# moneycanbuy.fun raw footage delivery

moneycanbuy.fun may use, change and publish these files without attribution.

## Format and provenance

All 14 clips were generated from separate image-model start stills with the Kling 2.6 Pro image-to-video model. Clips are separate continuous shots. The stills are the full-size images submitted as model start frames. No still was upscaled by us; videos were generated at native vertical dimensions. The delivered H.264 MP4s were trimmed from 121 to exactly 120 frames, encoded yuv420p with moov at the front, and contain **no audio track** as the clip brief requests. This conflicts with the background acceptance criterion that asks for AAC audio. Twelve clips are native 1080×1920; D04 and D07 are the model's native 1076×1920. Every clip is 5.000 s at 24 fps. The complete artifacts directory is about 52 MiB, with each clip well under 64 MiB and the contact sheet under 10 MiB.

The music is a real Google Lyria 3 instrumental generation. The delivered WAV is 30.000 s, 48 kHz, signed 16-bit stereo. Measured integrated loudness is -15.03 LUFS and true peak is -1.07 dBTP after allowed gain, trim, fade and MP3-to-WAV conversion. Model logs identified the output as instrumental. Listening verification of the absence of any vocal-like sound was unavailable.

## Per-file record

| File under artifacts/ | Tool and model/version | Exact generation or assembly prompt | Regeneration and reason / observed limitation |
| --- | --- | --- | --- |
| start/D01.png | OpenAI image_gen built-in (version not exposed) | Photoreal vertical live-action phone footage start still. In a clean generic subway car with absolutely blank walls and no signs, use the exact young man in the reference image: same face, short dark curly hair, dark charcoal hoodie, black jeans. He holds a modern phone vertically toward the camera in his right hand while looking down and scrolling with his left thumb. The entire phone display is a flat, solid, evenly lit pure bright chroma green #00FF00 rectangle with all four corners and edges visible, no glare, no reflection, no interface. Natural fluorescent daylight, everyday candid realism, slight handheld composition. No writing, letters, numbers, logos, print, symbols, advertising, brands, watermarks anywhere. | Original generated still. |
| start/D02.png | OpenAI image_gen built-in (version not exposed) | Photoreal live-action vertical 9:16 start still, candid phone footage aesthetic, natural true colors, real skin and fabric, slight handheld framing, no cinematic color grading. Every visible sign, screen, clothing item and package is blank. No writing, letters, numbers, brand marks, logos, watermarks or symbols anywhere. Every green phone display is a rigid flat solid pure #00FF00 panel, evenly lit, camera-facing, all four edges and corners unobstructed, no glare or interface. Fictional night city square with bright abstract colored light panels high above, none resembling real Times Square brands. The exact young man from the reference, same face hair charcoal hoodie jeans, has climbed onto a simple bench, holding up a completely blank rigid cardboard sign and opening his mouth to shout. Several ordinary passersby around him turn and hold phones with perfectly visible green screens toward camera. The sign is empty and flat. | Original generated still. |
| start/D03.png | OpenAI image_gen built-in (version not exposed) | Photoreal live-action vertical 9:16 start still, candid phone footage aesthetic, natural true colors, real skin and fabric, slight handheld framing, no cinematic color grading. Every visible sign, screen, clothing item and package is blank. No writing, letters, numbers, brand marks, logos, watermarks or symbols anywhere. Every green phone display is a rigid flat solid pure #00FF00 panel, evenly lit, camera-facing, all four edges and corners unobstructed, no glare or interface. The same fictional night square and bench, abstract colored lights. The exact young man from reference in charcoal hoodie and black jeans still stands on the bench, catching his breath, shoulders rising, and holds a phone toward camera with the green display fully visible, finger poised to tap its center. Maintain exact cast identity. | Original generated still. |
