Skip to content

Wan 2.2 / Text to video

Wan 2.2 fast-motion fixes: simplify direction and framing

Fast motion is difficult to evaluate when the camera, subject and background all change direction. A useful Wan retry reduces that ambiguity. Start with one subject moving along a clear path and keep enough space in the frame to follow it.

By JollyAIReviewed 2 min read

Identify the source of confusion

If the subject appears to teleport, the requested action may be too complex for the short shot. If its shape changes, a large viewing-angle change may be exposing uncertain geometry. If the camera loses it, the framing may be too tight. These are different problems and deserve separate revisions.

Do not confuse added motion blur with coherent movement. Blur can make an artifact less visible, but it does not establish that the underlying action is correct. Review the result at normal speed and inspect the difficult section slowly.

Try a simpler action

  1. Remove scene cuts and competing camera moves.
  2. State one travel direction and a steady speed.
  3. Use a wider framing with the entire subject visible.
  4. Reintroduce complexity only after the basic movement is acceptable.

Example repair prompt

PROMPT RECIPE / ADAPT TO YOUR SOURCE

One adult cyclist wearing a helmet rides steadily from left to right along a straight closed practice path. A fixed side-view camera shows the whole bicycle throughout the shot. Stable daylight, simple background, believable wheel rotation, no turn or scene change.

Choose a usable production approach

If a complex stunt is essential, real footage or controlled 3D animation may be more suitable. For a concept montage, a shorter coherent segment may be enough. Keep the accepted segment and document its limits so another editor does not use the distorted remainder. Add sound in post according to the actual output rather than assuming a visual prompt creates an audio track.

REVIEWED JOLLYAI IMPLEMENTATION

The workflow behind the guide

Written scene → configured Wan text conditioning and sampling → frame decoding → video encoding and any selected post-processing.

Inputs, presets & access
PRO for the reviewed standard Wan T2V route. Use the visible preset; do not infer file metadata from the prompt.
Implementation note
JollyAI runs configured Wan workflows with locally selected checkpoints and adapters. T2V, I2V and first/last-frame are separate routes. Presets and post-processing differ by route. Plan sound separately for the demonstrated silent workflow, and inspect the downloaded file before quoting dimensions or duration.

Configuration reviewed September 20, 2026. Your current playground controls and plans page determine availability. Prompt recipes above are suggestions, not guarantees or claimed test results.

Open Text to video ↗

Sources and useful next steps

The implementation notes describe JollyAI's reviewed local integration. Upstream capabilities and licences may differ from the settings available here.

Found an outdated setting or a reproducible problem? Contact JollyAI with the workflow and job reference. Keep account tokens and private input files out of public reports.

KEEP BUILDING

Continue the workflow.