DEV Community

Cover image for How to add captions to a vertical video so TikTok and Reels buttons don't cover them
Hanako H.
Hanako H.

Posted on

How to add captions to a vertical video so TikTok and Reels buttons don't cover them

Most editors put captions near the bottom of the frame, about where they sit on a TV. On a vertical feed that spot is taken. TikTok, Instagram Reels and YouTube Shorts draw the account name, the post text and the music line across the bottom of the video, and a column of like, comment and share buttons down the right side. A caption placed at the usual height ends up underneath all of that, and the viewer sees half a sentence.

The fix is to know where the covered band is before you place anything, and to set the caption line above it. The steps below use the video editor in the osFoundry desktop app, which shows that band on the preview and then burns captions in above it. The clip is a half-minute answer from NASA's 2020 call with the crew of the International Space Station, already cropped to 9:16.

Turn on the safe area guide

Open the video editor in a 9:16 project and add the clip with "Add a shot from a file" above the Shots list. Under the preview, click Safe area.

The video editor with the Safe area guide on: the dashed box is the part of the frame a feed leaves alone

The shaded edges are what a vertical feed draws over. For 9:16 the app marks the top 8% of the frame, the bottom 20% and 6% down each side. On a 1080×1920 video that is 154 pixels at the top, 384 at the bottom and 65 on each side. Anything that matters, a face, a product, a line of text, belongs inside the dashed box.

These figures fit ordinary posts. Ads get more UI on top of them: Meta's own guide for Reels ads asks you to keep the top 14%, the bottom 35% and 6% of each side free of text and logos, because the call-to-action button sits there too.

Transcribe the clip

Select the clip on the timeline and click Transcript in the timeline toolbar. A panel opens under the editor with a single button.

The transcript panel before the first run, with the Transcribe button

Click Transcribe. Only the sound is sent, to the transcription model behind your own API key, and every word comes back with its start and end time. The panel says it in plain words: transcribing uses your model, while editing, captions and clip suggestions are free.

Burn the captions in

When the words appear, click Captions in the bar above them.

Captions clicked: the toast reports 13 captions burned in

The app first groups the words into caption lines. A line takes at most 38 characters and at most 4 seconds, and any pause longer than 0.9 seconds starts a new line, because a pause usually means a new sentence. Then ffmpeg burns the lines into a new file, and the line height comes from the safe area of the frame rather than a fixed margin: for 9:16 the captions are set 72% of the way down, above the covered fifth.

For this clip that made 13 captions, burned in about 5 seconds. The toast confirms it: "13 captions burned in, clear of the 9:16 safe area." The original file is left alone. The captioned copy is written next to it with .captioned.mp4 added to the name, and the caption lines are saved beside it as an .srt file, which is handy if you'd rather upload captions separately on YouTube.

Before you post

Play the result once with the platform's own preview, since apps move their buttons between versions and between ads and organic posts. If a caption runs to two lines in a large font, its top edge climbs toward the middle of the frame, so short lines matter as much as the baseline.

The editor runs on your own machine, as described on the page about osFoundry's local-first design, and the app is free from the osFoundry download page for macOS, Windows and Linux.

Top comments (0)

⌂ 换网址