| start/D04.png | OpenAI image_gen built-in (version not exposed) | Photoreal live-action vertical 9:16 start still, candid phone footage aesthetic, natural true colors, real skin and fabric, slight handheld framing, no cinematic color grading. Every visible sign, screen, clothing item and package is blank. No writing, letters, numbers, brand marks, logos, watermarks or symbols anywhere. Every green phone display is a rigid flat solid pure #00FF00 panel, evenly lit, camera-facing, all four edges and corners unobstructed, no glare or interface. Warm everyday living room with blank walls. The exact adult man from the reference, short black hair, modest beard, plain rust t-shirt and dark blue jeans, sits on a sofa with a completely blank metallic chips bag beside him. He smirks and holds a camera-facing green-screen phone while a finger approaches the middle to tap it. Keep exact face and clothes. | Original generated still. |
| start/D05a.png | OpenAI image_gen built-in (version not exposed) | Photoreal live-action vertical 9:16 start still, candid phone footage aesthetic, natural true colors, real skin and fabric, slight handheld framing, no cinematic color grading. Every visible sign, screen, clothing item and package is blank. No writing, letters, numbers, brand marks, logos, watermarks or symbols anywhere. Every green phone display is a rigid flat solid pure #00FF00 panel, evenly lit, camera-facing, all four edges and corners unobstructed, no glare or interface. An ordinary fictional adult woman with dark skin, short natural hair, plain cream sweater and navy trousers sits on a park bench in natural daylight. She raises a phone toward the camera and her finger is near the center of the visible green screen, about to tap once. Trees and walkway in the background. | Original generated still. |
| start/D05b.png | OpenAI image_gen built-in (version not exposed) | Photoreal live-action vertical 9:16 start still, candid phone footage aesthetic, natural true colors, real skin and fabric, slight handheld framing, no cinematic color grading. Every visible sign, screen, clothing item and package is blank. No writing, letters, numbers, brand marks, logos, watermarks or symbols anywhere. Every green phone display is a rigid flat solid pure #00FF00 panel, evenly lit, camera-facing, all four edges and corners unobstructed, no glare or interface. A different ordinary fictional adult man with East Asian appearance, straight dark hair, plain light blue shirt and gray trousers sits safely in the passenger seat of a parked car in daylight. He raises a phone toward the camera and his fingertip approaches the middle of the visible green screen for one tap. No driving or dashboard labels. | Original generated still. |
| start/D05c.png | OpenAI image_gen built-in (version not exposed) | Photoreal live-action vertical 9:16 start still, candid phone footage aesthetic, natural true colors, real skin and fabric, slight handheld framing, no cinematic color grading. Every visible sign, screen, clothing item and package is blank. No writing, letters, numbers, brand marks, logos, watermarks or symbols anywhere. Every green phone display is a rigid flat solid pure #00FF00 panel, evenly lit, camera-facing, all four edges and corners unobstructed, no glare or interface. A different ordinary fictional adult woman with light skin, brown curly bob hair, plain olive top sits in a small cafe at a plain table in warm daylight. She raises a phone toward camera and her finger is near the middle of its visible bright green screen for one tap. All menus and cups blank. | Original generated still. |
