Automatically Keep a Presenter Centered in Vertical Video
When horizontal or loosely framed footage is adapted to vertical video, a presenter can drift toward the edges of the crop. In this workflow IntelliTrack follows a stable feature on the person and Tracker Match Move compensates the movement. A downstream Transform defines the final scale and framing, keeping the presenter visually anchored in the chosen part of the vertical canvas without dozens of manual position keyframes.
Why this graph is built this way
Workflow logic
The task only needs two main processing nodes. Tracker analyzes motion and converts it into positional correction; Transform handles final composition, scale and edge safety. Keeping those jobs separate makes the setup easy to understand and reuse. If the presenter should sit off-center to leave room for captions, change Transform Center without retracking the shot.
Node tree
Connection diagram
The cable colors follow Fusion conventions: blue is an Effect Mask, neutral lines carry images, and colored ports indicate specific node inputs. Node-page links are intentionally disabled on the English diagram until the matching Fusion Encyclopedia pages are translated.
Node order
How to read the chain
MediaIn1 → Tracker1
Feed the source clip to Tracker and choose a high-contrast stable feature on the presenter in IntelliTrack, usually part of the face, head or clothing that is not frequently occluded.
Tracker1: Track Forward / Backward
Track through the required range and inspect turns, fast movement and partial occlusions. If the track drifts, retrack that section from a more stable feature.
Tracker1: Operation → Match Move
Match Move converts the recorded feature movement into frame compensation, so the image shifts relative to the presenter and keeps the tracked point stable.
Tracker1 → Transform1
Use Transform after stabilization for the final vertical crop. Increase Size and adjust Center so the face and body sit correctly while preserving room for captions or social UI.
Transform1 → MediaOut1
MediaOut receives the completed vertical reframing. The composition point can still be changed in Transform without recalculating the track.
Result
What the shot looks like
The presenter remains in a stable part of the vertical frame while speaking or walking. Framing becomes cleaner and repetitive manual position keyframes are replaced by one track plus a Transform adjustment.
When to use it
Where it works
- Adapting interviews and horizontal footage for YouTube Shorts, Reels and TikTok.
- Automatic talking-head reframing without manual Position keyframes.
- Keeping a walking presenter inside a chosen vertical-frame zone.
- Reserving side or top space for subtitles, titles and interface elements.
- Batch production of short vertical videos with a consistent framing rule.
Node parameters
What changes from the defaults
Tracker1
The source composition uses IntelliTrack with a tracking point on the presenter. Operation is set to Match Move so subject motion drives the framing correction.
- Tracker: IntelliTrack 1
- Operation: Match Move
- Merge: disabled
- Pattern Center: 0.683136 / 0.769286
- Pattern Width: 0.03
- Pattern Height: 0.0534
- Track recorded across frames 0–91
Transform1
Transform performs the final vertical framing after Match Move. Exact values depend on source material, but the method is consistent: scale until the vertical frame is covered, then place the person at the intended composition point.
- Center: X 0.13 / Y 0.32
- Pivot: X 0.495 / Y 0.5
- Size: 1.76
- Input: Tracker1.Output
- Offset Center when space is needed for text or graphics
Common mistakes
What not to break
- Do not track the mouth, eyes or another small feature that constantly changes shape; prefer a more stable area of the head or clothing.
- Avoid points that are frequently covered by a hand, microphone or another person.
- Do not scale without safety margin; compensation can expose empty edges if the crop is too tight.
- Do not try to fix a bad track only with Transform. Repair the Tracker path first.
- Do not force the face into the mathematical center when subtitles, logos or graphics need space on one side.
- Avoid overly rigid compensation on handheld footage; retaining a little natural camera movement can look better than perfect lock-off.