You picked a caption colour that looked great in the first shot. Then the clip cut to a bright window, or a white shirt, or a patch of sky, and the words disappeared for a second. That isn’t bad luck with one colour. Every text colour has a background it can’t be read on, so the colour on its own is never the whole answer.
The short answer. There’s no colour that reads on every background, so don’t rely on the colour alone. Use white or yellow text with a black outline for most videos, and switch to a dark box behind the text when the footage is bright or busy. What makes a caption readable is a difference in brightness between the letters and whatever sits right next to them, not the hue.
Why no single colour works everywhere
The eye picks out text mainly by brightness, not by colour. A yellow word on a light grey wall can look vivid in a still, but the two are close in brightness, so the edges of the letters are hard to find at a glance. A dull grey word on a black background reads instantly, because the brightness difference is large.
Short-form video makes this worse because the background never sits still. A talking-head clip might have a dark room behind you for ten seconds and a bright window for the next three. White text works in the first part and fails in the second. Black text does the opposite. Any colour in between is a compromise that fails a bit everywhere.
This is why the colour choice and the edge choice belong together. The colour sets how the caption looks. The outline, shadow or box is what keeps it readable when the shot changes underneath it.
Contrast is measurable
Web accessibility guidance gives a useful reference point. The W3C’s Web Content Accessibility Guidelines ask for a contrast ratio of at least 4.5:1 between normal text and its background, and 3:1 for large text. That ratio is worked out from the brightness of the two colours, which is why yellow on white scores badly even though they’re different colours. Caption text over video can’t be measured against one fixed background, but the principle carries over: if the letters and what’s touching them are close in brightness, the words will be hard to read.
Light text with a dark edge
For most videos, the safest setup is light text with a dark edge round every letter. White or yellow letters with a black outline, a shadow, or both.
It works because the outline brings its own background. Whatever is behind the caption, every letter has a ring of black next to it, so there’s always a large brightness difference at the edge where the eye looks for one. Over dark footage the outline blends in and the bright letters do the work. Over bright footage the outline does the work.
How thick the outline should be
Thin outlines look tidy in an editor and vanish on a phone. As a starting point, make the outline a noticeable fraction of the letter stroke, thick enough that you can see it clearly when you hold the phone at arm’s length. If you can only see it when you zoom in, it won’t survive the platform’s compression. A heavy font gives the outline more to wrap around, which is one more reason to pick a bold or black weight.
White or yellow
White is the neutral choice and goes with any brand. Yellow is as bright as white on most screens but stands out from white objects in the shot, such as walls, sky and clothing. Broadcast subtitles have long used both. A common pattern on short-form video is white for the line and yellow for the word being spoken.
When to use a box instead
An outline has limits. Over very busy footage, like a crowd, foliage, patterned fabric or fast motion, the outline is surrounded by so much detail that the text still feels noisy. That’s when a solid or semi-transparent box behind the caption earns its place.
A box swaps the background out for one you control. Dark box, light text, and the contrast is the same on every frame of the video. The trade-off is that it covers more of the picture, so keep the padding tight and the corners rounded if you want it to feel lighter.
A rough rule:
- Outline only: talking heads, steady shots, footage with a mostly even background.
- Outline plus shadow: mixed footage that swings between light and dark.
- Box: bright scenes, busy scenes, screen recordings, anything with text already in the frame.
In Capster, text colour and outline colour and width are in the style panel on the free tier, and the caption background sits with the Pro controls. The caption styles page lists which presets ship with a box and which use an outline.
Colours that hold up after upload
TikTok, Instagram and YouTube all re-encode your video, and that process treats colours unevenly. Brightness detail is kept at full resolution, while colour detail is stored at lower resolution. So white, yellow and other light colours keep crisp edges, and saturated reds, blues and magentas tend to smear at the edges of letters. The full reason is in why captions look blurry after upload.
In practice:
- Keep the letters themselves light: white, yellow, pale cyan, pale green.
- Put strong brand colours in the highlight block, the box or the outline, where a soft edge matters less.
- Avoid pairing two saturated colours, like red text on a blue box. Even when the hues clash, their brightness can be close.
Picking a highlight colour
Word highlight, where the spoken word changes colour or gets a coloured block behind it, adds a second colour to think about. The highlight needs to differ from the rest of the line, and both need to read against the background.
Two things keep it working for everyone:
- Change brightness, not only hue. Red-green colour blindness is the most common kind, so a red active word in a line of green text may not stand out at all for some viewers. Yellow against white, or white against a dark coloured block, changes brightness or shape, which everyone can see.
- Let a block carry strong colours. A pink or red block behind a white word gives you the brand colour without making the letters themselves saturated.
A quick colour check before you export
| Check | Why |
|---|---|
| Text is white, yellow or another light colour | Light letters keep sharp edges after compression |
| Every letter has a dark outline, shadow or box | The edge brings its own contrast to any shot |
| Outline is visible at arm’s length | Thin outlines disappear on a phone |
| Box used over bright or busy shots | One controlled background beats a noisy one |
| Highlight changes brightness, not just hue | Readable for colour-blind viewers |
| Checked on the brightest shot in the clip | That’s where captions usually fail |
The last row matters most. Scrub to the brightest, busiest moment in the video and read the caption there. If it holds up, it’ll hold up everywhere else. Then burn the subtitles into the video and check placement against each platform’s buttons, starting with where TikTok puts its interface in captions for a TikTok video.