| start/D05d.png | OpenAI image_gen built-in (version not exposed) | Photoreal live-action vertical 9:16 start still, candid phone footage aesthetic, natural true colors, real skin and fabric, slight handheld framing, no cinematic color grading. Every visible sign, screen, clothing item and package is blank. No writing, letters, numbers, brand marks, logos, watermarks or symbols anywhere. Every green phone display is a rigid flat solid pure #00FF00 panel, evenly lit, camera-facing, all four edges and corners unobstructed, no glare or interface. A different ordinary fictional adult man with medium brown skin, shaved head, plain light gray t-shirt stands on a simple apartment balcony in daylight. He raises a phone toward camera and his finger is near the center of its visible bright green display for one tap. No signage in city background. | Original animated still restored; a later shirt edit was not used for animation, so its print remains. |
| start/D06.png | OpenAI image_gen built-in (version not exposed) | Photoreal live-action vertical 9:16 start still, candid phone footage aesthetic, natural true colors, real skin and fabric, slight handheld framing, no cinematic color grading. Every visible sign, screen, clothing item and package is blank. No writing, letters, numbers, brand marks, logos, watermarks or symbols anywhere. Every green phone display is a rigid flat solid pure #00FF00 panel, evenly lit, camera-facing, all four edges and corners unobstructed, no glare or interface. The exact young man from reference, same face and hair, charcoal hoodie and black jeans, at home in a plain room, angry expression. He holds a phone toward camera, its full green display visible, and prepares to jab the middle of the screen firmly with a finger. No other screens or text. | Original generated still. |
| start/D07.png | OpenAI image_gen built-in (version not exposed) | Photoreal vertical candid phone photograph of the same five green frog-mask judges and courtroom as the attached images. Recompose this start still so the rectangular bright green wall screen is smaller and completely INSIDE the upper half of frame, with generous visible wall space around all four sides; all four screen corners must be visible. Five judges remain in one row behind one long bench in the lower half, black robes, all five rigid chest plaques completely visible with no hand covering any edge or corner. The bench front green rectangle must also be fully visible. Every green area is uniformly lit solid pure #00FF00, no reflections or glare. Central judge holds a rigid blank gavel. Real warm indoor light, real costume masks, no words, letters, numbers, marks, logos, symbols or watermarks anywhere. | Regenerated once because the first wall screen was cropped. |
| start/D08.png | OpenAI image_gen built-in (version not exposed) | Photoreal live-action vertical 9:16 start still, candid phone footage aesthetic, natural true colors, real skin and fabric, slight handheld framing, no cinematic color grading. No writing, letters, numbers, brand marks, logos, watermarks or symbols anywhere except the single '$' on the toad's chest tag. Keep faces and clothing exactly as character references. All blank props are rigid, clean, and free of print. Same five distinct simple green frog-mask judges in black robes from reference behind one long courtroom bench, the central head judge raising a completely blank rigid wooden gavel for one strike, the others attentive. Each rigid chest plaque and the entire bench front panel are flat solid #00FF00 green, camera facing, evenly lit, edges and corners visible, no glare. No wall writing, seals or signs. | Original generated still. |
| start/D09.png | OpenAI image_gen built-in (version not exposed) | Photoreal live-action vertical 9:16 start still, candid phone footage aesthetic, natural true colors, real skin and fabric, slight handheld framing, no cinematic color grading. No writing, letters, numbers, brand marks, logos, watermarks or symbols anywhere except the single '$' on the toad's chest tag. Keep faces and clothing exactly as character references. All blank props are rigid, clean, and free of print. In the fictional night square with only abstract colored light panels, the exact human toad mascot from reference stands in front three-quarter view, large flat-green plush head, heavy half-closed lids over cream eyes, wide flat smirk, nostrils, dark cheek spots, plain dark hoodie, thick gold chain and '$' tag always visible. Beside him the exact young man from reference in charcoal hoodie and black jeans, cupping his hands. The toad holds a rigid clear locked box full of completely blank plain gold coins, preparing to pour them into the doer's hands. True-color street phone photo, no other signage or writing. | Original generated still. |
