Subtitle Formatting Best Practices: Line Length, Reading Speed and Timing

The best subtitles are the ones viewers never think about. They appear when the speech starts, they can be read comfortably before they vanish, they break lines where a person would naturally pause, and they stay out of the way of the picture. Achieving that is not magic. It follows a set of fairly consistent rules that broadcasters, streaming services, and professional subtitlers have developed over decades.

This guide collects those formatting best practices in one place and explains the reasoning behind them, so you can apply them with judgment instead of memorising numbers. Whether you are polishing a merged file, preparing captions for your own video, or reviewing a translation, these guidelines will help you produce text that is easy on the eyes.

A note before we start: there is no single universal standard. Different organisations publish slightly different limits, and the right numbers depend on your audience and platform. Treat the figures below as widely used starting points, and check the current style guide of any service you deliver to.

Why the rules exist

A viewer watching a subtitled film is doing two jobs at once: following the image and reading the text. Every extra word, every awkward line break, and every subtitle that vanishes too quickly takes attention away from the picture. The guidelines are all ways of protecting that attention. Keep that principle in mind and most decisions become easier.

Line length

The most common recommendation is a maximum of about 42 characters per line, spaces and punctuation included. This number comes from how much text fits comfortably across a standard screen at a readable size without crowding the edges. On narrow phone screens, shorter lines are easier still.

Alongside that, keep subtitles to two lines at most. A third line covers too much of the picture and takes longer to read than the speech lasts. If the dialogue is too long for two lines, split it across two consecutive subtitles instead.

Reading speed

Reading speed is measured in characters per second, usually shortened to CPS, or in words per minute. A widely used target for adult programmes is around 17 characters per second, and children's content is usually set lower because younger viewers read more slowly. Some publicly available guidelines for broadcast give a range of roughly 160 to 180 words per minute.

To check a subtitle, count its characters and divide by the number of seconds it is displayed. A 40-character line shown for two seconds has a reading speed of 20 characters per second, which is fast. Showing it for 2.4 seconds brings it down to about 17.

Many subtitle editors calculate this automatically and highlight the lines that are too fast. If a line is too fast, you have three options: lengthen the display time if the neighbouring subtitles allow it, shorten the text, or split it into two subtitles with their own timing.

Duration

Every subtitle needs a minimum time on screen, or it will flicker past before it can be read. A frequently cited minimum is about five-sixths of a second, and many people use one second as an easy rule. At the other end, a subtitle should not stay up too long, or viewers will start to wonder whether it has frozen. A maximum of about seven seconds is typical.

Short interjections, such as "No!" or "What?", are the usual exception to the minimum, and even then it is wise to hold them for at least a full second when the surrounding dialogue allows.

Timing relative to speech

The text should appear when the speech begins, or a fraction of a second after, and disappear at about the moment the speech ends. Slightly late is better than slightly early, because a subtitle that appears before the line is spoken spoils the delivery of jokes and surprises.

You can let a subtitle linger for a short time after the speech ends, particularly if it is long, but avoid leaving it on screen through a cut to a new scene, because viewers will think it belongs to the new scene. When a subtitle must cross a shot change, many guidelines suggest ending it just before or on the cut.

Gaps between subtitles

Leave a small gap between consecutive subtitles, so the eye registers that the text has changed. Professional guidelines often specify a minimum gap of two frames, which is about 80 milliseconds at 24 frames per second. Without a gap, two subtitles that follow one another can look like one subtitle that has simply changed its words.

Line breaks

Where you break a line is one of the most noticeable differences between amateur and professional subtitles. The goal is to break at a natural pause so each line forms a unit of meaning. A few practical rules:

• Break after punctuation when you can, such as after a comma or at the end of a clause.
• Keep grammatical units together. Avoid separating an article from its noun, a preposition from its object, an auxiliary verb from its main verb, or a first name from a surname.
• Prefer a pyramid shape. When two lines are needed, a shorter top line over a longer bottom line (or two lines of similar length) looks tidier than a long top line and a tiny bottom one.
• Do not leave a single word on a line unless it is the end of a sentence.

