![\"45fc8e69jnnr0s00j3d000nbp\"](\"https:\/\/www.1ai.net\/wp-content\/uploads\/2026\/10\/45fc8e69j00tnnr0s00j3d000ki00nbp.jpg\" "\\"45fc8e69jnnr0s00j3d000nbp\\"")
<\/p>\nAn ordinary video, and perhaps a story for short films:<\/p>\n
The staff members entered the back cook with containers, dropped them unexpectedly and then panicked; a man stood in front of the door looking at his mobile phone and suddenly turned around and followed the dog behind him\u3002<\/p>\n
These are not complex, but they have actions, changes and reasons to keep them alive\u3002<\/p>\n
The problem is, it's hard to say what we really want to change by writing \"to make it more filmy.\" This section, from material organization, story-breaking, mirror design to generation, leads you to vague aesthetic requirements, and turns them into operative hints. Even without a photographic basis, we can start with a clear little story\u3002<\/p>\n
<\/p>\n
Correlation<\/p>\n
First, to make sure that the story remains the same and the method of filming can change<\/strong><\/p>\n Before you start writing the hints, you have to distinguish between two things: what happened to the video and how the camera records it. The former is a story, the latter is a picture. When recreated, the person, scene, props and event sequence can be preserved, and the slot, scene, lens and rhythm adjusted\u3002<\/p>\n In the case of a post-cooking accident, the story is that a staff member enters a container with red sauce and is ready to put it on the operating table, the container is unbalanced, the sauce is spilled, he looks down and he panics. This causal link must be retained. As for filming from the door, from the side of the table, or cutting to the side of the shoe to observe the fall, it could be rescheduled\u3002<\/p>\n Reference Video<\/p>\n The so-called sense of film needs to be put to the attention of the audience: first, to let people know where the role is, then to draw attention to the key items, then to the accident and finally to the reaction of the person. A good image is just one part of it, and the order in which information appears is equally important\u3002<\/p>\n First, we write about a creative goal: \u201cRetention of events and characters, rearrange the camera so that the audience can clearly see actions, accidents and reactions.\u201d This sentence helps you to judge whether the next hint is off the subject\u3002<\/p>\n II. Three extracts: characters, scenes and props<\/strong><\/p>\n The purpose of material extraction is to prepare a set of reusable visual references. The character chart answers \u201cWho\u201d, the scene map answers \u201cWhere\u201d, and the prop chart answers \u201cWhat to hold\u201d. In practice, these three types of material need to be treated separately: first, the corresponding screenshot is selected, then the extraction request is submitted separately, and only one type of material is processed each time. Once all checks have been completed, they will be used for subsequent spectroscopy and video generation\u3002<\/p>\n First step: individual extraction\u3002<\/p>\n The figures in the original video may be small, vague or obscured, and a single screenshot can easily miss the appearance. Thus, the character extraction is divided into two stages: an analysis of the whole video and the production of reference maps in conjunction with the results of the analysis\u3002<\/p>\n LET'S UPLOAD THE FULL VIDEO AND GET AI TO SORT OUT THE CHARACTER\u3002<\/p>\n Identify the person that needs to be analysed, and allow AI to organize a stable appearance in a picture of him at different times, angles and actions. You can send the following command (this is a language model that can analyze videos, such as ChatGPT, Gemini):<\/p>\n Please analyze the complete video that I uploaded, combine the images of the target person at different times, angles and actions, and organize a list of the visual features used to generate the figure reference\u3002 Then hand over the feature table and video screenshot to the image generation tool (this can be generated using GPT image 2.0)\u3002<\/p>\n iamge2.0 in ChatGPT<\/p>\n Based on the points of time listed in the analysis, relatively clear images of the face, body and other angles are taken. The following instructions are used to create the person(s):<\/p>\n Please create a clear character reference board for the [target person] with a pure white background, with the following four independent views, in conjunction with the video character profile and the screenshot I have provided: After generation, the