| start/D10.png | OpenAI image_gen built-in (version not exposed) | Photoreal live-action vertical 9:16 start still, candid phone footage aesthetic, natural true colors, real skin and fabric, slight handheld framing, no cinematic color grading. No writing, letters, numbers, brand marks, logos, watermarks or symbols anywhere except the single '$' on the toad's chest tag. Keep faces and clothing exactly as character references. All blank props are rigid, clean, and free of print. Fictional night square with abstract colored lights only. Three people together: exact reference toad mascot from front three-quarter view, gold chain and '$' tag clearly visible; exact reference hater in plain rust t-shirt holding a rigid clear coin box and a completely blank metallic chips bag; exact reference doer in charcoal hoodie waiting to receive the box. The toad reaches to take the box from the hater and hand it to the doer. No printed designs, all coins blank. Natural candid phone photo. | Original generated still. |
| start/D11.png | OpenAI image_gen built-in (version not exposed) | Photoreal live-action vertical 9:16 start still, candid phone footage aesthetic, natural true colors, real skin and fabric, slight handheld framing, no cinematic color grading. No writing, letters, numbers, brand marks, logos, watermarks or symbols anywhere except the single '$' on the toad's chest tag. Keep faces and clothing exactly as character references. All blank props are rigid, clean, and free of print. Fictional night square with abstract colored lights only. Exact human toad mascot from reference in a plain dark hoodie, front three-quarter view, same flat-green plush head with heavy half-closed lids, cream eyes, dark cheek spots, flat smirk, small nostrils, thick gold chain and '$' tag clearly visible. The exact young man in charcoal hoodie stands beside him. In front is a simple metal bowl containing real live fire; the toad holds a handful of rigid plain blank gold coins above the flames, with a small completely blank rigid golden trophy cup in his other hand ready to hand over. Photoreal true-color candid phone image. | Original generated still. |
| clips/D01.mp4 | Kling 2.6 Pro / kwaivgi/kling-v2.6, b13f36d030496dd78d2986ba8b2b22a44222b3f58c15fb63ef7d6b4aa3a53319 | One continuous candid handheld phone shot, photoreal real life. The young man in the dark charcoal hoodie inside a subway car scrolls his bright flat green phone screen toward the camera with one thumb, then stops suddenly and stares at the phone in surprise. His face, hair, clothes, body, and the blank green phone remain unchanged throughout. Natural motion at normal speed for the full 5 seconds; subtle moving train and people in the background. No audio, no cuts, no camera pan or zoom, no writing or logos anywhere. | Trimmed one model frame and re-encoded to 120 frames at native size. Phone display gains dark UI marks late in the clip. |
| clips/D02.mp4 | Kling 2.6 Pro / kwaivgi/kling-v2.6, b13f36d030496dd78d2986ba8b2b22a44222b3f58c15fb63ef7d6b4aa3a53319 | Photoreal ordinary handheld phone footage, vertical one continuous real-time shot for 5 seconds with active natural movement throughout. Keep every person's face, hair, body and clothes fixed exactly as the start image. Keep all props rigid. No cuts, no slow motion, no camera zoom or pan, no audio, no words, letters, numbers, logos, branding or watermarks. Every phone screen stays flat pure bright #00FF00, camera-facing, all four corners visible, no glare or interface. The doer in charcoal hoodie climbs fully onto the bench holding the rigid blank cardboard sign and shouts YOLO aloud with visible mouth movement while passersby turn toward him and raise their bright green-screen phones to film. The square lights are abstract color only. | Trimmed one model frame and re-encoded to 120 frames at native size. Some passerby phone displays do not remain flat green. |
