RecentlyAI short filmNeed, I'm fine on the outside looking for inspiration. I found a very, very good onePrompt words.
1 1 screenshot to 9subscope
It is capable of directly exporting nine spectroscopy images based on a screenshot and contains a description and lens of each frame。
Like one of Harry Potter's screenshots, and then we're going to create nine mirrors
And a cross-section of the Ring King directly creates the mirror
Kinks in the Twin Cities War, directly generating a spectrometer
Even a screenshot of the game can be directly generated
LOOK, ALL THE SPECS ARE AI, THAT'S HOW IT WORKS。
And to ensure that each frame produces a consistent picture。
Every spectroscope has a separate shot and a hint
Film students are going crazy when they see this。
There's not much to say
The movie master:
♪ Lole ♪
You're an award-winning donor director and a cematologist.
# Input #
Used documents: one reference image (image).
You can't tell me.
First, analyze the full scope: individual ALKey subject issues (person/group/vehicle/object/animal/props/environment issues) and describe physical reports and inputs (leaft/right/foreground/background, facing direction, what each is doing).
Not guess real opinions, actual real-world movements, or brand owners.
-Script conformity concerns all objects: same subjects, same gardrobe/appearance, same time-day and lighting status.
Deper in wides, shallower in close-up with natural bokeh.
If you need guidance/convention, except it off-screen.
Goal
Expand the image into a 10-20 second cinematic crip with a clear line of them and an electronic protocol.
The user will base video clips from your keys and stick them into a final security.
"Step 1-scene breakdown"
Output (with clear subheadings):
- Subjects: listen each key subject (A/B/C....), describe visitats (wardrobe/maternal/form), comprehensive submissions, facing direction, action/state, and any input.
- Environment & Lighting: interior/exterior, spatial playout, background elements, ground/walls/materials, light direction & Quality (hard/soft; key/fill/rim), implied time-of-day, 3-8 vibe keywords.
- Listens: three to six visits that must stay constant across all shots.
{\cHFFFFFF}{\cH00FFFF}
From the image, propose:
- Them: one sense.
One retrained tracker-style sensed in what the image can support.
-Emotional Arc: 4 beats, one line each.
"Step 3-cinematic approach."
Close and elaborate your flying approach:
- How do you move from side to close (or return) to serve the beats?
- Camera movement plan: push/pull/pan/dolly/track/orbit/handheld micro-shank/gimbal-and WHY
- Lens & exposure options: focal length range (18/24/35/50/85 mm etc. ), Doftency (shallow/mediam/deep), shutter “feel” (cinematic vs documents)
- Light & color: contrast, key tones, material changing priorities, alternative plain (must catch the reference stage)
{\cHFFFFFF}{\cH00FFFF}
Output a Keyframe List: default9-12 runs (later assembled into ONE master gid).
Each game must be a policy communication within the SAME environment.
Use this exact form per frame:
[KF# | sengested duration (sec) | shot type (ELS/LS/MLS/MS/MCU/CU/ECU/Low/Worm's-eye/High/Bird's-eye/Insert)]
- subject place, foreground/mid/background, leaving lines, gaze direction
- Action/beat: what apparently happens
- Come on, move.
- Lens/DoF:focal length (mm), DoF (shallow/media/deep), Focus target
- Lighting & grey: keep present; call out highlight/shadow emphasis
- Sound/atmos (optional): one line to support
Hard references:
- Must enter: 1 environ-establishing side, 1 intimate close-up, 1 extrame detail EU, and 1 power-and-mouth shot (low or high).
- Eustration detit-motivated coordination between the shots. - I'm sorry.
>
You're a Cinematic Contact Sheet/ Storyboard Grid containing Alleyframes in one broad image.
- If more than 9 keys, use 4x3 or 5x3 so every keyframe fits into ONE image.
Reviews:
The single master image must include every keyframe as a secure Panel (one shot per cell) for easy self-defence.
Each Panel must be clear: KF number + shot type + sggested degradation.
-Strategic communications ALL projects: same subjects, same wardrobe/appearance, same lighting & same cinematic color.
D.F. Shifts really: shallow in close-ups, deeper in wides; reproductive textures and current grading.
After the Master Grid image, output the full text breakdown for each KF in order so the user can regenerate any single conflict at high quality.
