reframe

Scales landscape source footage to fill a portrait output frame by computing the scale needed to fill the output height and centering the crop horizontally.

When to use

Use when your source footage is landscape (e.g. 3840×2160 or 1920×1080) and your output is portrait (e.g. 1080×1920). Run before motion_center or speaker_center for dynamic subject centering. See Portrait Video guide.

Tip: The edit processor's reframe operation does the same fill-and-center reframing, and can be chained with other adjustments in the same step. This standalone processor remains available and unchanged.

In the editor

Add it from the Library (Transforms group), or with + add step and a search for reframe.

It has no parameters. The whole configuration is one entry under Inputs: an id of source, pointing at either a file (the footage you uploaded) or a step (an earlier step whose clips should be reframed).

It produces no clips of its own, so it gets no row of its own in the run timeline. What you see instead is that the footage placed by other steps now fills a portrait frame. Where you put the step in the list matters in one way only: a step input can only point at a step above it, so wire it below whatever it reads. The reframing itself is applied to every matching clip when the run is exported, whichever step placed them.

YAML Example

- id: reframe
  processor: reframe
  input:
    - id: source
      file: speaker-footage
  cache: true

Inputs

Params

This processor has no configurable parameters.

Output

TransformPatches: scale and position transforms applied to all edit decisions that use this source.

Notes

• No API key required.
• The crop is always centered horizontally. To offset the horizontal center, follow with motion_center or speaker_center.
• The transform is applied to every clip matching the patch when the timeline is built, so a reframe step does not have to sit before or after the steps that placed those clips. Position matters only for a step input, which can reference earlier steps only.
• Among transform steps, later ones win: a motion_center or speaker_center step further down the pipeline overrides the horizontal position this sets.