| clips/D03.mp4 | Kling 2.6 Pro / kwaivgi/kling-v2.6, b13f36d030496dd78d2986ba8b2b22a44222b3f58c15fb63ef7d6b4aa3a53319 | Photoreal ordinary handheld phone footage, vertical one continuous real-time shot for 5 seconds with active natural movement throughout. Keep every person's face, hair, body and clothes fixed exactly as the start image. Keep all props rigid. No cuts, no slow motion, no camera zoom or pan, no audio, no words, letters, numbers, logos, branding or watermarks. Every phone screen stays flat pure bright #00FF00, camera-facing, all four corners visible, no glare or interface. The doer remains on the bench catching his breath, chest moving, then reaches forward and taps once in the middle of his bright green phone display. Night square pedestrians keep moving naturally. | Trimmed one model frame and re-encoded to 120 frames at native size. No major visual failure seen in three sampled frames; tap timing is not fully certified. |
| clips/D04.mp4 | Kling 2.6 Pro / kwaivgi/kling-v2.6, b13f36d030496dd78d2986ba8b2b22a44222b3f58c15fb63ef7d6b4aa3a53319 | Photoreal ordinary handheld phone footage, vertical one continuous real-time shot for 5 seconds with active natural movement throughout. Keep every person's face, hair, body and clothes fixed exactly as the start image. Keep all props rigid. No cuts, no slow motion, no camera zoom or pan, no audio, no words, letters, numbers, logos, branding or watermarks. Every phone screen stays flat pure bright #00FF00, camera-facing, all four corners visible, no glare or interface. The hater on the sofa smirks, lifts the blank chips bag slightly, then taps once in the middle of his bright green phone screen. Natural living room movement. | Trimmed one model frame and re-encoded to 120 frames at native size. No major visual failure seen in three sampled frames. |
| clips/D05a.mp4 | Kling 2.6 Pro / kwaivgi/kling-v2.6, b13f36d030496dd78d2986ba8b2b22a44222b3f58c15fb63ef7d6b4aa3a53319 | Photoreal ordinary handheld phone footage, vertical one continuous real-time shot for 5 seconds with active natural movement throughout. Keep every person's face, hair, body and clothes fixed exactly as the start image. Keep all props rigid. No cuts, no slow motion, no camera zoom or pan, no audio, no words, letters, numbers, logos, branding or watermarks. Every phone screen stays flat pure bright #00FF00, camera-facing, all four corners visible, no glare or interface. The woman on the park bench raises her bright green-screen phone toward the camera and taps it exactly once with a natural hand motion; people and leaves move in the background. | Trimmed one model frame and re-encoded to 120 frames at native size. The model introduces small lettering on the phone at middle frame. |
| clips/D05b.mp4 | Kling 2.6 Pro / kwaivgi/kling-v2.6, b13f36d030496dd78d2986ba8b2b22a44222b3f58c15fb63ef7d6b4aa3a53319 | Photoreal ordinary handheld phone footage, vertical one continuous real-time shot for 5 seconds with active natural movement throughout. Keep every person's face, hair, body and clothes fixed exactly as the start image. Keep all props rigid. No cuts, no slow motion, no camera zoom or pan, no audio, no words, letters, numbers, logos, branding or watermarks. Every green screen or plaque stays flat pure bright #00FF00, camera-facing, all four corners visible, no glare or interface. The man in the parked car passenger seat raises his bright green-screen phone toward the camera and taps it exactly once with a natural hand motion, then lowers his tapping finger. Subtle everyday movement through the window. | Trimmed one model frame and re-encoded to 120 frames at native size. The phone acquires a pale reflected shape at middle frame and lowers by the end. |