face, hair, body and clothing are checked for conformity with the original person and the details of the completion are checked. Where cross-references or back references are required, the corresponding screenshot is created separately and the identified character set is followed. The entire video was not visible to the five officials and could still only be confirmed as a production design\u3002<\/p>\n Step 2: Take the scene alone\u3002<\/p>\n Re-select a screenshot that can see the environment layout and submit the scene extraction request separately\u3002<\/p>\n Videoshot<\/p>\n The location of the main environmental elements in the original map should be maintained, the person should be removed and a space scenario reference map should be available for repositioning, whether it be indoor space, streets, natural landscapes or fantasy scenarios\u3002<\/p>\n select the image generation, using the feedream 5.0pro model<\/p>\n this is how the scene extracts (seedream 5.0pro-generated):<\/p>\n Please use the image I uploaded as the only visual basis to extract the scene and generate a complete picture of the scene without a character. Automatically identify the scene type of the original map and keep the actually visible buildings, topography, vegetation, furniture, facilities and environmental objects; elements not found in the original map are not added. Maintains the same style of original structure, shot angle, perspective, space layout, object position, ratio, material, colour, light and image. To remove all persons and their clothing, handheld or carried with them, and to keep the displays and objects that belong to the scene itself. The area covered by the person and his\/her articles is naturally completed in accordance with the visible structure, texture and light surrounding it, and is reasonably connected and does not add buildings, entrances or other facilities to the space. Remove superstitious subtitles, interfaces and borders, and keep the actual internal identity of the scene. Do not re-engineer the environment, do not change the type of scene and do not delete the original set for the sake of cleanliness. Only the scenes are exported, and displays of characters or props are not spelled\u3002<\/p>\n After this is done, the scene is adjusted to the target movie style\u3002<\/p>\n the space scenario is first used to confirm the layout and then stylish it. hand it over to the image generation tool to illustrate the type of film you want; it can be accompanied by style reference maps, or it can describe the light, tone and atmosphere in text. replace the [target style] in the template below with its own requirements, such as \u201cnatural sunlight, soft shadows, low saturated and warm colours, small film-writing styles\u201d (seedream 5.0pro can be used to generate pictures)\u3002<\/p>\n Please base the structure on my uploading of the blank scene, and re-engineer it as [target style] and generate a film style reference for video production. If style reference maps are provided, only the light, colour, material performance and atmosphere in them will not be reproduced in their construction or layout. The location and proportion of the original scene type, structure, vision, spatial dimensions, main structures, corridors and objects are maintained. Reorienting the light to the soft, dark, dark, colour and material senses in target style to give a clear spatial dimension to the outlook, medium, background, and to keep visible details of the critical action areas. Film senses do not amount to overall dimness, and do not uniformly add fog, beams or exaggerates; effects should be consistent with the scene and target style. Do not move or add or delete critical facilities, do not change the space necessary for a person to walk and operate, and do not add new characters, features or words. Only one full film-style empty scene reference map is exported\u3002<\/p>\n Step 3: Separate pick-up of props\u3002<\/p>\n A clear and visible cut-off of the key props is selected and the container is extracted separately. Focus on the shape, material, colour, size, open state and content. Large packagings and snacks require different techniques and cannot be readily replaced at extraction\u3002<\/p>\n the prop extract is written as follows (seedream 5.0pro generation):<\/p>\n Based on the images that I uploaded, please single-handedly extract the \"target props\" and make a reference panel for multiple viewing props with a pure white background, each showing the same props\u3002 After three extractions have been completed, the person