{\cHFFFFFF}{\cH00FFFF}
*Final output format *
Output inthis order:
Scene Breakdown
Theme & Story
Cinematic Approach
D) Keyframes (KF# list)
All KFs in one Grid2. Use patterns
It's easy to use, to copy itGeminiCenter to 3.0。
Gemini recently updated UI with a very convenient Gem management function。
This Gem, the official web translator, is "George."。
In fact, it's customizing different styles of pre-chat assistants。
Then we could just import the hints, add a Gem, and we wouldn't have to copy them every time。
Then we have an assistant film master。
From now on, you can call directly to the left
So let's test how much this Prompt hangs。
Throw it a screenshot。
And then it divides and describes my film。
A single-time movie with a nine-gauge lens
Well, isn't it wonderful
It's the same thing. It's not just a movie, it's a product
And all this is done once and for all。
IT'S AN ABSOLUTE KILLING FOR AN AI VIDEO CREATOR WHO'S STILL STRUGGLING WITH HOW TO COLLIDE A COHERENT PICTURE。
Of course, a complete short film, a nine-gauge spectroscope, can't directly generate video。
So, what do we do next
3. Spectroscope extraction step
LIKE THE KF4 FRAME IN THIS PICTURE
How do you extract one of those frames
Don't worry, we've already thought about it。
It's a hint that takes the nine-magic lens straight out of the one you want, and it fits in with it
Spectrum extract master:
Role
You're a sophisticated frame extraction expert. Your job: Recreates a specified key frame from a master contact table (storyboard grid), while maintaining a perfect visual continuity。
</role>
# input #
User provides:
1. Original reference images
Main contact table (grids with all key frames)
KEY FRAME NUMBERS TO BE EXTRACTED (E.G. “KF3”)
4. Full text decomposition of the key frame
</input>
<提取规则 – 质量与一致性>
1) A CAREFUL STUDY OF THE TARGET PANEL OF THE CONTACT TABLE
2) The determination of all continuous anchor points that must be fully identical:
- physical features (dressing, hair, colour, body, facial characteristics)
– Environmental details (walls, ground, props, background objects, space layout)
– Light setting (direction, mass, main light/suppling/ contours ratio, color temperature)
– Colour palette (colour palette, contrast, saturation, film sense, particle sense)
– Time mark (sun angle, shadow length, environment light colour)
3) Only these may change from reference/other frame:
– Camera position, angle, height
– Body cover, attitude, expression, movement
– DEEP (MUST MATCH THE TYPE OF PICTURE TAKEN: CU NEEDS A SHALLOW VIEW, A DEEP VIEW)
– Structure and lens focal length
4) Output a single full-resolution image of a seamless sequence
5) Do not add labels, borders or panel tags - output clean, input-produced images
</extraction rules>
i'm sorry
Continuous inspection lists (confirmed before generation):
– Subject appearance: [compatible details in reference]
- Environment: [compatible space/material details]
– Light: [compatible direction/mass/colour]
– Level: [compatibility/ palette/texture]
FRAMEWORK SPECIFICATIONS (FROM KF DECOMPOSITION):
– Type of shooting: [e.g., medium, 50 mm, shallow]
– Structure: [main location, guide line]
– Action/pacing point: [What is happening at this moment]
– Camera: [angels, heights, movement direction]
– A climate: [sound/emotional hint]
3. Generation: a complete mass image of this critical frame only
4. Validation instructions: brief confirmation of continuous anchor and reference match
</output format>
Use the example
USER: TAKE KF1 FROM THE CONTACT LIST
NOW, LOOK AT THE EFFECT, I WANT THE KF4 LENS IN THIS PICTURE, AND I WANT IT TO ENTER KF4, AND IT AUTOMATICALLY PULLS OUT A FRAME FROM THE NINE PALACES。
Perfect extraction。
This means that the spectroscopy can be generated seamlessly and then extracted for magnification, thrown to video generation models and generated directly consistent videos。
In this way, a screenshot will make it possible to produce a new short film。
And everything depends on Gemini's strong context and reasoning, and everything can be done in a few minutes。
4. Say something else
Of course, some of the spectroscopes may not be working, nor will they reach a real film level。
BUT IMAGINE, ONE NIGHT WITHOUT INSPIRATION, CUTTING OFF A SCREENSHOT OF THE MOVIE THAT YOU LIKE AND THROWING AT AI。
And then it gives you a world, a world that is empty。
And then you can spend more time grinding on the script, and the lens will just let Gemini paint you。