| clips/D05c.mp4 | Kling 2.6 Pro / kwaivgi/kling-v2.6, b13f36d030496dd78d2986ba8b2b22a44222b3f58c15fb63ef7d6b4aa3a53319 | Photoreal ordinary handheld phone footage, vertical one continuous real-time shot for 5 seconds with active natural movement throughout. Keep every person's face, hair, body and clothes fixed exactly as the start image. Keep all props rigid. No cuts, no slow motion, no camera zoom or pan, no audio, no words, letters, numbers, logos, branding or watermarks. Every green screen or plaque stays flat pure bright #00FF00, camera-facing, all four corners visible, no glare or interface. The woman in the cafe raises her bright green-screen phone toward the camera and taps it exactly once with a natural hand motion; other people move softly behind her. | Trimmed one model frame and re-encoded to 120 frames at native size. No major visual failure seen in three sampled frames. |
| clips/D05d.mp4 | Kling 2.6 Pro / kwaivgi/kling-v2.6, b13f36d030496dd78d2986ba8b2b22a44222b3f58c15fb63ef7d6b4aa3a53319 | Photoreal ordinary handheld phone footage, vertical one continuous real-time shot for 5 seconds with active natural movement throughout. Keep every person's face, hair, body and clothes fixed exactly as the start image. Keep all props rigid. No cuts, no slow motion, no camera zoom or pan, no audio, no words, letters, numbers, logos, branding or watermarks. Every green screen or plaque stays flat pure bright #00FF00, camera-facing, all four corners visible, no glare or interface. The man on the balcony raises his bright green-screen phone toward the camera and taps it exactly once with a natural hand motion; slight real breeze and city background movement. | Trimmed one model frame and re-encoded to 120 frames at native size. A printed emblem is visible on the shirt and UI appears on the phone. |
| clips/D06.mp4 | Kling 2.6 Pro / kwaivgi/kling-v2.6, b13f36d030496dd78d2986ba8b2b22a44222b3f58c15fb63ef7d6b4aa3a53319 | Photoreal ordinary handheld phone footage, vertical one continuous real-time shot for 5 seconds with active natural movement throughout. Keep every person's face, hair, body and clothes fixed exactly as the start image. Keep all props rigid. No cuts, no slow motion, no camera zoom or pan, no audio, no words, letters, numbers, logos, branding or watermarks. Every green screen or plaque stays flat pure bright #00FF00, camera-facing, all four corners visible, no glare or interface. The young man in his charcoal hoodie at home looks angry and jabs the center of the bright green phone screen once, then recoils a little. His face and clothes stay identical. | Trimmed one model frame and re-encoded to 120 frames at native size. No major visual failure seen in three sampled frames. |
| clips/D07.mp4 | Kling 2.6 Pro / kwaivgi/kling-v2.6, b13f36d030496dd78d2986ba8b2b22a44222b3f58c15fb63ef7d6b4aa3a53319 | Photoreal ordinary handheld phone footage, vertical one continuous real-time shot for 5 seconds with active natural movement throughout. Keep every person's face, hair, body and clothes fixed exactly as the start image. Keep all props rigid. No cuts, no slow motion, no camera zoom or pan, no audio, no words, letters, numbers, logos, branding or watermarks. Every green screen or plaque stays flat pure bright #00FF00, camera-facing, all four corners visible, no glare or interface. Five judges wearing simple green frog masks sit behind one long bench. They turn their heads to watch the fully visible pure green wall screen, then turn toward each other and confer with small hand gestures. Each of five chest plaques and the bench front stay camera-facing and fully unobstructed. The wall screen remains within frame. | Trimmed one model frame and re-encoded to 120 frames at native size. The model introduces tiny lettering on the bench panel by the end; a plaque edge is partly covered. |
| clips/D08.mp4 | Kling 2.6 Pro / kwaivgi/kling-v2.6, b13f36d030496dd78d2986ba8b2b22a44222b3f58c15fb63ef7d6b4aa3a53319 | One continuous uncut five-second photoreal vertical phone-camera courtroom video at normal speed. The central head judge in a plain green frog mask lowers the rigid wooden gavel decisively from above shoulder to the bench surface during the first two seconds, makes one unmistakable strike with contact, then keeps the gavel on the bench. The other four green frog-mask judges nod once after the strike. Keep all five judges the same, robes fixed, no face morphs. Keep five chest plaques and front bench panel bright solid flat #00FF00 with no UI, reflections, marks or text; all edges and corners visible. No camera move, no sound, no cuts, no writing, numbers, brands, logos or watermark. | Trimmed one model frame and re-encoded to 120 frames at native size. The sampled frames do not show clear gavel contact with the bench. A Kling v3 regeneration was rejected with HTTP 402 Insufficient credit. |
