Captions Blocking the Speaker's Face? Reposition Them

Move the caption band down to the lower third of the frame so it sits below the speaker's chin, or push it above their head, anywhere except across the eyes and mouth. The mouth carries lip-sync cues a watcher reads unconsciously, and the eyes carry the connection that holds a scroll. Both must stay clear. Re-crop so the speaker's head sits in the upper-middle, then drop the caption into the open space beneath.

The fix is almost always a placement problem, not a caption problem. Your editor shows the speaker centered with text below, looking fine. Then auto-reframe pushes the face down to fill a 9:16 crop, the caption follows the old position, and now the band cuts across the mouth. This guide gives you the two facial regions a caption must never touch, a band-by-band decision rule (lower third vs. top vs. the middle band that almost always fails), and how to reconcile "off the face" with "clear of the app's buttons."

Why captions on the face hurt the clip

A caption across the face costs you on two fronts at once: it blocks the lip-sync the silent majority leans on, and it covers the human connection that earns the watch. Most social video plays on mute, Digiday in 2016 cited publishers reporting silent-view rates anywhere from 50% to 85%, so treat it as a wide range, not a hard law. Either way, two things drive a muted watch: the words and the face. Cover the face with the words and you cancel one with the other.

The mouth is the costlier of the two to block. Even on mute, viewers read a speaker's lips to follow rhythm and emphasis; a caption band parked on the mouth fights that signal and the clip reads as off, even when the words are technically legible. The eyes are second, they carry the parasocial pull that makes someone stop scrolling. And the window is short: the first three seconds of a clip are, as castmagic puts it, "absolutely critical", viewers make a split-second decision to stay or scroll. A face you can't see in those three seconds is three seconds spent arguing with your own caption.

The two facial regions captions must never cover

There are exactly two no-go regions on a face: the eye line and the mouth. Everything else, forehead, cheeks, chin, the space below the chin, the space above the head, is fair game for a caption. Most "captions on the face" complaints are really "captions on the mouth," because the default lower-position band drifts up into the mouth when the face is cropped lower than expected.

The practical rule: line the top of your caption block to the speaker's chin, never higher. If the chin sits low in the frame and there's no room beneath it for the band, you have a cropping problem, not a captioning one, fix the crop first (see below), then the caption has somewhere to go.

Lower third, top, or middle: the band decision

Three horizontal bands are available on a vertical frame, and they are not equal. The lower third is the default and the best in most clips; the top is the rescue position when the face fills the lower half; the middle band, text running across the center where the face usually lives, is the one that creates the problem in the first place.

Here is the decision rule, by where the speaker's head lands in your crop:

Where the face sits	Use this band	The tradeoff
Head high, chin above center	Lower third (default)	Best readability; risk is the platform's bottom UI
Face fills the lower half	Top band, above the head	Clears the face; reads slightly less naturally
Two speakers side by side	Lower third, single line	Captions stay shared; keep them short

Three notes that keep the rule honest. The lower third wins by default because it's the band viewers' eyes already drift toward and it reads as part of the shot, not pasted over it, but it collides with the app's heaviest UI, which is its real cost. The top band is genuinely useful and underused: when a guest's face fills the bottom two-thirds of a tight crop, text above the head is the only spot that's clear of both the face and the bottom buttons. And the middle band has no legitimate use for a talking-head clip, if your captions are landing center-frame, the cause is almost always a crop that dropped the face too low.

Fix the crop first, then the caption

When captions keep landing on the mouth no matter where you set the band, the crop is the cause. Auto-reframe centers on the loudest face, and on a talking-head clip that often means the face sits low and large in the 9:16 frame, leaving no room below the chin. Repositioning the caption alone just trades one collision for another. Re-crop so the head sits in the upper-middle, eyes around the upper-third line, and the lower third opens up for text.

This is the same discipline as keeping a face from getting cut off in a vertical crop: the reframe decides everything downstream, so get the head placement right before you touch the caption. If the speaker moves around or there are two people on a wide shot, lock a manual reframe rather than trusting the auto-tracker to hold a stable head position, a face that drifts will drag your "safe" caption band onto the mouth halfway through the clip.

Off the face is only half the job: clear the UI too

A caption can sit perfectly clear of the face and still get covered, by the app's own buttons and progress bar. Placement has two jobs that pull in different directions: clear the face (a per-clip decision based on the crop) and clear the platform UI (a fixed margin that's the same every time). The lower third solves the face but runs straight into the bottom dead zone where TikTok, Reels, and Shorts stack their controls.

Reconcile them by sitting the band in the lower-middle, not the lower edge, below the chin but above the app's bottom UI. For the measured per-platform margins and the one baseline that clears every feed's buttons and scrubber at once, use the caption safe-zone map; this article handles the face, that one handles the interface, and a caption has to satisfy both. Contrast does the rest of the work: even a well-placed caption over a busy face needs a background plate or stroke so the text separates from the skin and shadows, see caption color and contrast for readability.

Common mistakes (and the fix)

• Moving the caption without re-cropping. The band slides off the mouth and onto the eyes, or off the eyes and into the platform UI. Fix: re-crop so the head sits high in the frame first, then place the caption in the space that opens below the chin.
• Leaving captions center-frame on a talking head. The middle band lands exactly where the face lives, blocking both lip-sync and connection. Fix: the middle band is never correct for a talking-head clip; default to the lower third and move to the top only when the face fills the lower half.
• Trusting auto-reframe to hold a moving speaker. A drifting tracker pulls the "safe" caption onto the mouth midway through. Fix: lock a manual reframe for clips where the speaker shifts position or two people share the frame.
• Solving the face but ignoring the buttons. A caption clear of the chin still disappears under the like button or the progress scrubber. Fix: sit the band in the lower-middle, clear of both the face and the bottom UI; verify against the live app, not the editor.
• Thin text straight over skin. Even a well-placed caption blurs into a face without separation. Fix: add a semi-opaque plate or a stroke so the words read against any frame behind them.

FAQ

How do I stop captions from blocking the speaker's face?

Re-crop so the head sits in the upper-middle of the frame, then place the caption band below the chin in the lower third. Keep the band off the two no-go regions, the eye line and the mouth, and clear of the app's bottom UI. If the face fills the lower half, move captions above the head instead.

Where should captions go if the speaker's face is large in the frame?

Put them above the head, in the top band. When a tight 9:16 crop leaves a face filling the lower two-thirds, the top is the only spot clear of both the face and the bottom buttons. It reads slightly less naturally than the lower third, but a visible face beats a perfectly natural caption that covers the mouth.

Why shouldn't captions cover the mouth specifically?

On mute, how most social video plays, a range Digiday once put as high as 85% in 2016, viewers read lips to follow rhythm and emphasis even when the audio is off. A caption on the mouth fights that read, so the clip feels off even when the words are legible. The eyes are second: they carry the connection that holds the scroll.

Should I use the lower third or the top of the frame?

Lower third by default, it reads as part of the shot and viewers' eyes already drift there. Switch to the top only when the face fills the lower half of the crop. Never use the middle band on a talking-head clip; the face lives there, and captions over it create the exact problem you're trying to fix.

Does fixing the caption position help if my clips already get no views?

Placement is one factor, not the whole story. If a clip covers the speaker's face, fixing it removes a real drag on watch time. But views depend on the hook, the moment, and the cut too, work through the full no-views diagnostic checklist rather than expecting one placement change to carry a weak clip.

For the work around placement, see how to keep clips from getting boring by adding movement and how to pick the AI-suggested clips worth captioning in the first place, a great placement on a weak moment is still a weak clip.