is checked for consistency, the layout of the scene is correct, the props are accurate and their numbers are kept. When the video is subsequently generated, these separately extracted reference maps can be provided together and the purpose of each can be specified\u3002<\/p>\n iii. Write the incident chain, then the graphic adjective<\/strong><\/p>\n With reference material, the next step is to organize the action into an event chain. Not to add the word \u201cshocking\u201d to the word \u201chigh\u201d, but to write what happened and what was the connection\u3002<\/p>\n The post-cooking case can be organized as \"Entering with two boxes of containers -- near the operating table -- the top packaging is unbalanced at the time of placement -- the sauce in the top container is spilling over the desktop -- the person is preparing to sort out the spilling container -- At the same time, the lower packaging fell on the ground because it was on the edge of the desktop - it was ready to go down to pick up the sauce while it bounced through the ground and reached the person's face - and the person was in a state of panic\u201d\u3002<\/p>\n Here's the difference between observation and speculation. The fact that the person was seen turning around did not prove that he was avoiding danger\u3002<\/p>\n Each key action requires a starting point and results. The \u201ccontainer tilt\u201d is incomplete and explains how it was held up, where it was imbalanced and where the sauce ended up. \u201cPersonal panic\u201d also translates into visible actions, such as pauses, heads down, shrunk shoulders or backsliding, rather than allowing the role to remain exaggerated\u3002<\/p>\n This is how the event chain asks for a large-language model like AI (ChatGPT\/gemini):<\/p>\n The visible actions in the video (story chain) are organized in chronological order, and each action begins, changes and results. Distinguishing between events that have taken place and motives that cannot be identified. The contact between the person and the props is maintained, the chain of events is first exported, and no mirrors, filters or new scenes are added for the time being\u3002<\/p>\n Upload reference video<\/p>\n This step avoids producing a beautiful picture of the results, but instead turns the word \u201cunfortunately\u201d into the word \u201cinvolved\u201d or the word \u201cturning\u201d into the word \u201cstand-up\u201d\u3002<\/p>\n IV. LET AI RE-ENGINEER THE LENS ACCORDING TO THE SEQUENCE OF EVENTS<\/strong><\/p>\n WHAT IS AVAILABLE IN THE PREVIOUS SECTION IS A SEQUENCE OF EVENTS, INDICATING \u201cWHAT HAPPENED\u201d. THIS SECTION GIVES IT TO AI, REQUESTING A REARRANGEMENT OF \u201cWHERE AND WHAT EACH PART IS FILMED\u201d. WHEN WE'RE DONE, WE'LL GET A SPECTROSCOPE FOR THE NEXT SECTION TO ORGANIZE THE VIDEO MESSAGE\u3002<\/p>\n In the same conversation that just analysed the video, the following instruction was sent:<\/p>\n Please re-engineer a film-sensitive spectroscopy for this paragraph, according to the sequence of events just organized. The sequence of persons, scenes, key props and events is preserved and no new dramas are added. Rearrange the angles of the shots, the scope of the images and the required lens movements, without following the fixed view of the original video from head to end. In the first place, the audience will see the person entering, the key movement, the accident and the reaction. Each shot focuses on one major action, which can be followed by back and back. Please use a table to list the lens number, the time period, the content of the picture and the way it is taken, and first to output the spectrometer\u3002<\/p>\n This is a redesigned lens. It retains the same event and presents the environment, the props, the failure process and the reaction of people from different angles. Time is an example of teaching and can be adjusted to the actual action rhythm\u3002<\/p>\n ONCE THEY HAVE THE SPECTROSCOPE, CHECK IF KEY ACTIONS HAVE BEEN OMITTED, THE STORY CHANGED OR THE SCENE NOT FOUND IN THE MATERIAL. IF A SHOT IS NOT APPROPRIATE, INDICATE THE CAMERA NUMBER AND ALLOW AI TO MODIFY THIS LINE. UPON CONFIRMATION, THE SPECTROSCOPE IS USED FOR THE NEXT STEP\u3002<\/p>\n V. HAND OVER THE MIRROR AND THE MATERIAL TO AI AND ORGANIZE THE VIDEO ALERTS<\/strong><\/p>\n REFERENCE MAPS OF INDIVIDUAL FIGURES, SCENES AND PROPS TAKEN SEPARATELY FROM SECTION II ARE NOW HANDED OVER TO AI, ALONG WITH THE MIRROR CHART IDENTIFIED IN SECTION IV, TO ENABLE IT TO BE ORGANIZED AS A REMINDER FOR VIDEO PRODUCTION\u3002<\/p>\n When uploading reference