Zoom in from one fixed viewpoint: enlarge the arrangement without changing its perspective

Original, untested video brief. No input image or generated result is supplied.

1. Task and boundary

This shot asks for one slow increase in image magnification while the camera position, viewing direction and objects stay fixed. A small foreground block and a more distant patterned board provide two depth layers. Their image sizes grow together; the block must not independently swell or move toward the lens. The useful evidence is a continuous change across the whole picture, followed by a short steady ending.

Keep this a modest zoom with both objects visible throughout. It is a visual consistency exercise, not a calibrated lens simulation or a demonstration of a particular focal length. A synthetic zoom and a crop can share the same perspective cues, so those cues alone cannot establish the physical optical process. Do not describe the result as proof that a real lens moved correctly.

2. Authorized first frame

Create your own photograph or original drawing, or use an image you are authorized to adapt. Show an unbranded rectangular block near the camera and a larger upright board behind it, both planted on one tabletop. Give the board several widely spaced dots and leave its outer edges inside the initial frame. Include a recognizable block corner that overlaps one board mark, plus a second mark visible beside the block. Avoid tiny lettering, reflections and people.

Choose an initial view that already contains the complete scene needed at the end. Leave enough space around the block that a restrained enlargement will not crop away its base or all background marks. Keep the camera approximately level and the central aim near the block. Use soft steady light and a plain table seam as another fixed reference. A still image with ambiguous object edges will make enlargement difficult to distinguish from reshaping.

3. Use the generator

Open the current image-to-video generator in a mode accepting your authorized first frame. Paste the complete prompt, check the available settings and cost estimate, then submit only when ready. This page supplies text; it does not upload an image or start a generation.

4. Prompt and timing

Use the supplied first frame. Keep the rectangular block, distant dotted board and tabletop completely stationary. Keep the camera at the same position and keep its viewing direction fixed. Make one slow, modest zoom in: the block, board marks and table details enlarge together around the same central aim. Preserve the relative size relationship of the near block and far board, and preserve which board mark is hidden behind the block corner. Do not move the camera forward or sideways, rotate it, move either object, reveal a new side of the block, or stretch the block alone. Keep the block base and enough of the board visible to compare them. Maintain steady lighting, recognizable rigid edges and a stable focus impression. Ease the enlargement to a stop, then hold the larger view without a reverse zoom or cut.

1. Opening: hold the original framing briefly so the block corner, hidden board mark, visible neighboring mark and table contact can be identified before enlargement begins.
2. Middle: enlarge the entire visible arrangement gradually. Follow the near and far references together; a background edge should not slide out from behind the block as though the viewpoint had translated.
3. Ending: stop at the modest larger framing while the block base and at least two useful background references remain visible. Hold long enough to compare their relationships without adding a second camera move.

These beats are author suggestions, not exact model timing guarantees. Adapt their proportions to the clip length actually available.

5. Inspect the output

After inspecting your own output, record a time or frame, visible observation and pass, fail or inconclusive judgment for each condition. A convincing ending alone does not verify the preceding action.

• The block and the board remain the same rigid objects, with stable edge shapes and no newly revealed block face during the action.
• Near and far image features enlarge together; their relative sizes remain approximately stable rather than the foreground alone growing much faster.
• The same board mark stays occluded by the chosen block corner, while the neighboring visible mark stays on the same side of that corner.
• Image movement belongs to the shared enlargement around a stable central aim, without a separate sideways drift, pan or forward-parallax event.
• The block base remains planted on the table, and the final larger view settles without rebound, reverse zoom or a discontinuous cut.
• If cropping, blur or overlapping references prevent a near-and-far comparison, record inconclusive; a larger foreground object by itself is not enough evidence.

6. Fix one failure

• If only the block grows, reduce the requested zoom and explicitly name the board dots and table seam as features that must enlarge at the same time.
• If the camera appears to advance, emphasize preserved overlap and relative scale. Use a clearer foreground corner against two distant marks; do not repair the clip merely by freezing the background.
• If the central aim wanders, remove any pan, orbit or dramatic-camera language. Ask for one fixed viewing direction and a shorter enlargement.
• If the endpoint loses every reference, choose a wider starting frame or reduce magnification. Keep enough of both depth layers in view instead of inventing new objects near the edges.
• If edges breathe or sharpen into a different shape, simplify the input texture and avoid a simultaneous focus pull. Reinspect the complete transition after changing one instruction.

7. Ending and limits

The ending is the same arrangement seen larger, with the same overlap and approximately the same near-to-far scale relationship. It does not include a return to the opening view. Keep your inspection notes tied to the changing frames as well as the hold, because a plausible final composition can conceal an incorrect intermediate camera move.

Save your own generated result through the available result controls, together with the authorized input, exact prompt, settings and inspection notes.

Output may imitate zoom imperfectly or replace it with cropping, object scaling or forward motion. No generated clip, lens measurement, model setting or success rate has been tested for this brief. Use only controls actually available in your chosen mode; timing beats do not guarantee frame-accurate execution. Keep the authorized source and your own observations with any saved result.

The dolly-zoom task offers a useful comparison: it asks the subject to stay similar in size while background perspective changes. Here the whole arrangement grows while perspective cues stay stable. Separating those two goals makes the inspection useful even when the output looks cinematic.