s | audio-seam | major | the track steps +13.8 dB at this cut with speech running on both sides of it, which is heard as a jump in level | | 71.833s | audio-seam | major | the track steps +18.7 dB at this cut with speech running on both sides of it, which is heard as a jump in level | | 76.533s | audio-seam | major | the track steps +8.4 dB at this cut with speech running on both sides of it, which is heard as a jump in level | | 79.167s | audio-seam | major | the track steps +17.4 dB at this cut with speech running on both sides of it, which is heard as a jump in level | | 80.467s | audio-seam | major | the track steps +12.0 dB at this cut with speech running on both sides of it, which is heard as a jump in level | | 82.333s | audio-seam | major | the track steps -8.8 dB at this cut with speech running on both sides of it, which is heard as a jump in level | | 84.5s | audio-seam | major | the track steps -19.2 dB at this cut with speech running on both sides of it, which is heard as a jump in level | | 86.1s | speech | minor | SWEEP (unverified): Speech audio is unnaturally flat with no room acoustic echo, sounding like generated voice audio | | 12.0s | editing | minor | SWEEP (frames inconclusive): Continuity break: a small hoop earring appears out of nowhere on the woman's left ear; it was completely absent in all preceding frames | | 6.467s | editing | minor | the cut lands 67 ms after the word "name" ends, so the line has no tail before the shot changes | | 9.667s | editing | minor | the cut lands 107 ms after the word "she'd" ends, so the line has no tail before the shot changes | | 28.633s | editing | minor | the cut lands 113 ms after the word "darker." ends, so the line has no tail before the shot changes | | 31.933s | editing | minor | the cut lands 93 ms after the word "okay?" ends, so the line has no tail before the shot changes | | 40.133s | editing | minor | the cut lands 113 ms after the word "Fine," ends, so the line has no tail before the shot changes | | 47.8s | editing | minor | SWEEP (unverified): Jump cut mid-gesture: the left hand appears instantly holding the removed cap, with no frames showing the full twisting motion to unscrew it. | | 56.3s | editing | minor | the cut lands 120 ms after the word "on." ends, so the line has no tail before the shot changes | | 64.6s | editing | minor | the cut lands 60 ms after the word "cone." ends, so the line has no tail before the shot changes | | 71.833s | editing | minor | the cut lands 107 ms after the word "The" ends, so the line has no tail before the shot changes | | 77.167s | editing | minor | the cut lands 47 ms after the word "eight," ends, so the line has no tail before the shot changes | | 77.167s | editing | minor | the cut lands 93 ms after the word "it" ends, so the line has no tail before the shot changes | | 79.167s | editing | minor | the cut lands 47 ms after the word "me." ends, so the line has no tail before the shot changes | | 81.333s | editing | minor | the cut lands 7 ms after the word "morning," ends, so the line has no tail before the shot changes | | 82.333s | editing | minor | the cut lands 93 ms after the word "sting." ends, so the line has no tail before the shot changes | | 84.5s | editing | minor | the cut lands 100 ms after the word "hours." ends, so the line has no tail before the shot changes | | 87.5s | editing | minor | the cut lands 100 ms after the word "They're" ends, so the line has no tail before the shot changes | | 66.6s | lipsync | minor | SWEEP (unverified): Lipsync mismatch: the woman's mouth movements do not align correctly with the spoken phrase "for the past two months" | | 75.1s | lipsync | minor | SWEEP (unverified): Lip sync fails: spoken words 'That border started to soften and blur' do not match mouth movements, mouth is closed during open spoken vowels | | 89.1s | lipsync | minor | SWEEP (unverified): The woman's mouth movements do not properly align with the spoken words, lip sync timing is off | | 70.1s | action | minor | SWEEP (frames inconclusive): Impossible object state: the deodorant was shown fully capped closed 1 second earlier; it is now uncapped with balm extended, no cap removal action was shown, and the cap is entirely missing | | 77.2s | action | minor | SWEEP (frames