maps, the \u201cpersons\u201d \u201csituation\u201d \u201cprotocols\u201d are simply marked, followed by an identified lens. This step follows the previous stories and lenses, and they are organized into directly usable texts\u3002<\/p>\n The following instructions can be sent directly:<\/p>\n This is a reference map of the person, scene and props taken, and a reference video, which is used to refer to the sequence of events, the movement of the person, the movement route and the interaction of the person and props. , in conjunction with the confirmed spectroscopy, make a complete reminder for video generation. The appearance of the person is based on a figure map, the environmental layout is based on a scene map and the key items are based on a propographic chart. In the time order of the spectroscopy, the movement of the person, the angle of the photograph and the change in the lens are written, the event is kept and no new drama is added. Follow the new spectrometry and remove the old space requirements that conflict with it. Output is the final hint that you can copy directly\u3002<\/p>\n Keep asking about the last chat<\/p>\n Once the results have been obtained, three things are checked: whether the person and the street are correctly written, whether the sequence of the action has changed, and whether the lens is aligned to the spectrometry\u3002<\/p>\n If you find a problem, fix it here and get into video production\u3002<\/p>\n VI. Input of original video and reference maps to generate the first version of the video<\/strong><\/p>\n to generate videos, upload original videos, character reference maps, confirmed film style maps and proprietors, and paste the previously organized video alerts. checks whether the video, image identification in the hint corresponds to the actual input, and then sets the proportion and length of the image to start generating\u3002<\/p>\n The original video supplements the action information that static images do not provide in full, such as how people hold their props, where they lose balance and how to retreat\u3002<\/p>\n Video alerts are as follows:<\/p>\n Please generate a continuous film-sensitization video of about 9 seconds and 4 shots of characters, scene-situations, props and original videos\u3002 Once the first edition has been produced, look at it from the beginning and see if the story is complete and check for key actions. Future cooking stories, for example, focus on whether the person enters with the right container, whether he or she accidentally fails when placed, and whether he or she reacts when he or she falls behind. The content should be confirmed and the colour and light should be adjusted\u3002<\/p>\n When you complete this process, you get a short film that keeps the event and rearranges the shoot. From the individual extraction of material to the design of the spectroscopy, to the preparation of the hints, to the generation and modification, each step has clear results and materials for the next step\u3002<\/p>","protected":false},"excerpt":{"rendered":" An ordinary video may also contain stories that are suitable for short films: staff members carry containers into the back of the kitchen, drop them down unexpectedly and then panic and look; a person stands in front of the door, looking at his cell phone, and suddenly turns around, and the dog behind them follows. These are not complex, but they have actions, changes and reasons to keep them alive. The problem is, it's hard to say what we really want to change by writing \"to make it more filmy.\" This section, from material organization, story-breaking, mirror design to production, leads you to vague aesthetic requirements, and turns them into operative hints. Even without a photographic basis, we can start with a clear little story. One, first, make sure that the story stays the same and the camera changes<\/p>","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":""},"categories":[149,144],"tags":[802,3149,956,5321,192,6481],"collection":[],"class_list":["post-59490","post","type-post","status-publish","format-standard","hentry","category-jiaocheng","category-baike","tag-ai","tag-prompt","tag-6481"],"acf":[],"_links":{"self":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/59490","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/comments?post=59490"}],"version-history":[{"count":0,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/59490\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/media?parent=59490"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/categories?post=59490"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/tags?post=59490"},{"taxonomy":"collection","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/collection?post=59490"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}![\"54eb1af2j00tnnr1s008gd000kj00bkp\"](\"https:\/\/www.1ai.net\/wp-content\/uploads\/2026\/10\/54eb1af2j00tnnr1s008gd000kj00bkp.jpg\" "\\"54eb1af2j00tnnr1s008gd000kj00bkp\\"")
