Why captions stay at two lines

Captions stay at two lines because a third line collides with platform chrome. TikTok, Reels, and Shorts already own the bottom of the 9:16 frame. The two-line cap is safe area, not a taste rule. Split the cue. Do not shrink the type until a paragraph fits.

Chrome is the third line

Broadcast parked two-line captions in a band that 16:9 could spare. Vertical feeds parked interface in that band: username, caption field, likes, share, shop, a CTA on paid cuts. YouTube Shorts, as of the 27 May 2026 YouTube Blog post on AI labels, can also draw a photoreal / meaningfully AI-altered overlay on the picture. That is more furniture, not more room for type.

A three-line block is not “more complete.” It is a rectangle that occupies the same pixels as the UI. The platform will win. Your third line becomes a username. Your second line becomes a caption field. The face you framed for the hook sits under a paragraph.

This is why the two-line limit survived the move from television to phones. Television had overscan. Phones have buttons. The constraint changed names. The geometry did not get kinder.

Burned-in captions make the collision permanent. A closed-caption player might restack a long cue. Pixels will not. If you burn three lines, every view is three lines, including the ones where chrome covers two of them.

A lower third, a legal row, and karaoke pop are also lines. Two-line speech plus a name bar is already three. Two-line speech plus a disclaimer is already three. The cap is per frame, not per tool. Count everything that is type in the lower band.

Split, do not shrink

Auto captions emit a 90-character cue when the speaker does not pause. That is a paragraph waiting for an editor. The wrong fix is a smaller font. At phone size the paragraph is still a paragraph, and now it is also too small. The right fix is a split at a clause.

Latin-script practice has long used a band around 32–42 characters per line, two lines, a hold long enough to read. That is a guideline, not a file-format law. CJK scripts count glyphs differently. The geometry is the same: if the block enters the chrome, it is too tall, whatever the script.

Do not hyphenate names, SKUs, or URLs to win a line. Hold the string and split somewhere else. A short second line is legal. A clever wrap that breaks “New York” or “15.99” is not.

Reading speed is the other half of the split. Two lines that stay for half a second still fail. Two lines that stay long enough to finish, then yield, is the object. Reading speed is characters per second versus hold. The two-line cap is height versus chrome. You need both.

All-caps and emoji push you into a third line without adding a third sentence. Caps are taller. Emoji is wide. Both are how a two-line transcript becomes a three-line burn. Sentence case, no sticker, then split.

DYNAMIC word-pop does not grant a third line. The live token scales. Scale plus two full rows is a collision with the mouth and with the UI. If you need karaoke, you need shorter cues, not taller ones.

Count the other graphics

Before you burn, list what else will sit in the lower third of this file:

• Platform chrome (assume it; recheck current UI)
• Paid CTA or shop row
• Legal / disclosure / Article 50 mark
• Name super
• YouTube Shorts AI overlay when the cut is photoreal generated or meaningfully altered

Each of those is a reason to use one line of speech, or to park speech higher, or to sequence the graphics so they do not share a frame. “Two lines” is the maximum for the caption track in a clean organic talking-head. It is not a budget you still have after a chyron and a disclaimer.

Headline-style title rows belong at the top, briefly, or not on a speech cut. They are not a way to sneak a third line above the chrome. Top chrome exists too: search, status, sometimes a banner. Safe area is a hole in the middle, not a promise about the edges.

Reframe changes the cap. A 16:9 webinar with comfortable two-line subs becomes a 9:16 clip whose two lines now sit in the TikTok caption field. Caption after crop. Recaptioning is cheaper than arguing with a comment that the words were “under the buttons.”

The mute geometry pass

Lock picture and ratio. Transcribe. Split every cue that is more than two lines or more than a finishable hold. Burn. Then watch on a phone, in the platform preview if you have it, with the sound off.

You are looking for three failures:

1. A line under a username, caption field, or CTA.
2. A block covering the mouth or the product.
3. A wrap you did not see on the desktop timeline.

If you see (1), you did not have two lines. You had two lines plus chrome. Move up or cut a line. If you see (2), the block is too tall or too low; two lines can still cover a face if the talent is framed tight. If you see (3), the font, caps, or emoji lied about width.

The AI caption generator and add captions to a video will burn the transcript they are given. 165 language codes, 21 BASIC at 1×, 9 DYNAMIC at 2×. None of those settings is a three-line mode. Caption styles that look big in a preview still have to live in the hole the UI leaves.

Two lines is not aesthetics. It is the number of rows a 9:16 feed will typically let you keep after it has taken its cut. Split the cue. Leave the type large enough to read. Let the third line go.

Keep the next burn to two lines in Versely.

FAQ

Why not three short lines instead of two longer ones?

Height is the scarce resource, not character count. Three rows occupy the chrome band even when each row is short. Split into two cues over time, not three rows at once.

Does shrinking the font let me keep a paragraph?

No. At phone size the paragraph is unreadable and still tall enough to hit UI. Split at a clause. Keep the type large.

Do YouTube Shorts change the two-line rule?

They tighten it. Shorts already have bottom chrome; as of 27 May 2026 photoreal AI labels overlay the picture as well. Two lines, higher than the overlay, or one line if a legal row is also in the frame.