inconclusive): Deodorant cap snaps onto the tube unnaturally fast without proper alignment; no physical fitting motion occurs, the cap just appears attached | | 62.4s | action | minor | SWEEP (unverified): Action mismatch: the woman holds the deodorant 1-2cm away from her nose, not touching it, but acts as if she is directly smelling the product | | 78.7s | physics | minor | SWEEP (unverified): The deodorant tube warps and bulges slightly unnaturally while pressed against her chest, plastic loses consistent rigid shape | | 58.7s | anatomy | minor | SWEEP (frames inconclusive): Right hand fingers are anatomically broken during gesturing: fingers merge together, appear duplicated then vanish, wrong finger count is visible | | 19.7s | anatomy | minor | SWEEP (unverified): Anatomical error: gesturing left hand displays warped nail shapes and extra fused abnormal fingers that change form between adjacent frames | | 68.4s | continuity | minor | SWEEP (frames inconclusive): Continuity break: a metal ring appears on the woman's right hand at this timestamp, this ring was not present on that hand in all prior frames of the clip | | 90.1s | product | minor | SWEEP (confirmed on the frames): The printed text label on the deodorant product warps, flickers, and changes completely every frame as the bottle moves. Letters morph, vanish, and rearrange unnaturally; the brand and product name never stay consistent. | | 0.1s | text | minor | SWEEP (confirmed on the frames): The upper text bubble "Glycolic acid made mine darker" appears instantly with no transition, and overlaps the existing lower subtitle before the speech references this line | | 55.8s | product | minor | SWEEP (confirmed on the frames): Printed text on the deodorant bottle warps distorts, lower line letters change shape and misalign between consecutive frames | | 70.3s | text | minor | SWEEP (confirmed on the frames): Product text error: the printed labeling on the deodorant bottle warps, distorts and becomes illegible as the woman tilts it toward the camera | | 70.4s | text | minor | SWEEP (confirmed on the frames): All smaller text below the 'Bright Days' product label is garbled, illegible nonsense letter shapes, not real readable text | | 71.8s | product | minor | SWEEP (confirmed on the frames): The white deodorant roller is already fully exposed outside the tube *before* the cap is pulled off; the cap did not cover the product head correctly | | 88.2s | text | minor | SWEEP (confirmed on the frames): Text printed on the deodorant product warps, shifts, and distorts unnaturally while the product is held perfectly still; smaller label text is illegible garbage characters | | 10.7s | text | minor | SWEEP (unverified): On-screen subtitle displays broken unnatural phrasing "with women who had exactly this." that does not match natural spoken sentence structure | | 43.0s | text | minor | SWEEP (unverified): Printed label text on the deodorant bottle warps slightly; letter shapes and spacing shift inconsistently between adjacent frames. | | 60.1s | text | minor | SWEEP (unverified): On-screen caption text overlaps incorrectly: old caption text does not fully erase before new text draws over it, creating doubled illegible letters | | 62.2s | text | minor | SWEEP (unverified): Text timing error: the "vanilla ice cream cone" on-screen caption appears 0.2 seconds before the corresponding spoken audio starts | | 50.3s | editing | cosmetic | SWEEP (the frame check disagreed, read it yourself): The deodorant cap vanishes abruptly mid-motion, it does not smoothly move out of frame, it disappears instantly with no transition | | 54.1s | editing | cosmetic | SWEEP (the frame check disagreed, read it yourself): The deodorant cap suddenly materialises out of thin air in the raised left hand, it was not present in the immediately preceding frame | | 84.4s | editing | cosmetic | SWEEP (the frame check disagreed, read it yourself): The blue deodorant product reappears instantly in the woman's hand out of nowhere, with no motion bringing it into frame | | 57.4s | lipsync | cosmetic | SWEEP (the frame check disagreed, read it yourself): Lip sync mismatch: spoken words for 'vanilla bergamot' play while the woman's mouth remains closed and does not form the correct speech