<\/p>\n
\nThe following are analysed separately:
\n1. Overall appearance: broad age perception, visible colour, hair, hat, beard and other identifying characteristics. 2. **Face and five officials:** Face profile, eyebrow, eyes, nose, lips, jaws and identifiable facial details. 3. ** Body size and body size:** Fat weight, shoulder width, torso and limb form, and overall body size ratio. 4. ** Clothing and accessories:** Clothing style, colour, pattern, pants, shoes and accessories\u3002
\nA comprehensive cross-check is made to distinguish between the stable characteristics of the person itself and the changes caused by light, perceiving and blurring the movement. Each indication is \u201cconfirmable\u201d \u201cgenerally legible\u201d or \u201cunconfirmable\u201d. Details that cannot be confirmed are not made up and do not imply precise height or weight\u3002
\nFinally, the time points suitable for the screenshot are listed and are used for face, body and other visible angles\u3002
\nThis step is only an output of the character profile analysis and cross-graph recommendations and does not produce a figure image\u3002<\/p>\n![\"ed9f746cj00tnnr2c001ad000mi00fwp\"](\"https:\/\/www.1ai.net\/wp-content\/uploads\/2026\/10\/ed9f746cj00tnnr2c001ad000mi00fwp.jpg\" "\\"ed9f746cj00tnnr2c001ad000mi00fwp\\"")
<\/p>\n![\"5a537b10j00tnnr37002jd000m7000esp\"](\"https:\/\/www.1ai.net\/wp-content\/uploads\/2026\/10\/5a537b10j00tnnr37002jd000m700esp.jpg\" "\\"5a537b10j00tnnr37002jd000m7000esp\\"")
<\/p>\n
\n1. A 3\/4-side human face view: approximately 45 degrees to the side of the head, visible with both eyes, with a complete display of the head, hair or hat and the neck shoulder, with a clear profile of the face, identifying the five officials and skin. 2. Frontal body chart: The person is facing the lens, naturally standing up and presenting the whole face, clothing and shoes. 3. The body of the side: the body and face of the person are 90 degrees and the profile of the head, the thickness of the body, the body size and the side structure of the clothing are clearly shown. 4. Back-to-back body chart: back-to-back lens with complete representation of the back hair, back structure of the clothing and the back side of the shoe\u3002
\nBased on the visible appearance in the screenshot, the information identified from other video images is supplemented by a feature sheet. The four views must maintain the same person ' s face characteristics, hair, colour, body size, body size ratio, dress colours and accessories. No beautification, no change of weight, no stretching\u3002
\nThe three whole-body view uses a uniform scale and natural straight-on attitude, relaxes both arms and leaves a little space for the torso, and presents a complete picture of the head to the sole of the shoe, which is aligned with the bottom baseline. Faces are magnified in isolation, organized in separate groups with three body-views, leaving blanks between the views, which do not overlap or vary\u3002
\nTo preserve the true quality of skin and clothing, using the same soft and light light. Remove original scenes and hand-held props, and keep the clothes and accessories that people wear\u3002
\nFor details marked as \u201cgenerally legible\u201d or \u201cunconfirmable\u201d in the feature table, as well as for the missing side, back side, based on the visible structure, which is naturally and restrainly completed, no new decoration, pattern or exaggerated identification feature. When there is a conflict between the text and the screenshot, the conflict is identified and a check is requested\u3002
\nMake the picture clearer, but do not consider the completion of details as a confirmed real look\u3002<\/p>\n![\"ead84ef4j00tnnr4000kd000v9000hmp\"](\"https:\/\/www.1ai.net\/wp-content\/uploads\/2026\/10\/ead84ef4j00tnnr4000kdd000v900hmp.jpg\" "\\"ead84ef4j00tnnr4000kd000v9000hmp\\"")
<\/p>\n![\"679c0d95j00tnnr4708fd000hc00a0p\"](\"https:\/\/www.1ai.net\/wp-content\/uploads\/2026\/10\/679c0d95j00tnnr47008fd000hc00a0p.jpg\" "\\"679c0d95j00tnnr4708fd000hc00a0p\\"")
<\/p>\n![\"dd1c24f4j00tnnr4b0037d000v9006qp\"](\"https:\/\/www.1ai.net\/wp-content\/uploads\/2026\/10\/dd1c24f4j00tnnr4b0037d000v9006qp.jpg\" "\\"dd1c24f4j00tnnr4b0037d000v9006qp\\"")