Compare these two ways of breaking the same subtitle:

I told him that we would
meet at the station before noon.

I told him that we would meet
at the station before noon.

The second version keeps "we would meet" together and flows more naturally when read quickly.

Punctuation and style

Subtitles use ordinary punctuation, with a few conventions.

• Ellipses for trailing off or hesitation. Use three dots, and use them sparingly.
• Dashes to show a change of speaker within the same subtitle. A common style puts a hyphen at the start of each speaker's line, with each speaker on their own line.
• Quotation marks only when quoting something, not for dialogue itself.
• Numbers. A common convention is to spell out numbers from one to ten and use digits for larger ones, but follow your style guide if you have one.
• Consistency. Choose a spelling variant and stick to it, for example colour versus color.

Avoid unusual symbols that some fonts or players cannot display. Stick to characters your text encoding supports. If a file will be viewed on many devices, UTF-8 is the safe choice; see our encoding guide.

Italics, bold and colour

Italics are the most useful and most widely supported styling. They are conventionally used for:

• voices from off screen or through a device, such as a phone or radio,
• thoughts and narration,
• foreign words that are not translated,
• titles of works.

Use <i> and </i> tags in an SRT file. Bold and underline are rarely necessary. Colour is sometimes used to identify speakers, but it is poorly supported on televisions and web players and does not work for viewers who override subtitle colours for accessibility. If you depend on colour, add a back-up such as a speaker label.

Speaker identification and sound

If you are creating captions for deaf and hard-of-hearing viewers, you need more than dialogue. Add speaker names when the speaker is off screen or unclear, and describe meaningful sounds in square brackets, such as [door slams] or [phone ringing]. Music is often marked with a note symbol or a bracketed description. For standard translation subtitles aimed at hearing audiences, sound descriptions are normally left out, and a speaker name is only added when confusion is likely.

Position

By default, subtitles sit at the bottom centre of the screen. Moving them to the top is acceptable when important text or action is at the bottom, such as on-screen captions in a news report. Plain SRT does not define positioning, so if you need it, you may need a format that does, as explained in SRT vs VTT vs ASS.

A checklist for reviewing a file

Run through this list before delivering a file:

1. No line longer than about 42 characters, and no more than two lines.
2. Reading speed at or below your target for every subtitle.
3. No subtitle shorter than the minimum or longer than the maximum.
4. A small gap between consecutive subtitles.
5. Lines broken at natural pauses.
6. Consistent spelling of names and terms.
7. Italics only where they have a meaning.
8. Numbering starts at 1 and runs without gaps.
9. File saved as UTF-8 with a .srt extension.
10. The file tested in a real player with the video.

Many of these checks can be automated. Subtitle Edit, for instance, offers a "fix common errors" function and a list of lines that break your chosen limits. Use automation to find the problems and your own eyes to decide how to fix them.

When rules collide

Sometimes rules conflict. A line that is accurate may be too long. A good line break may produce an uneven shape. A speaker may talk so fast that no subtitle can keep to the reading speed. In those cases, protect the meaning first, then readability, then neatness. Condense by cutting words that add little, such as repeated phrases and filler, but never cut information that the viewer needs to follow the story.

How this relates to merging

When you merge two translations with the merge tool, apply these same criteria to decide between versions. A line that fits the time available is better than one that needs a speed-reading champion, even if both are accurate. Check the final file at the end, particularly the places where lines from the two sources meet, since the combined rhythm can differ from what either translator planned. Our article on comparing subtitle translations shows how to build that into your review.

Summary

Readable subtitles are short, well timed, and broken at natural pauses. Aim for about 42 characters per line, two lines at most, a reading speed around 17 characters per second, a duration between roughly one and seven seconds, and a small gap between subtitles. Use italics sparingly and consistently, identify speakers only when needed, and always test the result in a player. The numbers are guidelines, not laws, but they exist because they protect the viewer's attention, which is the one resource every subtitle depends on.