shapes | | 10.8s | anatomy | cosmetic | SWEEP (the frame check disagreed, read it yourself): Anatomical error: when raising her right hand, fingers are fused, malformed and do not appear as distinct human fingers | | 18.9s | text | cosmetic | SWEEP (the frame check disagreed, read it yourself): On-screen text is broken: the full sentence line is duplicated twice, stacked and overlapping incorrectly on the frame | | 0s | lipsync | unknown | sync NOT judged: peak MFCC R2 0.143 is below the 0.15 floor every control cleared, so the mouth is not readable here; best lag -100 ms; loudness correlation -0.021 at zero lag, best 0.056 at -200 ms, 72 frames loud-with-closed-mouth, 130 open-over-silence, NOT judged (the donor reads 0.087 / 26 / 1); LipLeak 0.793 | | 0.0s | overlay | unknown | PERSISTENT GRAPHIC not on the edit track: 279x58 px at (225,285), 1 shot(s) from 0.0s to 3.63s, edge persistence 0.922, pixel std inside 50.16 vs frame motion 5.85, reads " ade mine darke"; crop /home/hq/Claude/offers/doe-deodorant/creative/awareness-doors-20261107/_qa/_overlay/overlay-225x285-0.3s.png | | 28.63s | overlay | unknown | PERSISTENT GRAPHIC not on the edit track: 13x324 px at (706,308), 4 shot(s) from 28.63s to 42.23s, edge persistence 0.993, pixel std inside 1.63 vs frame motion 11.94; crop /home/hq/Claude/offers/doe-deodorant/creative/awareness-doors-20261107/_qa/_overlay/overlay-706x308-28.9s.png | | 32.8s | overlay | unknown | PERSISTENT GRAPHIC not on the edit track: 92x470 px at (101,189), 3 shot(s) from 32.8s to 84.5s, edge persistence 0.975, pixel std inside 12.77 vs frame motion 18.64, reads "The"; crop /home/hq/Claude/offers/doe-deodorant/creative/awareness-doors-20261107/_qa/_overlay/overlay-101x189-33.1s.png | | 32.8s | overlay | unknown | PERSISTENT GRAPHIC not on the edit track: 114x405 px at (537,234), 3 shot(s) from 32.8s to 87.17s, edge persistence 0.986, pixel std inside 7.81 vs frame motion 18.64, reads "teb"; crop /home/hq/Claude/offers/doe-deodorant/creative/awareness-doors-20261107/_qa/_overlay/overlay-537x234-33.1s.png | | 82.33s | overlay | unknown | PERSISTENT GRAPHIC not on the edit track: 139x479 px at (479,245), 1 shot(s) from 82.33s to 84.5s, edge persistence 0.993, pixel std inside 6.95 vs frame motion 8.28, reads "A QUS."; crop /home/hq/Claude/offers/doe-deodorant/creative/awareness-doors-20261107/_qa/_overlay/overlay-479x245-82.6s.png | | 84.5s | overlay | unknown | PERSISTENT GRAPHIC not on the edit track: 180x414 px at (63,288), 2 shot(s) from 84.5s to 92.6s, edge persistence 0.993, pixel std inside 8.78 vs frame motion 10.23, reads "ifyo"; crop /home/hq/Claude/offers/doe-deodorant/creative/awareness-doors-20261107/_qa/_overlay/overlay-63x288-84.8s.png | | 4.34s | skin-tone | unknown | BODY OVERLAY skin a*b* (136.0, 142.4) vs the presenter's cheek (136.7, 132.9): distance 9.5 (a*-0.7, b*+9.5), reported only until an accepted pair is measured | | 0s | caption | unknown | CAPTION SYNC, reported only: median -0.34 s from the first spoken word to the caption appearing over 77 events (90th percentile 0.84 s); the correct cut reads +0.00 / 0.19 and unshifted controls read -0.24 to -0.29, so no verdict is drawn | | 0s | text | unknown | OTHER ON-SCREEN TEXT, read for a human: Days | | 3.63s | skin-tone | unknown | skin NOT JUDGED: the shot from 3.63s to 6.47s sits a*+3.6 b*-7.0 from the other shots (distance 8.1). 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. | | 5.06s | pronunciation | unknown | REDUB SHORTLIST: "glycolic" is heard by an acoustic model with no vocabulary as "emin glack coll" (overlap 0.62). 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. | | 7.0s | audio-seam | unknown | the track steps +19.4 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 | | 7.0s | skin-tone | unknown | skin NOT JUDGED: the shot from 7.0s to 9.67s sits a*+1.2 b*-3.6 from the other shots (distance 3.8). 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. | | 9.67s | skin-tone | unknown | skin NOT JUDGED: the shot from 9.67s to 12.07s sits a*+0.8 b*-3.2 from the other shots (distance 3.3). 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. | | 12.067s | audio-seam | unknown | the track steps +14.1 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 | | 13.74s | pronunciation | unknown | REDUB SHORTLIST: "treating" is heard by an acoustic model with no vocabulary as "tre" (overlap 0.55). 