1. Typography that survives a small screen
Captions are read at arm's length on a phone, often in bright light, and frequently by someone who is only half paying attention. That rules out most of what works in print. You want a heavy sans-serif with open letterforms, generous size, and enough edge treatment that it never dissolves into the footage behind it. Elegance is not the goal — legibility at a glance is.
- Heavy weight. A caption that is merely bold usually is not bold enough.
- Sans-serif with open shapes. Condensed display faces lose legibility at speed.
- Big. Captions are almost always too small rather than too large.
- An outline, shadow or plate so the type never sits directly against busy footage.
2. Phrasing: how much text at once
How many words share the screen is a separate decision from how the captions look, and it changes the rhythm rather than the style. Short-form talking heads read best with one to three words landing with the voice, which creates constant motion and keeps the eye moving at speaking pace. Broadcast-style subtitles are the opposite: a full line, one or two lines deep, held long enough to read comfortably.
- One to three words at a time: short-form, vertical, high energy.
- A full line, one or two deep: explainers, interviews, anything widescreen.
- Never more than two lines on screen at once — a third line becomes a paragraph.
Changing the caption look does not change the phrasing, and changing the phrasing does not change the look. If you want both, ask for both.
3. Placement and safe areas
Where captions sit matters as much as how they look. On vertical video the bottom of the frame belongs to the platform — its own caption, the account handle, the progress bar — so anything you put there is either covered or competing. Keep captions in the central band, roughly at or just below the middle of the frame, where they sit under the speaker's face without colliding with the interface.
- Vertical: central band. Not the bottom edge.
- Widescreen: lower third is fine, since there is no platform interface over it.
- Keep captions clear of the right-hand column on vertical platforms.
- Check the actual placement on a phone, not just in the editor.
4. Contrast against moving footage
The hardest part of caption readability is that the background changes constantly. White type is invisible on a bright sky; black type disappears in shadow. The reliable fixes are all about separating the text from whatever is behind it: a thick outline, a hard offset shadow, or a solid plate behind the line. A solid block is the most readable treatment there is, which is why tutorials and busy footage tend to use it.
- Thick outline: works on almost anything and stays out of the way.
- Hard shadow: adds depth without a box, good on mid-tone footage.
- Solid block or plate: maximum readability, and a natural place for a brand colour.
- Glow: reads well on dark footage, disappears on bright.
5. Choosing a style family
Cubix groups its caption looks into families, and picking the right family is most of the decision. The families exist because the same caption treatment cannot serve an entertainment hook and a client interview. Choose the family from the content, then choose within it by what suits the footage.
- Viral: loud boxed or highlighted key words, built for short-form hooks.
- Highlight: a coloured marker or pill behind the active word — energetic but not shouting.
- Bold: heavy type, no colour trick. Good for podcasts and strong delivery.
- Minimal: clean, broadcast-looking. The right family for client, corporate, interview and documentary work.
- Keywords: the line stays calm and only the meaningful word changes colour. The editorial short-form look.
- Block: text on a solid strip. The most readable option and the natural home for a brand colour.
- Aesthetic, neon, retro, fun: lowercase moods, glowing type, throwbacks and novelty, for content whose style calls for them.
New captions arrive in a loud default style. If the video is restrained — an interview, a client piece, anything professional — name the style you want in the same breath as asking for captions.
✖ Using a full sentence per caption on vertical video
Why it fails: Long lines shrink the type and cover the speaker, and the eye cannot keep up at speaking pace.
✔ Better approach: Show a few words at a time so each one lands with the voice.
✖ Thin or decorative fonts
Why it fails: Light weights and elaborate letterforms vanish against footage at phone size.
✔ Better approach: Use a heavy sans-serif with a clear outline or plate behind it.
✖ Placing captions at the very bottom of a vertical frame
Why it fails: That region is occupied by the platform's own interface.
✔ Better approach: Move them into the central band of the frame.
✖ Leaving the default loud style on serious content
Why it fails: A bright box undercuts the authority of an interview or a client video.
✔ Better approach: Pick a restrained style by name for professional work.
Execute this workflow automatically in Cubix
Generate word-timed captions and pick from a full library of looks, from broadcast-clean to full viral.
Frequently Asked Questions
Common questions around this editing workflow.