CLEARFRAME
Make a short science explainer with AI assistance

Original English sample | Version 0.1 | 6 October 2026
Status: Draft, untested workflow. No finished video has been produced or evaluated.
Clearframe is a provisional brand for a private site concept. This sample is independently written. No Chinese PDF has been supplied, used, translated, or adapted.

WHAT YOU WILL MAKE

A planned 26-second educational video answering one question: why does a card's shadow change size when the card moves between a small light and a screen?

This lesson takes you from an idea to a script, storyboard, visual assets, narration, edit, and export. AI can help you propose and revise. You remain responsible for the science, geometry, rights, and final review. The steps describe a proposed workflow, not measured production time or guaranteed model behavior.

Audience: Beginners making their first explanatory video.
Planned delivery: 1920 × 1080, 16:9, 30 fps, approximately 26 seconds.
Style: Calm voice, clean shapes, dark blue card, warm light, pale screen.
Learning goal: The viewer can predict which direction to move the card to enlarge the shadow in this specific setup.

YOU NEED

- A text AI tool, if you want assistance with planning and revision.
- An editor that can place images, draw or import shapes, keyframe position and size, add audio, and export video. Button names vary by product.
- Your own recorded voice, or a synthetic voice you have permission to use. A voice-generation tool is optional.
- An optional small, low-power light, opaque card, and wall for a physical check. Keep the light and wall fixed, move only the card, and do not shine light into eyes. Avoid hot lamps and lasers.
- A place to save a brief, script, storyboard, assets, audio, captions, source notes, and export.

1. SET ONE ACCURATE CLAIM

Scope the video before prompting. Use a small light whose rays spread out, one flat opaque card, and a screen behind it. Keep the light and screen fixed. Avoid a tightly focused beam. The card stays perpendicular to the light-to-screen axis and centered on it.

Science check: In this arrangement, moving the card toward the light enlarges its shadow on the screen. The opposite movement reduces it. Science Buddies describes this effect in its original shadow-puppet activity: https://www.sciencebuddies.org/stem-activities/shadow-puppets
This link supports the science only. The script, prompts, layout, and workflow here are original. No third-party images or text are included.

Treat the diagram as an idealized point-source model. Real lamps have a finite emitting area and can create soft edges. This lesson does not explain the changing length of outdoor shadows across the day.

For a geometry check, use similar triangles: shadow height = card height × (light-to-screen distance ÷ light-to-card distance). This is a construction rule for the idealized diagram, not a measurement from an experiment. If the card is halfway to the screen, its shadow is twice its height. At one quarter of the distance, the shadow is four times its height.

Prompt A — Turn a bounded idea into a brief

Act as an educational video planner. Plan a 26-second English explainer for beginners. Topic: a flat opaque card moving between a fixed small light source and a fixed screen. Use an idealized point-source diagram. The single learning goal is to predict which card movement enlarges the shadow. Propose three opening questions, then select one that can be answered visually. Return a one-sentence brief, an audience description, one learning goal, a list of assumptions, and a list of claims that need checking. Do not introduce outdoor sunlight, extra science topics, invented sources, statistics, or promises about learning results. Flag uncertainty instead of guessing. The final production will use editor-drawn geometry rather than generated footage as scientific evidence.

Decision for this sample: use the opening question “Why does this shadow grow?” It names the visible change without adding another topic.

2. WRITE AND TIME THE SCRIPT

Prompt B — Draft the narration

Write one English narration script of 50–60 words for the brief below. Use short, speakable sentences and one question at the start. Describe the controlled setup before the change. Explain the effect using spreading rays and the area blocked on the screen. Finish with a recap of what changes and what stays the same. Keep the tone warm and matter-of-fact. Do not add a call to subscribe, claims about outdoor shadows, exact real-world measurements, or invented experimental results. Return only the narration, its word count, and any scientific ambiguity you notice. Aim for 26 seconds with pauses, but say that actual duration must be checked with recorded audio.

Brief: Show a card first halfway between a fixed small light and screen, then closer to the light, then halfway again. Use an idealized point-source diagram with correctly drawn straight rays. The card's physical dimensions never change.

Selected original voice script

Why does this shadow grow? Keep the lamp and screen still. Move the card toward the light: its shadow gets bigger. Move it toward the screen: the shadow gets smaller. Light spreads outward, so the card blocks a wider area on the screen when it's nearer the lamp. Same card, different position, different shadow.

Read it aloud before making visual assets. Record one natural take, then measure it. The planned shot lengths below are provisional. If the narration exceeds 30 seconds, trim words or simplify a sentence. If it is under 20 seconds, add breathing room around the two changes. Do not speed the narration until it becomes difficult to follow.

3. PLAN THE SHOTS

Prompt C — Make a usable storyboard

Create a five-shot storyboard totaling 26 seconds for the exact script pasted below. Return one numbered paragraph per shot with start and end time, visible elements, movement, matching narration, and on-screen text of no more than six words. Keep the camera locked. Use one light, one card, and one screen in a consistent side-view diagram. The light and screen must never move; the card keeps the same height and moves only horizontally. Preserve the script's meaning. Flag any action that cannot fit its assigned time. Do not invent extra narration or physical behavior. Use these segments: 0–3, 3–6, 6–11, 11–16, 16–26 seconds.

