# How to Convert Landscape Video to Vertical Without Losing Quality > A center crop loses subjects who move. How reframers that follow a speaker work, what breaks them, and how to frame footage for ClipForge's default crop. - URL: https://clip-forge.io/blog/smart-reframing-landscape-to-vertical - Published: 2026-01-08 - Updated: 2026-10-26 - Author: ClipForge Editorial Team, Editorial Team, ClipForge AI (https://clip-forge.io/about) - Tags: reframing, vertical video, video editing, technology ## The Vertical Video Requirement Every major short-form platform expects 9:16 vertical video. TikTok, YouTube Shorts, and Instagram Reels all display content in full-screen vertical format. Viewers scrolling through these feeds expect vertical-native content. Horizontal video with letterboxing (black bars above and below) looks out of place and receives lower engagement. If your source material is recorded in 16:9 landscape — which is the case for most YouTube videos, podcast recordings, webinars, and conference talks — you need to reframe it before distributing as short-form clips. ## Why Center Cropping Fails The simplest approach to converting landscape to vertical is to crop the center of the frame. This works when the subject is perfectly centered and stationary. But in real-world content, it fails frequently: - **Speakers move.** In a podcast or interview, speakers gesture, lean, and shift position. A static center crop cuts off heads, hands, and faces at the worst moments. - **Multi-person content.** A two-person interview with speakers on the left and right of frame will crop both of them badly with a center cut. - **Presentations and screen shares.** Slides and screen content are designed for wide format. Cropping the center loses the margins where key information often lives. ## How Speaker-Tracking Reframers Work A tracking reframer uses computer vision to move the crop window frame by frame, following the visual focus of the content. ClipForge does a narrower version of this, on Pro and for one face only (see What ClipForge Does Today below), but it is worth knowing what the approach solves and where it breaks. ### Speaker Tracking A tracking reframer identifies faces and tracks them throughout the clip. When one person is speaking, the frame centers on them. When a different person starts speaking, the frame smoothly transitions to follow the new speaker. In multi-person scenarios, the frame may widen to include both speakers during exchanges and narrow to the active speaker during monologues. ### Motion Smoothing Raw tracking data produces a jittery frame because it reacts to every small movement. Good reframers apply motion smoothing to create natural, cinematic panning rather than abrupt jumps. The result feels like a camera operator is following the action rather than an algorithm snapping between positions. ### Scene-Aware Framing Not all content is talking-head footage. When a scene-aware reframer detects a presentation slide, screen share, or B-roll footage, it adjusts the framing strategy. For slides, it may zoom into the content area rather than tracking a speaker. For B-roll, it identifies the focal point of the scene. ## Common Reframing Challenges ### Side-by-Side Speakers When two people sit at opposite sides of a wide frame, the reframer must decide whether to show both (at a smaller scale) or follow the active speaker. The best approach depends on the conversation dynamic. Quick back-and-forth exchanges benefit from showing both speakers. Longer monologues benefit from focusing on the speaker. ### Fast Speaker Switches Rapid exchanges where speakers interrupt each other can cause distracting frame movements if the tracking switches too aggressively. Motion smoothing and minimum-hold-time settings prevent the frame from bouncing between speakers during quick exchanges. ### Tall Subjects Subjects who stand up, use a whiteboard, or present on stage may exceed the vertical frame when the crop is tight. The system adjusts zoom level to keep the full subject in frame when vertical space is limited. ## Reviewing and Adjusting Whatever reframes your footage, review each clip before it ships. Check for: - Faces or gestures cut off at the frame edge. - Frame transitions that feel too fast or too slow. - Moments where the wrong person is in frame during a conversation. ## What ClipForge Does Today ClipForge does not move the crop to whoever is speaking. By default every export is scaled to fill 9:16, 1:1, 4:5 or 16:9 and cropped from the center. On Pro, two tools track a face: Follow faces, a switch in the export dialog that is off by default, and the reframe tool in Pro Studio. Both run in your browser, follow one face (the widest in frame) and can zoom in up to 1.5x on a small one, so in a two-shot they hold one person rather than cutting to the one talking. In Pro Studio you can also keyframe the crop position by hand. That makes framing at the source the reliable fix: - Frame a single speaker near the middle of the shot. - Record each person in a two-person conversation on their own camera, rather than one wide shot with a speaker at each edge. - Export slide-heavy or screen-share segments at 1:1, which keeps more of the width than 9:16. ## Keep Reading - How AI Clip Detection Works: The Technology Behind Viral Moments (/blog/how-ai-clip-detection-works) - How to Make YouTube Shorts from Existing Videos: Step-by-Step (/blog/how-to-make-youtube-shorts-from-existing-videos) - Social Media Video Dimensions Guide 2026: Sizes and Specs for Every Platform (/blog/social-media-video-dimensions-guide-2026) ## Getting Started Upload a landscape video to ClipForge, keep the detected clips worth publishing, and export them at 9:16. Scrub each export once: when the subject sits near the middle of the source frame, the center crop holds them. ## Frequently asked questions ### How do you convert landscape video to vertical? Reframe rather than crop. A static centre crop only works when the subject is centred and still, which almost no real footage is. Speaker-tracking reframing follows the active subject through the frame, motion smoothing prevents the crop jittering as they move, and scene-aware framing adjusts when the shot changes. ### Why does centre cropping fail on real footage? Because speakers move. In a podcast or interview people gesture, lean and shift position, so a fixed centre crop cuts off heads and hands at the exact moments worth clipping. Two-person interviews are worse: with speakers on the left and right, the centre of the frame is the empty space between them. ### What footage is hardest to reframe automatically? Three cases. Side-by-side speakers, where no single crop holds both. Fast speaker switches, where the tracking has to cut rather than pan or it looks seasick. And tall subjects such as full-body or standing presentations, where a 9:16 crop tight enough to hold the face loses the gesture that made the moment work. ### Do you need to review automatic reframing? Yes, and it is quick. Scrub the reframed clip once at speed and watch for the crop losing the subject or swinging between two people. Most clips need no change. The ones that do usually need a fixed crop for a few seconds rather than a full manual pass, so the review costs far less than framing by hand. --- Published by ClipForge AI (https://clip-forge.io). Every post: https://clip-forge.io/long-llms.txt