Auto Crop Video to 9:16 and Keep the Speaker in Frame
Face tracking plus on-screen speaker detection drives the crop: a dead zone absorbs jitter, pan speed is capped, and the camera cuts cleanly at every scene change.
Auto Crop Video to 9:16 and Keep the Speaker in Frame Features
EchoWave's auto crop tools are used by creators, video editors and streamers worldwide
Crop that tracks
A fixed center rectangle loses the subject. Tracking does not.
Auto crop video with a static center box and the speaker slides out of frame the first time they lean. EchoWave crops by watching instead: face detection finds everyone on camera, on-screen speaker detection (lip movement matched to the voice) works out who is talking, and the crop follows them with a dead zone for small movement, a capped pan speed, and smoothing so nothing whips. When the footage hits a scene cut, the crop cuts with it, never pans across it. Target 9:16 with the vertical video converter workflow, or 1:1, 4:5, and 16:9, and if you want a manual pass, the plain crop tool and AI reframe sit one click away.
-
7 layouts
Follow speaker, Split, Grid, Presenter + slides, Picture in picture, Fill, and Letterbox over blur
-
4 ratios
Crop the same source to 9:16, 1:1, 4:5, or a tighter 16:9 without re-tracking
-
0 pans
Across scene cuts: the camera always cuts cleanly when the source footage does
-
30 min
Of source video analyzed free every month; Creator covers 180 and Pro covers 600
What you get
Everything that breaks a naive auto crop, handled
Cut-off faces, whip pans, and trackers that jump at every camera change: each one has a specific fix here.
-
Speaker-Led Framing
On-screen speaker detection (lip movement matched to the voice) tells the crop who matters right now, so the frame centers on the person talking instead of a generic face box or the middle of the shot.
-
Capped Pan Speed
The camera moves like an operator, not a servo. Pan speed is limited and smoothed, so even a fast-moving subject produces a deliberate follow rather than a whip pan across the frame.
-
Dead Zone Steadiness
Head turns, gestures, and small shifts inside the dead zone do not move the crop at all. Only a real change of position triggers a pan, which keeps a talking-head shot calm for its full length.
-
Scene-Cut Aware
When the source cuts to another angle, the crop cuts with it. It never drags a pan across a scene change, the single most visible flaw in this category of tool.
-
Every Aspect Ratio You Post
Crop to 9:16 for TikTok, Reels, and Shorts, 1:1 or 4:5 for feeds, or a reframed 16:9. Pair it with resize for TikTok when a platform needs exact specs.
-
Layouts for More Than One Face
Split stacks two speakers, Grid handles three or four, and Presenter + slides keeps a screen share on top with the speaker below. Nobody gets cropped out when the other person starts talking.
-
Adjustable Camera Behavior
Tune steadiness and zoom, set a cut-versus-pan preference, or pin focus on one chosen speaker. If you want a straight center crop, the Fill layout does exactly that, no tracking involved.
-
Ordinary, Editable Keyframes
Every camera move lands in your project as plain keyframes. If the crop drifts in one shot, open the editor and fix that shot by hand. Nothing about the auto crop is locked in.
How it works
How to auto crop video
Landscape in, tracked vertical out, in four steps.
-
Add Your Footage
Upload a landscape video, pick one from your library, paste a direct media link, or start from an existing project. Podcasts, interviews, webinars, lectures, and gaming VODs all work.
-
Pick a Ratio and Layout
Choose 9:16, 1:1, 4:5, or 16:9, then a layout: Follow speaker for one person, Split for two, Grid for panels, Letterbox over blur for gameplay with facecam.
-
Review the Camera Moves
Watch the tracked crop, then adjust steadiness, zoom, or the cut-versus-pan balance. Pin focus on a specific speaker, or edit any individual keyframe in the full timeline.
-
Export the Vertical Cut
Render through the normal pipeline. Free exports carry a small watermark and a 720p cap, Creator renders 1080p, and Pro renders 4K. Bulk export several ratios at once.
Who uses it
Who auto crops video with EchoWave
-
Shorts and Reels creators
Turn landscape podcast and vlog footage into 9:16 posts without the speaker drifting out of frame mid-sentence.
-
Interview editors
Crop two-person conversations with Split so both guests stay visible, or let the camera cut between speakers on their turns.
-
Streamers and gamers
Fit facecam plus gameplay into one vertical frame with Letterbox over blur or Picture in picture, no manual masking pass.