Script: Why does this shadow grow? Keep the lamp and screen still. Move the card toward the light: its shadow gets bigger. Move it toward the screen: the shadow gets smaller. Light spreads outward, so the card blocks a wider area on the screen when it's nearer the lamp. Same card, different position, different shadow.

Planned shot list

Shot 1 | 0–3 s | The question
Show the complete simplified side view. Briefly alternate two clearly labeled card positions and their corresponding shadow sizes, with a cut between them. Keep the card the same height.
Narration: “Why does this shadow grow?”
Text: “Why does the shadow grow?”

Shot 2 | 3–6 s | Fix the setup
Hold the card halfway between the source and screen. Label the left point “Light,” the center rectangle “Card,” and the right vertical surface “Screen.” Add small lock symbols beside the light and screen if useful.
Narration: “Keep the lamp and screen still.”
Text: “Light and screen stay fixed”

Shot 3 | 6–11 s | Move toward the source
Move the card left to the quarter-distance position. Extend straight boundary rays through the card's top and bottom corners. Increase the dark segment on the screen according to the geometry below. Hold the end state long enough to compare it.
Narration: “Move the card toward the light: its shadow gets bigger.”
Text: “Toward the light → bigger”

Shot 4 | 11–16 s | Reverse the change
Return the card to its original halfway position. The projected segment reduces while the light and screen remain fixed.
Narration: “Move it toward the screen: the shadow gets smaller.”
Text: “Toward the screen → smaller”

Shot 5 | 16–26 s | Explain and recap
Show the two endpoint diagrams side by side or cut between them. Highlight the blocked region behind the card, bounded by straight rays. Label the screen segment as the shadow. Finish on the two states with the identical card dimensions obvious.
Narration: “Light spreads outward, so the card blocks a wider area on the screen when it's nearer the lamp. Same card, different position, different shadow.”
Text: “Same card. Different position.”

Timing note: The selected narration has 54 words. Shot 5 receives ten seconds for its explanation and recap. Adjust the holds or extend the whole edit up to 30 seconds after measuring the voice track. The five-shot timing plan is a starting point, not a claim that a voice take has been tested.

4. BUILD VISUALS YOU CAN CHECK

Use simple shapes for the factual diagram. AI-generated pictures often produce plausible but incorrect rays, labels, or object sizes. A generated still is a style proposal until checked and rebuilt as needed.

Diagram construction on a 1200 × 720 design canvas

- Origin is the top-left corner. x increases rightward; y increases downward.
- Light source center: x = 80, y = 360.
- Screen: vertical line at x = 1120.
- Card height: 120 units, from y = 300 to y = 420.
- State A: card x = 600. Rays from the light through the top and bottom card edges reach the screen at y = 240 and y = 480. Shadow height: 240.
- State B: card x = 340. The same construction reaches the screen at y = 120 and y = 600. Shadow height: 480.
- Draw each boundary ray from the source to its matching screen endpoint. Make the segment behind the card dashed to indicate a geometric continuation into the blocked region. Solid rays must not appear to pass through the opaque card.
- Shade the wedge between the card and screen subtly. Make the dark vertical segment on the screen the strongest shadow mark. State “Simplified side view” on the first diagram.
- Scale the complete diagram uniformly into the final 16:9 frame. Add labels in the editor, never as baked-in AI text.

For intermediate motion, calculate shadow endpoints at each card position using the same ratio. Simple linear resizing of the shadow will not stay geometrically correct throughout the move. If the editor cannot maintain the relationship, use accurate endpoint stills and a direct cut rather than an incorrect animated ray diagram. A short cut is a complete, acceptable version of this lesson.

Prompt D — Optional visual direction

Create a visual style proposal for a beginner science explainer. Use a pale warm-gray background, a small amber light icon on the left, a dark blue flat card in the center, and a pale vertical screen on the right. Use clean, simple shapes with generous empty space. Keep a fixed side view. No people, hands, brand marks, text, numerals, light rays, shadows, or photorealistic detail. The scientific rays, shadow geometry, labels, and animation will be drawn separately by a human editor. Return a single 16:9 reference image if image output is supported; otherwise describe the palette, composition, and shapes in plain English. Do not claim the result is a scientifically accurate diagram.

Keep asset names descriptive: diagram_state_a, diagram_state_b, background, narration_take_01, captions, export_review_01. Record which elements were generated and any relevant usage terms. Do not upload private reference files to an AI service unless you have permission.