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. | | 16.07s | skin-tone | unknown | skin NOT JUDGED: the shot from 16.07s to 19.47s sits a*+0.9 b*-3.2 from the other shots (distance 4.9). 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. | | 17.26s | pronunciation | unknown | REDUB SHORTLIST: "something" is heard by an acoustic model with no vocabulary as "ping sp" (overlap 0.4). 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. | | 19.47s | skin-tone | unknown | skin NOT JUDGED: the shot from 19.47s to 21.57s sits a*-0.3 b*-3.6 from the other shots (distance 4.2). 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. | | 20.82s | pronunciation | unknown | REDUB SHORTLIST: "aggressive" is heard by an acoustic model with no vocabulary as "more gress" (overlap 0.56). 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. | | 21.567s | audio-seam | unknown | the track steps +18.9 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 | | 21.57s | skin-tone | unknown | skin NOT JUDGED: the shot from 21.57s to 23.2s sits a*-0.6 b*-3.2 from the other shots (distance 3.8). 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. | | 23.867s | audio-seam | unknown | the track steps +21.1 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 | | 23.87s | skin-tone | unknown | skin NOT JUDGED: the shot from 23.87s to 27.17s sits a*-0.2 b*-3.3 from the other shots (distance 5.2). 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. | | 24.7s | pronunciation | unknown | REDUB SHORTLIST: "assuming" is heard by an acoustic model with no vocabulary as "s su" (overlap 0.55). 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. | | 27.17s | skin-tone | unknown | skin NOT JUDGED: the shot from 27.17s to 28.63s sits a*-2.2 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. | | 28.63s | skin-tone | unknown | skin NOT JUDGED: the shot from 28.63s to 31.93s sits a*-3.2 b*-0.5 from the other shots (distance 3.3). 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. | | 33.34s | pronunciation | unknown | REDUB SHORTLIST: "brightening" is heard by an acoustic model with no vocabulary as "r" (overlap 0.17). 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. | | 38.63s | skin-tone | unknown | skin NOT JUDGED: the shot from 38.63s to 40.13s sits a*-3.1 b*+2.2 from the other shots (distance 4.4). 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. | | 42.23s | skin-tone | unknown | skin NOT JUDGED: the shot from 42.23s to 44.63s sits a*+3.5 b*+0.6 from the other shots (distance 5.2). 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. | | 44.633s | audio-seam | unknown | the track steps +20.0 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 | | 48.37s | skin-tone | unknown | skin NOT JUDGED: the shot from 48.37s to 50.3s sits a*+6.0 b*+3.6 from the other shots (distance 7.1). 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. | | 64.93s | skin-tone | unknown | skin NOT JUDGED: the shot from 64.93s to 66.77s sits a*+4.8 b*+2.8 from the other shots (distance 5.6). 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. | | 74.77s | skin-tone | unknown | skin NOT JUDGED: the shot from 74.77s to 76.53s sits a*-4.7 b*+1.1 from the other shots (distance 6.4). 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. | | 77.17s | skin-tone | unknown | skin NOT JUDGED: the shot from 77.17s to 79.17s sits a*-3.1 b*+0.7 from the other shots (distance 5.4). 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. | | 82.33s | skin-tone | unknown | skin NOT JUDGED: the shot from 82.33s to 84.5s sits a*-3.8 b*+1.6 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. | | 84.5s | skin-tone | unknown | skin NOT JUDGED: the shot from 84.5s to 87.17s sits a*-0.7 b*-3.0 from the other shots (distance 3.0). 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. | | 87.167s | audio-seam | unknown | the track steps +28.6 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 | | 88.067s | audio-seam | unknown | the track steps +24.6 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 | | 92.6s | audio-seam | unknown | the track steps +32.1 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 |