Caption Placement So Subtitles Never Cover On-Screen Text

Your captions sit at the bottom of the frame, and then a lower third slides in with a guest's name and title, and for four seconds the viewer can read neither. Poor caption placement turns two useful pieces of text into one unreadable smudge, and it happens on almost every interview, tutorial and product video that uses on-screen graphics.

The fix is not complicated, but it does need a plan before you export. I manage captions for an archive of roughly 3,000 videos, and placement errors are the single most common thing our reviewers flag after spelling. This guide covers the default positions, when to move captions, how each file format handles position, and a quick review routine that catches collisions before your audience does.

Default caption placement and why it exists

The standard position is bottom centre, inside the title safe area. On a 1920 by 1080 frame, that means keeping caption text above roughly the bottom 10 percent of the picture, so the lowest line sits around 970 pixels from the top. Broadcast guidelines and most streaming style guides agree on this for a simple reason: viewers' eyes return to the same spot every time a new caption appears, and a consistent home position lowers reading effort.

Keep to two lines whenever you can. A third line pushes the block up into the face of whoever is speaking in a typical medium shot. If you find yourself needing three lines often, the problem is usually line length and timing, which I cover in the guide to caption reading speed and line length.

When to move captions off the bottom

Moving a caption is the exception, and every move should have a reason. These are the situations that justify one.

• Lower thirds and name straps. Move the caption to the top of the frame for the duration of the graphic, then return it to the bottom.
• Burned-in text in the footage itself, such as a screen recording with a menu bar at the bottom, a chyron in news footage, or a product label being read aloud.
• Action at the bottom of the frame, like hands on a keyboard in a tutorial or feet in a dance clip, where the caption would hide what the video is teaching.
• Platform interface overlays. Vertical video apps place buttons, usernames and descriptions over the lower 20 to 25 percent of the screen.

When you move to the top, keep the same title safe margin, so the first line starts around 10 percent down from the top edge. Do not float captions in the middle of the frame unless you are positioning them next to a specific speaker in a crowded scene.

Vertical video needs its own rules

For a 1080 by 1920 short, the bottom home position used on horizontal video is usually covered by the app's caption and username. A safer default for vertical clips is about 60 to 70 percent of the way down the frame, roughly 1,150 to 1,350 pixels from the top. Also keep text clear of the right edge, where like and share buttons stack. Test on an actual phone, not just your editor preview, since the overlays differ between apps and change every few months.

How file formats handle caption placement

Not every format can carry position data, and this is where a lot of careful work gets lost on upload.

In WebVTT, a cue with the setting line:10% places it near the top, while line:85% keeps it near the bottom. If you deliver only SRT, any placement you set in your editor will disappear the moment a platform reads the file. The guide to choosing a caption file format that survives re-uploads covers which hosts keep what.

Why "always burn them in" is the wrong answer to placement problems

A common suggestion in editing groups is to stop worrying about placement and burn the captions into the picture, where you control every pixel. It sounds tidy. In practice it trades one problem for three.

Burned-in text cannot be turned off, resized, or restyled by a viewer who needs larger type or higher contrast. It cannot be translated without a full re-export. And when a platform later adds its own overlay, your carefully placed text ends up under a button anyway, with no way to move it. Open captions have a place, mainly in muted social clips, but for most videos a positioned WebVTT file does the job better. The longer argument is in why burned-in subtitles are usually the wrong choice.

A ten minute placement review before export

This is the routine our reviewers use on every file. On a 12 minute interview it takes about ten minutes.

1. List every on-screen graphic with its in and out time. Most editors can export markers, or you can scrub the timeline at 4x speed and note them.
2. Check each graphic's window in the caption track. Any cue that overlaps a lower third by even half a second gets moved to the top.
3. Split cues at graphic boundaries. If a caption starts before a name strap appears and ends after, split it so only the overlapping part moves. Jumping text mid-sentence is distracting, so cut at a natural pause.
4. Watch with captions on at full screen and again at a small player size, since overlays scale differently.
5. Preview on the target platform using an unlisted or private upload before you publish.

Keep a note of which graphics package your team uses. If lower thirds always occupy the same region, you can build a preset position and apply it in seconds next time.

Frequently asked questions

Where should captions be placed on a video?

Bottom centre, inside the title safe area, is the standard for horizontal video. Move captions to the top only when something important occupies the bottom of the frame, such as a lower third or on-screen text. For vertical video, place them higher, around 60 to 70 percent down, to avoid app overlays.

Can you change caption position in an SRT file?

Not reliably. The SRT standard has no position field, and while a few players read unofficial tags, most platforms ignore them. If position matters, export WebVTT or TTML instead, both of which store placement in the file.

Should captions move when a speaker changes?

Usually not. Viewers read faster when captions stay in one place, so speaker changes are better shown with a dash or a name label. Positioning captions beside each speaker is only worth the effort in crowded scenes where it is otherwise unclear who is talking.