Why does slow motion bend my product label even though the original clip is correct?

Locate the first changed frame

Keep the original-speed clip and identify the interval where the label is readable. Compare it with the retimed output at the same product position. A warped label already present in the source is a source problem; a defect appearing only in the edited version directs investigation to the processing chain, including interpolation, effects and export compression.

Inspect fine lettering, cap edges and nearby occlusions rather than only the large product silhouette. Pause at the transition and step through frames. A smooth wide shot can conceal a brief change in the brand name that matters when the buyer watches closely.

Compare interpolation with one source interval

Export short labeled tests using the same source, timing and framing. Frame sampling, blending and optical flow solve retiming differently; none should be approved solely because its name sounds more advanced. Compare motion quality together with the product evidence that must remain unchanged.

If slower playback cannot preserve the label acceptably, reduce the slowdown or choose another supported interval. A simpler edit may be preferable to a smooth but false product mark. The native source and every test stay separate so an unsuccessful experiment does not replace the approved file.

1. Save the source and mark the label’s correct interval.
2. Confirm the retiming method in the external editor.
3. Make same-interval tests without changing framing or color.
4. Step through intermediate lettering and occlusion frames.
5. Choose the least misleading edit or hold the slowdown.

Match source time, not the same timeline timestamp

For a fictional 50% speed test, a source moment one second after the selected in-point appears two seconds after the retimed in-point. The mapping assumes a constant speed with no ramp or freeze; it is not a universal formula for an edited sequence. Compare the same source action before judging which method changed the label.

Keep the source in/out, speed, sequence/output frame rate and export settings with each test. Sampling does not protect against an unrelated effect or export defect. If all tests fail, locate that common stage before blaming the interpolation method.

Fictional exercise: a turning bottle cap

In this fictional exercise, a bottle turns and its label passes behind a highlight. The original frames retain the name, while one optical-flow test creates a bent character between them. The example is invented and does not measure any model’s error rate.

A sampling test preserves the lettering but looks less fluid. The fictional reviewer chooses a smaller slowdown and evaluates both label fidelity and motion again. If neither version meets the brief, the stop decision is to keep the normal-speed shot, not claim that generated lettering has been corrected universally.

Approve the final edit after export

Replay the chosen exported file, including the transition into and out of the retimed interval. Check the actual delivery scale as well as paused label frames. Save the source interval, method and speed with the acceptance record rather than approving an editor preview alone.

CartVisual creates visual material; retiming and postproduction are separate external work. Stop when the final export preserves the approved name and visible product details while meeting the motion brief. If later edits change the speed or interpolation, reopen this check instead of relying on the previous approval.

Retiming comparison

Use this matrix as a manual review record and add your own real evidence. Its conditions are not completed test results.

Download this question’s review worksheet (CSV)