<\/p>\n![\"ced91ff6j00tnnr4o010td000v90hp\"](\"https:\/\/www.1ai.net\/wp-content\/uploads\/2026\/10\/ced91ff6j00tnnr4o010td000v900hhp.jpg\" "\\"ced91ff6j00tnnr4o010td000v90hp\\"")
<\/p>\n![\"7147439bj00tnnr4x0127d000v9000hfp\"](\"https:\/\/www.1ai.net\/wp-content\/uploads\/2026\/10\/7147439bj00tnnr4x0127d000v900hfp.jpg\" "\\"7147439bj00tnnr4x0127d000v9000hfp\\"")
<\/p>\n![\"8cad9df7j00tnnr511004ld000v9098p\"](\"https:\/\/www.1ai.net\/wp-content\/uploads\/2026\/10\/8cad9df7j00tnnr51004ld000v90098p.jpg\" "\\"8cad9df7j00tnnr511004ld000v9098p\\"")
<\/p>\n
\nThe reference board contains the following:
\n1. Positive view: A complete display of the exterior profile of the props, the positive structure and the main components. Side view: A complete display of the thickness, side shape and widget connection of the props. Top view: Show top shape, opening or lid structure from above. 4. Entirely written: The main theme of the road is clearly presented in terms of appearance, material and colour. 5. Local detail: Depending on the actual object, multiple details can be arranged by displaying material texture, edges, seams, interfaces, handles, buttons or other identifying details. 6. Internal detailed view: displays the inner walls, liners, partitions, openings or contents that can be identified in the reference diagram, highlighting internal materials and structures\u3002
\nAll views shall be consistent with the same shape, material, colour, texture, proportion of the parts and their existing markings, and shall not be replaced or redesigned. Maintains a uniform ratio between the complete view, with a close-up only for the corresponding part, without changing the structure\u3002
\nRemoval of persons, hands, background and unrelated items and retention of parts and accessories of the props. Use a pure white background, a soft light and a slight shadow, with view partitions organized in order to leave space, without overlap and indistinguishment\u3002
\nThe perspective not presented in the original map can only be carefully completed on the basis of visible structures. When the internal structure cannot be confirmed, it is noted that additional reference is required, that no internal parts are made up, that closed items are not excreted and that no internal cavities are added to items without internal cavities\u3002
\nDo not add text, identification or functional widgets not found in the original map\u3002<\/p>\n![\"b28bf4baj00tnnr5800j4d000v90hjp\"](\"https:\/\/www.1ai.net\/wp-content\/uploads\/2026\/10\/b28bf4baj00tnnr5800j4d000v900hjp.jpg\" "\\"b28bf4baj00tnnr5800j4d000v90hjp\\"")
<\/p>\n![\"2b54b7eaj00tnnr5f001fd000n100c1p\"](\"https:\/\/www.1ai.net\/wp-content\/uploads\/2026\/10\/2b54b7eaj00tnnr5f001fd000n100c1p.jpg\" "\\"2b54b7eaj00tnnr5f001fd000n100c1p\\"")
<\/p>\n![\"f244693j00tnnr5n00jhd000sg00tip\"](\"https:\/\/www.1ai.net\/wp-content\/uploads\/2026\/10\/fe244693j00tnnr5n00jhd000sg00tip.jpg\" "\\"f244693j00tnnr5n00jhd000sg00tip\\"")
<\/p>\n![\"ddb2dcf3j00tnnr5z016pd000m201a0p\"](\"https:\/\/www.1ai.net\/wp-content\/uploads\/2026\/10\/ddb2dcf3j00tnnr5z016pd000m201a0p.jpg\" "\\"ddb2dcf3j00tnnr5z016pd000m201a0p\\"")
<\/p>\n![\"e236796ej00tnnr6800e6d000v9000gqp\"](\"https:\/\/www.1ai.net\/wp-content\/uploads\/2026\/10\/e236796ej00tnnr6800e6d000v900gqp.jpg\" "\\"e236796ej00tnnr6800e6d000v9000gqp\\"")
<\/p>\n
\n(a) Reference priority: Image 2 of the person The appearance of the person is strictly based on the person ' s reference map; The original video 1 is used only to refer to the sequence of events, the movement of persons, the movement route, the contact relationship between the person and the two containers and the course of the accident, and does not follow the high-level stationary monitoring, graphics and clippings of the original video\u3002
\nThe characters remain the same: black baseball caps, black short-sleeve Polo shirts, blue collages on both sides, black aprons, black shorts, black socks and black shoes, and the same body, face and clothing ratios as the figure's reference maps. Both containers are always the same set of transparent plastic sauce containers, which contain red-brown thick sauce. The accident must be performed in accordance with the original video real event to show that the sauce is capable of spilling from the oscillating container opening and that the lid structure in the reference map does not prevent the spilling\u3002
\n00:00-00:02, camera 1: Pouch entry
