For the lesson you can no longer unflatten
Keep the equation.
Slow the distraction.
When a repeating decoration competes with the lesson, changing the whole clip changes too much. Mark the text, diagram or counter that matters. Review a local 5× retime, then check the actual export. This page shows static images only.
One small edit. A specific promise to check.
With 5× local retiming, the important areas stay on the original timeline and supported audio is copied and checked. The decoration gets an explicit edit plan. A failed preservation check withholds the edited download.
Start with the owned synthetic workshop fixture in the repository. It uses generated shapes and a tone; real lesson benefit is still untested.
Choose a clip
The server validates the file against one of these two exact profiles before analysis.
- Resolution
- 320 × 180 or 640 × 360
- Frame rate
- 30 fps
- Length
- 1.0–12.0 s in 0.1 s steps
at either resolution - Audio
- Optional AAC-LC mono 48 kHz only: copy unchanged, verify packets, timing and decoded samples. No extra tracks or offsets; local retiming only. Canonical pinned FFmpeg mux metadata is required; many otherwise-playable AAC MP4 files are refused.
H.264 · YUV420 · limited-range BT.709
Maximum input: 4 MiB / 8 MiB by profile
Server profile & limitations
Connecting to the local server…
Review the proposal
Your review starts here
After analysis, inspect a static overlay and its exact frame interval. Nothing is approved automatically.
- Analyze
- Review
- Encode
- Result
Choose the explicit action below. Only one candidate can be approved for this job. Retiming repeats one observed cycle and can discard subtle variations; it does not smooth motion. Encoding can change pixels outside the region. The legacy interval hold, if selected, reuses the first sample in each 10-frame block. It changes the animation and can cause an abrupt change when the original frames resume.
Generating a comparison does not accept the output for use. Review the encoded measurements before choosing an output.
Encoded-result measurements
These values come from the server’s result report. Read the scope and limitations alongside the measurements.
Static boundary comparison
Four fixed moments, with the original input, unchanged re-encode control and generated edit side by side. Every image is a static PNG crop of the same region. These selected moments do not show the full motion.
Frame, crop & hash provenance
Exact frame indexes, times in seconds, crop coordinates [x, y, width, height] and SHA-256 values reported by the server. The browser does not independently verify file hashes.