On-Screen Text Too Small on Phones? Size and Contrast Fixes
Set on-screen text to at least 5% of the frame's height, roughly 54px tall on a standard 1080×1920 vertical clip, use a bold or heavy weight, and put a contrast separator behind it. Text that looks fine on your laptop is usually too small on a phone because your editing canvas is four to six times larger than the screen people actually watch on. Size first, then contrast.
That last sentence is the trap. Most "unreadable text" problems are sizing problems wearing a contrast costume. You stare at a 27-inch monitor, the caption looks balanced, you ship it, and on a 6-inch phone held at arm's length it shrinks to something the viewer has to squint at, so they scroll. Contrast matters, and we'll get to it, but it can't rescue text that's simply too small. Fix the scale and half of these clips become readable on their own.
Why text size decides whether the clip lands
Most short-form video plays without sound, so the text on screen carries the message, and if it's too small to read in the first second, the message never arrives. The commonly cited mute rate sits in the 75–85% range: it's publisher- and survey-reported and directional, not one audited number, but the direction isn't in dispute. Ad-tech firm Sharethrough put it at about 75%, and Digiday reported as far back as 2016 that about 85% of Facebook video is watched without sound. When the audio is off, undersized text is a silent clip with the subtitle illegible.
There's a measurement reason this keeps happening. A 1080×1920 vertical clip is enormous on your monitor and small in a feed. The same caption that fills a comfortable strip on your screen renders at a fraction of a phone's physical height when someone watches it at a normal viewing distance. Editors design for the canvas; viewers watch on the phone. The gap between those two is where readable text goes to die.
What size should on-screen text be? The minimum that survives a phone
Set body captions to a minimum of 5% of the frame height, and treat that as a floor, not a target, 6–8% reads better for fast-talking clips. On a 1080×1920 vertical frame, 5% is about 96px of frame height, which lands a bold cap-height of roughly 54–60px once you account for line spacing. The rule scales: it's the percentage that matters, because it stays correct whether you export at 1080p or 720p.
Working in percentages instead of fixed pixels is the single change that fixes this for good. A 48px font is large on a 720-wide clip and small on a 1080-wide one; "5% of frame height" is the same readable size on both. Size by ratio, export at any resolution, and the text holds. (If your text is sharp in the editor but mushy after upload, that's a separate export problem, see why clips come out blurry or low-res.)
This isn't arbitrary. Broadcast subtitling has carried minimum-size rules for decades, the BBC Subtitle Guidelines specify a minimum text height relative to the screen so subtitles stay legible on small displays. Phones are the smallest display most of your clips will ever play on, so the floor has to assume the worst case: a mid-size phone, held at arm's length, in a moving feed.
The arm's-length test: the only check that matters
There's one test that catches undersized text every time, and it takes ten seconds. Export the clip, send it to your own phone, hold the phone at a normal arm's-length viewing distance, not up close, and read the text without leaning in. If you have to squint or pull the phone closer, it's too small. Your editor's preview at 100% zoom is lying to you; the phone is the source of truth.
Do this with the harshest moment in the clip, not the calmest. Find the frame where the speaker is mid-gesture against a busy background and check that one. Text that survives your worst frame survives the whole clip. If it only works on the clean intro shot, it isn't done.
Size by role: a rule for every element on screen
Not all on-screen text is the same job, so it shouldn't be the same size. A spoken-word caption, a hook headline, and a "@handle" label have different reading priorities. Size them in that order, and never let the lowest-priority text drop below the floor where it becomes decorative noise nobody can read.
The honest implication: if you can't fit a label at a readable size, you have too much text on screen. Undersized text is often a symptom of cramming, three lines of caption, a handle, a logo, and a sticker all fighting for the same lower third. Cut the line that's least important and size up the rest. A clip with two readable elements beats one with five illegible ones.
Size is a separate question from where the text sits. You can size everything correctly and still lose the caption behind a platform's UI buttons if it's parked in the wrong part of the frame, that's the safe-zone problem, and it has its own fix in our guide to caption placement and platform safe zones. Get the size right here, then make sure it lives where every app will actually show it.