-
Webinar and demo teams
Presenter + slides keeps the screen share readable on top with the speaker framed below, sized for portrait viewing.
-
Educators and speakers
Crop lecture and talk recordings that cut between camera angles; the crop cuts cleanly at every angle change.
-
Editors who finish by hand
Use the tracked crop as a fast first pass, then adjust individual keyframes shot by shot in the editor.
Stop babysitting the crop
Upload a landscape video, pick 9:16, and get a tracked, steady crop you can still adjust shot by shot. Your first 30 minutes each month are free.
What creators say after trying EchoWave
About the EchoWave team
EchoWave is a browser-based video and audio editor built by Lemon Vault LLC since 2018. The people who build the editor write these guides and test every step in the production app before publishing.
Auto crop video FAQ
How do I auto crop video?
Upload your landscape video to EchoWave or pick one from your library. Choose the target aspect ratio, usually 9:16 for TikTok, Reels, and Shorts, then a layout such as Follow speaker or Split. The Smart camera detects faces and who is talking, then builds the crop with a dead zone, a capped pan speed, and clean cuts at scene changes. Review the result, tweak any shot you want, and export. Free exports carry a small watermark and a 720p cap; Creator renders 1080p and Pro renders 4K.
What is smart crop or auto crop for video?
Smart crop, also called auto crop or auto reframe, is cropping that follows the subject instead of holding a fixed rectangle over the middle of the frame. A static center crop works only when nobody moves; the moment a speaker leans or walks, they leave the box. A smart crop tracks faces, decides who to frame, and moves the crop window over time. In EchoWave that movement is written as ordinary keyframes, so the automatic result and a hand-made crop are the same kind of thing, and equally editable.
Can AI crop a video and keep the speaker's face in frame?
Yes. EchoWave detects every face on screen, then uses on-screen speaker detection (lip movement matched to the voice) to work out who is actually talking. The crop frames that person at eye level and follows them if they move, within a dead zone so small shifts do not cause motion. When another person starts speaking, the camera cuts to them on the turn rather than panning across the set, which reads like a switched multi-camera edit.
Why does auto crop make my video look jumpy?
Most auto crop tools move the frame every time the detected face shifts by a few pixels, and they keep panning even when the source footage cuts to a different shot, which produces the smeared whip-pan look people complain about. EchoWave was tuned against exactly that: a dead zone ignores small movement, pan speed is capped and smoothed, and the camera never pans across a scene cut, it always cuts cleanly instead. Steadiness is also a slider, so you can make the crop calmer or more responsive per project.
Can auto crop handle a video with more than one person?
Yes, and this is where a single-face tracker usually fails. Follow speaker cuts between people as the conversation moves, based on who is actually talking. Split stacks two speakers in one vertical frame so neither gets cropped out, and Grid fits three or four. If one person should always hold the frame, a host over a chatty guest for example, pin focus on that speaker and the camera stays with them regardless of who speaks.
Does auto crop work on gameplay and facecam footage?
Yes. Pick the gaming genre preset, then choose a layout that fits the footage. Letterbox over blur keeps the full widescreen gameplay visible over a blurred fill, so nothing important gets chopped off, while Picture in picture floats the facecam over the action. For screen-share content like tutorials and webinars, Presenter + slides puts the screen on top and the speaker below. Analysis is also screen share aware, so meeting and demo recordings frame sensibly.
Can I fix an auto-cropped video if the crop is wrong in one spot?
Yes, and you should expect to now and then; no tracker is perfect on every frame. Every camera move EchoWave makes is stored as ordinary keyframes in your project, not baked into the pixels. Open the clip in the full editor, find the shot where the framing drifts, and drag the crop where it belongs. The rest of the video keeps its automatic tracking. You can also re-run the crop with different steadiness, zoom, or layout settings.
Is EchoWave's auto crop free?
The free plan analyzes 30 minutes of source video per month, enough to crop several clips and judge the tracking quality on your own footage. Free exports include a small watermark and are capped at 720p. Creator raises the allowance to 180 minutes with 1080p exports, Pro covers 600 minutes with 4K, and Ultra includes 3,000 minutes with workspace-scoped jobs your whole team can see.
Crop smarter post vertical
A tracked 9:16 crop with no whip pans, editable down to the keyframe.
Get Started →