| clips/D09.mp4 | Kling 2.6 Pro / kwaivgi/kling-v2.6, b13f36d030496dd78d2986ba8b2b22a44222b3f58c15fb63ef7d6b4aa3a53319 | Photoreal ordinary handheld phone footage, vertical one continuous real-time shot for 5 seconds with active natural movement throughout. Keep every person's face, hair, body and clothes fixed exactly as the start image. Keep all props rigid. No cuts, no slow motion, no camera zoom or pan, no audio, no words, letters, numbers, logos, branding or watermarks. The toad mascot is the exact same giant flat green plush head with heavy half-closed lids, cream eyes, wide flat smirk, tiny nostrils, dark cheek spots; its thick gold chain and single '$' tag remain visible at all times. Never show the toad from behind. In the fictional night square the toad mascot tips the rigid clear locked coin box and pours plain gold coins into the doer's cupped hands. Coins tumble with real gravity, keeping rigid round shapes. The doer reacts happily, same charcoal hoodie and face. Abstract colored lights in the background. | Trimmed one model frame and re-encoded to 120 frames at native size. No major visual failure seen in three sampled frames. |
| clips/D10.mp4 | Kling 2.6 Pro / kwaivgi/kling-v2.6, b13f36d030496dd78d2986ba8b2b22a44222b3f58c15fb63ef7d6b4aa3a53319 | Photoreal ordinary handheld phone footage, vertical one continuous real-time shot for 5 seconds with active natural movement throughout. Keep every person's face, hair, body and clothes fixed exactly as the start image. Keep all props rigid. No cuts, no slow motion, no camera zoom or pan, no audio, no words, letters, numbers, logos, branding or watermarks. The toad mascot is the exact same giant flat green plush head with heavy half-closed lids, cream eyes, wide flat smirk, tiny nostrils, dark cheek spots; its thick gold chain and single '$' tag remain visible at all times. Never show the toad from behind. In the fictional night square the hater in rust t-shirt releases the clear rigid coin box into the toad's hands; the toad immediately passes that same box into the doer's hands. As the box changes hands, the hater drops his completely blank chips bag to the ground. Three identities and their clothes remain fixed. | Trimmed one model frame and re-encoded to 120 frames at native size. The sampled frames do not confirm the blank chips bag dropping. |
| clips/D11.mp4 | Kling 2.6 Pro / kwaivgi/kling-v2.6, b13f36d030496dd78d2986ba8b2b22a44222b3f58c15fb63ef7d6b4aa3a53319 | Photoreal ordinary handheld phone footage, vertical one continuous real-time shot for 5 seconds with active natural movement throughout. Keep every person's face, hair, body and clothes fixed exactly as the start image. Keep all props rigid. No cuts, no slow motion, no camera zoom or pan, no audio, no words, letters, numbers, logos, branding or watermarks. The toad mascot is the exact same giant flat green plush head with heavy half-closed lids, cream eyes, wide flat smirk, tiny nostrils, dark cheek spots; its thick gold chain and single '$' tag remain visible at all times. Never show the toad from behind. The front-facing toad throws a handful of plain rigid gold coins into the metal bowl of real live fire. The flames dance naturally. Then the toad reaches his small blank rigid gold trophy cup to the doer, who accepts it. The toad's chain and '$' tag remain visible. | Trimmed one model frame and re-encoded to 120 frames at native size. The toad rotates to a side/back view by the end and the trophy handoff is unclear. |