Size handles half of it, contrast handles the rest
Once the text is big enough, contrast is what keeps it readable over a moving background. Bump the size first, then make sure the letters separate from the footage: a near-white fill (#F5F5F5), a dark stroke at 8–12% of cap height, and a drop shadow or a semi-opaque box for the hardest frames. A large caption with no separator still vanishes the moment the speaker walks past a window.
For accessibility, captions count as "large text," so the WCAG 2.2 contrast minimum is only 3:1 at AA, and white text on a black stroke or box clears it by a mile, landing near 21:1 against the separator. You won't fail contrast if you use a separator. You'll fail it only by skipping one and letting raw text fight the frame. The full breakdown of stroke, shadow, and box settings, and which to use over which footage, lives in our caption color and contrast guide. This article's job is the size; that one's job is the separation. Together they're the whole fix.
Weight matters as much as the separator here. A thin font at the right size still reads poorly on a phone because the strokes of the letters are too delicate to survive both the small screen and the platform's re-compression. Use a bold or heavy weight for captions, never light or regular. Picking the right typeface and weight is its own decision; see the best caption fonts for podcast clips.
Common mistakes (and the fix)
These are the sizing failures I see most when reviewing clips, in order of how often they sink an otherwise good edit.
- Designing on a laptop, never checking on a phone. The number-one cause. The editor canvas is four to six times larger than the viewing screen, so text that looks right is small in the feed. Fix: run the arm's-length test on every clip before export, send it to your phone, no zooming in.
- Fixed pixel sizes instead of percentages. A font set to 48px is fine on a 720-wide clip and tiny on a 1080-wide one. Fix: size text as a percentage of frame height (5% floor for body captions) so it scales across resolutions.
- Too much text on screen at once. Three caption lines plus a handle plus a logo forces everything small. Fix: keep captions to one or two lines, cut the least important element, and size up what's left. Reading speed and line length have their own limits, that's a related fix in the boring-clips checklist of movement and pacing.
- Thin fonts at small sizes. A delicate weight loses its strokes on a phone and after compression. Fix: use bold or heavier for any caption; reserve light weights for large headlines only, if at all.
- Sizing for the clean frame, not the busy one. Text looks readable over the studio intro and disappears against the busy mid-clip background. Fix: set size and contrast for the worst frame in the clip, then apply it throughout.
- Trusting the platform to scale it up. It won't. The platform displays your text at the proportion you set. Fix: the size you export is the size viewers get, make it large at the source. If your clips are getting no traction, undersized text is one of the first things to rule out; work through the full no-views diagnostic checklist.
FAQ
Why does my video text look fine on my computer but too small on my phone?
Because your editing canvas is far larger than a phone screen. A 1080×1920 clip is big on a 27-inch monitor and small in a feed, so text that looks balanced on the canvas shrinks at a normal phone viewing distance. Size the text as a percentage of frame height (5% floor) and always check the export on an actual phone.
What font size should captions be for vertical video?
A minimum of 5% of frame height for body captions, which is roughly 54px of cap-height on a 1080×1920 clip; 6–8% reads better for fast-talking content. Size in percentages rather than fixed pixels so the text stays correct whether you export at 1080p or 720p. Use a bold weight, never thin.
How do I make text readable on a busy background?
Size is the first lever, make the text large enough to read instantly. Then add a separator: a near-white fill with a dark stroke at 8–12% of cap height, plus a drop shadow or a semi-opaque box for the hardest frames. Set both for the busiest frame in the clip, not the cleanest one.
Should I use pixels or percentages for caption size?
Percentages of frame height. A fixed pixel size is large on a low-resolution export and small on a high-resolution one, so it breaks the moment you change export settings. "5% of frame height" stays the same readable proportion at any resolution, which is why it's the reliable floor.
Does making text bigger hurt the clip's design?
Only if the frame is already crowded. If bigger text feels like too much, the real problem is usually too many on-screen elements competing for the lower third. Cut the least important label, keep captions to one or two lines, and the remaining text has room to be both large and clean.