5. RECORD OR GENERATE THE VOICE

Prompt E — Narration direction

Read the script below in clear, natural English. Sound like a calm teacher explaining one visual change. Use a neutral voice that does not imitate a real person. Add a short pause after the first question and after “still.” Emphasize “bigger” and “smaller” lightly. Read only the script, without headings, stage directions, extra words, or music. Aim for a natural pace that can fit approximately 26–30 seconds; do not guarantee a duration. If you cannot generate audio, return the script with suggested pause marks instead.

Script: Why does this shadow grow? Keep the lamp and screen still. Move the card toward the light: its shadow gets bigger. Move it toward the screen: the shadow gets smaller. Light spreads outward, so the card blocks a wider area on the screen when it's nearer the lamp. Same card, different position, different shadow.

Check every word in the actual audio. Re-record awkward phrases rather than hiding them under music. Using your own voice is a complete alternative. Music is optional; omit it for the first version unless it helps and you have the required rights.

6. EDIT AND EXPORT

Prompt F — Build the edit plan

Act as an editing assistant. Use the five-shot storyboard and exact narration below to propose a plain-language edit checklist for any basic video editor. Preserve the locked light and screen and the checked diagram states. Include how to align changes to the spoken words, add readable captions, check audio clarity, inspect the full exported file, and preserve an editable project. Do not invent menu names, tool compatibility, export guarantees, generated assets, or a completed test. Ask for the actual recorded duration if exact caption timestamps are needed. Do not create timestamps from text alone and claim they match recorded audio.

Storyboard: 0–3 question; 3–6 fixed setup; 6–11 card toward light; 11–16 card toward screen; 16–26 explanation and recap. These are provisional segments; the measured narration controls final timing.
Narration: Why does this shadow grow? Keep the lamp and screen still. Move the card toward the light: its shadow gets bigger. Move it toward the screen: the shadow gets smaller. Light spreads outward, so the card blocks a wider area on the screen when it's nearer the lamp. Same card, different position, different shadow.

Manual editing steps

1. Create a 1920 × 1080, 30 fps sequence. Import the checked diagrams and recorded narration. Save the editable project.
2. Put the voice track on the timeline first. Mark the question, fixed setup, each movement, and recap. Adjust shot boundaries to the actual read.
3. Place the two checked diagrams. Prefer clean cuts. Animate only if you can preserve the geometry at intermediate frames.
4. Align each state change with its spoken description. Leave a short hold after the change so the viewer can inspect it.
5. Add labels as editable text. Use high contrast and sufficient size for a phone-size preview. Avoid placing important material near the frame edges.
6. Make captions from the final audio, then correct them by listening. Use short phrases, usually one or two lines. Keep them away from the diagram labels. Export a separate caption file if supported; burn them in if the intended player will not use a caption file.
7. Listen on headphones and a phone speaker. Keep the voice intelligible and free of clipping. Lower or remove music if it competes with speech.
8. Export an MP4 using H.264 video and AAC audio if supported by the editor and destination. Use the sequence's 1080p dimensions and 30 fps. These are working defaults, not a promise about every platform's requirements.
9. Watch the actual exported file from start to finish with sound, then with sound off. Check the first frame, last frame, captions, state changes, audio sync, and total duration. Save the project, source notes, captions, and final file together.

7. QUALITY GATE BEFORE SHARING

Prompt G — Review without inventing a test

Review the script, storyboard, and asset descriptions I provide. Separate issues you can check from text from issues that require the actual video or a physical demonstration. Check factual scope, geometry assumptions, timing plausibility, reading difficulty, accessibility, and unsupported claims. Return up to five specific fixes ranked by impact. Do not state that you watched, listened to, tested, verified, or exported anything unless the relevant file was actually supplied and you can inspect it. Treat AI-generated visual output as unverified. If the video is unavailable, explicitly leave visual, audio, and export checks open.

Final checklist

[ ] The lesson explains the indoor small-source setup, with a clear simplified-model label.
[ ] The light and screen remain fixed. The card's size and orientation remain fixed.
[ ] All shown ray endpoints match the chosen card position; no solid ray passes through the card.
[ ] The shadow changes in the correct direction, including any intermediate animated frames.
[ ] The final narration matches the demonstrated movement and does not overgeneralize to sunlight.
[ ] The measured exported duration is 20–30 seconds; nothing depends on a claimed recording speed.
[ ] Captions are accurate and legible at the intended viewing size, with no overlapping labels.
[ ] The voice is clear and synchronized; the final export has been watched and listened to.
[ ] All visual, voice, music, and font rights are appropriate for the intended use.
[ ] AI assistance is described truthfully wherever disclosure is needed.
[ ] No draft asset, untested prompt, or planned export is represented as a finished video.

Current check status

Completed for this textual sample: original writing; science reference consulted; idealized endpoint geometry calculated; workflow and claim-scope review.
Still open: running the prompts in any model; producing visual and voice assets; building the edit; testing comprehension; checking actual timing; rendering and reviewing an exported video.

NEXT ITERATION

After a Chinese PDF is supplied, first establish what may be adapted, identify its actual sequence and terminology, and separate source material from this original sample. Only then compare or revise the lesson. Do not describe this version as an adaptation of a PDF.