Face Tracking: Auto-Reframe Horizontal Video for Shorts

Face Tracking lets you turn landscape (16:9) footage into vertical (9:16) Shorts without manually repositioning the crop. Bitcut detects faces in your video, follows them frame by frame, and produces a smooth virtual camera that keeps the speaker centered in the vertical frame.

This is essential when you have interview footage, podcast recordings, or any horizontal video you want to repurpose for YouTube Shorts, Instagram Reels, or TikTok.

How It Works

When you enable face tracking on a clip, Bitcut runs on-device analysis to:

  • Detect all faces in every frame of the clip
  • Identify scene changes — camera cuts, angle switches, wide-to-close transitions
  • Generate keyframes that define where the vertical crop should be at each moment
  • Apply spring physics to smooth the camera movement between keyframes

The result is a virtual camera that pans naturally to follow the active speaker, with instant cuts at scene boundaries and smooth motion everywhere else.

No jitter. The spring-physics engine uses dead zones and smoothing to prevent the camera from reacting to small head movements. The crop stays steady when the face is near center, and moves only when needed.

When to Use Face Tracking

Face tracking is most useful when:

  • Your source video is horizontal (16:9) and you need a vertical export
  • The subject moves around the frame or the camera cuts between angles
  • You have multiple speakers and the active face changes throughout

If your video is already vertical, or the subject is always centered, you may not need face tracking at all — a static center crop works fine.

Next Steps