| track/track.wav | Google Lyria 3, d21615d0cd8b418e80268533307745e90044b4ba15df423fdb489be6bc594a88 | Original instrumental hype trap / phonk music, 120 BPM, four four meter, 30 seconds. Bass heavy punchy drums and crisp hi hats. Quiet build then strong drop at 2.5 seconds. Rhythmic accents at 5.5, 7.5, 9.5, 12 seconds. Tense half time segment from 14 to 18 seconds. Hardest percussive gavel hit at 18 seconds. Coin-like metallic ticks at 20 and 23 seconds, fire-like burst at 25.5 seconds, final emphatic hit at 28.5 seconds. No vocals, words, lyrics, or vocal chops. No recognizable existing melody. | Original model MP3 was trimmed to 30.0 s, lowered by 1.7 dB, faded over the final 0.5 s, and converted to 48 kHz 16-bit stereo WAV. Final hit near 28.5 s is weak. |
| track/beats.json | Measured from track.wav by 5 ms RMS-onset analysis and 0.5 s autocorrelation | Find strongest onset within ±0.1 s of each requested cut; emit the 120 BPM beat grid and bar downbeats. | Measurement file; no generation model. The downbeat phase and accent semantics need human listening review. |
| contact-sheet.png | Pillow 10.2.0, ffmpeg 6.1.1 | Assemble cast reference stills and frames 0, 60, 119 from each delivered MP4, with labels only on this sheet. | Rebuilt after final video transcodes. |
| README.md | Manual production record | Describe sources, prompts, checks, and limitations. | This file. |

## Music timing measured from track.wav

A 5 ms energy-onset scan found a 0.500 s autocorrelation period, corresponding to 120 BPM. beats.json records the nominal half-second beat grid starting at zero and four-beat downbeats every 2.0 s. The nearest detected onset within 0.1 s of each requested cut was:

| Requested time | Measured onset | Offset | Type |
| ---: | ---: | ---: | --- |
| 2.500 | 2.460 | -0.040 | drop |
| 5.500 | 5.595 | +0.095 | hit |
| 7.500 | 7.545 | +0.045 | hit |
| 9.500 | 9.575 | +0.075 | hit |
| 12.000 | 11.960 | -0.040 | hit |
| 14.000 | 14.090 | +0.090 | tension hit |
| 18.000 | 18.045 | +0.045 | gavel |
| 20.000 | 19.960 | -0.040 | coin |
| 23.000 | 23.030 | +0.030 | coin |
| 25.500 | 25.550 | +0.050 | fire |
| 28.500 | 28.440 | -0.060 | final, weak |

Onset timing is measured; labels such as "gavel" describe the requested role, not proof that the model produced a literal sound effect. The 28.5 s onset is weak and may not serve as an emphatic final hit.

## Rejection-list review

1. Letters, numbers, logos: **fail**. Visible generated lettering appears in D05a and D07; D05d has a shirt emblem. Other frames were sampled at start, middle and end, so a full-frame guarantee is unavailable.
2. Green surfaces: **fail**. D01 and D05d gain UI marks; D05a gains lettering; D05b gains a pale reflected area; some D02 passerby phones are not flat green. Pixel-perfect #00FF00 is not certified across full motion.
3. Morphing identities, hands or props: **not fully certified** by the three-frame review. Main cast remains recognizable in sampled frames; complex handoffs may contain artifacts.
4. Near-still, slideshow or code animation: **pass on production method**. Every clip came from the Kling video generation model and shows motion in sampled frames. No still pan, slide, or code animation was used.
5. Wrong toad head: **fail in D11 end frame**. The referenced plush head, half-closed eyes, chain and tag appear, but the toad rotates to side/back near the end. D09 and D10 remain front/three-quarter in sampled frames.
6. Format, upscale, watermark: **pass for encoded files**. All are H.264/yuv420p, 120 frames, 24 fps, 5.000 s, native model dimensions, no audio stream, and no visible watermark in sampled frames.
7. Brand look or famous face: **not fully certified**. Cast is generated as fictional ordinary adults. D05d's emblem is an unacceptable print and no brand clearance is claimed.
8. Track voice, code generation, duration: **pass for model provenance and duration**. The track was generated by Lyria 3, not composed in code, and is exactly 30.000 s. Model logs say instrumental; a listening review is still needed.

A planned regeneration for D08 using Kling v3 failed with **HTTP 402 Payment Required: Insufficient credit**. Further model regeneration was stopped, leaving the visual issues above for review.