\nThe opening is based on two piled-up sauce containers in the person ' s arms. On the side, about 45 degrees of close-range close-cut close-up, taken only to the person ' s chest, arms and hands, without a full view. The person walked from the sceneal entrance to the operating table with two containers, with both arms holding them continuously, and the two containers had a slight real shake with their pace, but still maintained a super-down. The camera moved slowly and slowly with the person, with a shallow view and a focus always locked on the hands and two containers. As the person approached the operating table, the stainless steel desktop gradually entered the picture and the hands began to move the two containers forward and down\u3002
\n00:02-0:04.3 |: Placing imbalance and sauce spill
\nCuts to a close view in the front of the side at the height of the operating table and can see both the top half of the person, his hands, two upper and lower containers and the desktop. The person normally tried to place the two containers on stainless steel operating tables, not by active dumping. During the placement, the upper packaging suddenly lost its balance, leaning to one side and leaning to the side, the internal red-brown thick sauce flowed naturally from the opening of the container, spilling down and spreading on stainless steel desktops. The actions follow the real weight, inertia and liquid movement, and clearly demonstrate the continuous cause and effect of \"normally placing the unplanned imbalance in upper packagings and spilling the sauce.\" The person immediately leaned forward and his hands were stretched towards the upper container where the dumping took place for disposal. The lens remains stable, with only very minor natural propulsion, and no extra lenses\u3002
\n00:04.3-0:06.5 |3: Second container falling
\nCuts to the middle view behind the side of the person, and the image contains both the person ' s hands, the desktop area and another container near the edge of the operating table. The person ' s attention remained focused on the upper container, which had just been dumped, and the body had leaned forward and adjusted it with its hands; at that point another container was located near the edge of the table and was not kept by the person. The container gradually crossed the table line, lost desktop support and then fell naturally to the ground. Do not act as a person who pushes or deliberately falls. When the container starts to fall, the camera moves in a short, restrained, down-sliding direction, following the direction of the fall; the person then bows his head and leans over to the falling area, and the action is linked to the mirror\u3002
\n00:06.5-0:09 Zero 4: Back-face and Back
\nIt cuts into a near-ground low-level view and is slightly slapping from the front side of the falling area. Below the picture, the ground can be seen and the container dropped, above the person who is on his knees. The person heads down, the upper half moves forward, and the face moves closer to the ground accident area. In the process of the person's fall, the sauce resulting from the fall suddenly rebounded upwards and splattered in the vicinity of the ground, and the red brown sauce went down to the person's face. It's a continuous process of \"persons leaning down, and the ground sauce bounced up, and the sauce hit the face,\" and it can't be changed to people taking the sauce to their face. Once the sauce is dried up, the person is able to reflect on the upper half, with his hands close to his face, and move back in a continuous and fast manner, moving away from the operating table and the falling area. The camera moved slightly and left the person with a retreat space to keep the movement intact until its end. The entire piece shall be kept in the same person, in the same kitchen space, in the same operating table and in the same group of containers, and the camera shall be replaced by the person in the direction, position, position of the container, position of the desktop sauce and position of the surface shall be continuous. The movement is real, weighty, and the liquid has real viscosity, inertia, spilling and impact rebound. The whole is a real film-photographic sense, natural indoor light, restraining depth and stable photography, without exaggerating slow motion, no comedy-style deformation, no additional characters or additional accidents, no new drama and no change in the sequence of events\u3002<\/p>\n![\"1676d011jtnnr6m009td000km00b9p\"](\"https:\/\/www.1ai.net\/wp-content\/uploads\/2026\/10\/1676d011j00tnnr6m009td000km00b9p.jpg\" "\\"1676d011jtnnr6m009td000km00b9p\\"")
<\/p>\n