Italics, Dashes, Ellipses and Numbers: Subtitle Typography Explained

Why subtitle typography has its own conventions

Subtitles give viewers a few seconds to take in text while watching a picture, without the visual cues a printed page offers. Typography carries information that would otherwise need extra words: italics say a voice comes from somewhere else, a dash says a new person is talking, an ellipsis says a sentence was left hanging. Because viewers learn these signals unconsciously, inconsistency is noticeable even when nobody can name the rule that was broken.

The conventions also differ by language and by organization. A subtitle style guide, or the subtitle section of a broader translation style guide, records the choices for a project so that every translator and reviewer applies them the same way. If you do not have one, creating a translation style guide shows how to write a short one.

Italics and when to use them

Italics are the most heavily used formatting in subtitles, and the most heavily debated. The uses that appear in most guides:

• Voices from outside the scene: narration, voice-over, a character's thoughts, flashback voices heard over another scene.
• Voices coming through a device: telephone calls, radio, television, public address and intercom, when the speaker is not in the room.
• Song lyrics, in styles that italicize them as well as marking them with music notes.
• Titles of books, films, albums and similar works, following the print conventions of the language.
• Foreign words that are not established in the subtitle language, used sparingly.

A frequent point of confusion is the speaker who is simply out of shot. Many guides reserve italics for voices that are not physically present in the scene, and leave a speaker who is in the room but off camera in normal type. Others italicize any off-screen speaker. Pick one rule and apply it consistently.

Italics for emphasis are discouraged in many styles, because the same formatting already means an off-screen voice. Where emphasis is essential to the meaning, guides often allow it for a single word.

Dialogue dashes for two speakers

When two people speak within one subtitle, each speaker gets their own line, and a dash marks the change. Conventions differ on the details:

Dash on both lines

Common in many English-language styles: each line starts with a hyphen

Dash on the second line only

Used in some European traditions, where the first speaker's line has no dash

Space after the dash

Some guides require a space after the hyphen, others forbid it

One sentence per speaker

A frequent rule: each speaker gets a single, short line within the shared cue

A shared cue only works when both lines are short and the exchange is quick, such as a question and a one-word answer. If either speaker says more, give each their own subtitle. Do not use a dash for a single speaker continuing on a second line; that line break is a matter of segmentation, covered in subtitle segmentation.

Ellipses, pauses and continuing sentences

An ellipsis signals that speech trails off, hesitates or is interrupted. Typical uses are a sentence the speaker abandons, a noticeable pause in the middle of a line, or a voice that fades away.

An older convention also put ellipses at the end of a cue and the start of the next when a sentence ran across two subtitles. Many current guides have dropped this, because the absence of final punctuation already tells the viewer the sentence continues, and the extra dots take up space. Some guides use a dash instead of an ellipsis for an abrupt interruption, keeping the ellipsis for trailing off.

Guides also differ on whether to use the single ellipsis character or three periods, and on whether a space precedes it. The choice matters most when a platform counts characters for reading-speed checks, since the single character counts as one.

Numbers, dates and units

Number conventions in subtitles usually borrow from print style, with an eye on space:

• Many English guides spell out one to ten and use numerals above ten.
• A number that starts a sentence is often spelled out, or the sentence is rephrased.
• Times, dates, measurements, money and percentages typically use numerals because they are shorter and easier to scan.
• Decimal separators, digit grouping and date order follow the target language: a translation into German or French should not keep the English decimal point.
• Units may be converted for the audience, which is a translation decision; record it in the style guide.

Whatever the rule, consistency inside a file matters more than the rule itself. A subtitle that says "five" in one cue and "5" in the next looks careless.

Capitalization and punctuation

Subtitles use normal sentence case. All capitals are generally avoided for dialogue, including shouting, because capitals are slower to read and take more width; an exclamation mark usually carries the volume. Some guides use capitals for translating on-screen text such as signs, to distinguish it from dialogue.

Punctuation follows the target language. Spanish opens questions and exclamations with inverted marks, French typically uses guillemets for quotations, and German uses its own quotation marks. Many styles leave out the full stop at the end of a subtitle in some languages, while English guides usually keep standard punctuation. Quotation marks are needed for quoted speech and for a character reading something aloud.

Song markers

Lyrics that are subtitled are usually marked with a music note symbol at the start and end of each lyric line or subtitle. Some styles also italicize the lyrics; others keep them upright and let the note do the work. In environments that cannot display the note character, a hash sign has been used as a substitute.

Whether to subtitle a song at all is a separate question that depends on whether its words matter to the story, and lyrics can raise rights questions. Those decisions, and singable versus literal translation, are covered in translating song lyrics in videos. Descriptions of instrumental music and other sounds, such as a bracketed note that music is playing, belong to captioning rather than typography and are covered in captioning sound effects and music.

How italics are written in SRT and VTT

SRT has no formal specification, but most players understand a small set of HTML-style tags. Italic text is wrapped in an opening and a closing i tag, the same tag HTML uses; bold and underline use b and u. Some players also understand a font tag for color. Support is uneven: desktop players such as VLC generally render these tags, while some platforms and devices ignore them or, occasionally, display them as literal characters.

WebVTT defines formatting as part of its specification. It supports the same i, b and u tags, class spans that a web page can style with CSS, voice spans that name a speaker, and language spans. Browsers render VTT through their built-in caption support, and a site can style cues with CSS, but platforms that import VTT may keep only a subset.

A safe workflow is to keep tags minimal, close every tag in the same cue it opens, and test the file on the platform where it will be published. If the subtitles are burned into the picture, formatting is fixed at export and player support stops mattering.

Common formatting mistakes and platform limits

• Italicizing a character who is merely off camera in one scene and not in another.
• Unclosed italic tags, which in some players turn every following cue italic.
• Mixing dialogue dash styles in the same file, or using a dash for a single speaker.
• Copying source-language punctuation and number formats into a translation.
• Relying on bold, color or underline to carry meaning when many platforms strip them.
• Shouting in capitals, which slows reading and makes lines too wide.

The biggest limit is support. Formatting written correctly in the file can still disappear on a given player or platform, so test before relying on it. If a platform rejects a file after you add tags, check first for a tag left open or a stray angle bracket in the text.

Formatting mydubly subtitle files

mydubly produces SRT and VTT files with one single-line cue per recognized speech segment, numbered in order and saved as UTF-8. The files carry no italics, styling, positioning, song markers, sound tags or speaker labels, and they contain recognized speech only. Punctuation, capitalization and number formats appear as the speech recognizer, and the translation where you chose a target language, produced them, which may not match your style guide.

That makes mydubly's output a timed draft to which you apply your conventions: add italics for off-screen voices, join two speakers into a dashed cue where it helps, mark lyrics, and normalize numbers in a subtitle editor. Editing mechanics are covered in how to edit an SRT file. The subtitle generator costs 1 credit per minute with a 5-credit minimum, so a 20-minute video costs 20 credits (2¢).

Next step: write down five formatting decisions

Before your next subtitle job, note your choices for off-screen italics, dialogue dashes, ellipses, numbers and song markers, then apply them to a short file and check it in the player your audience will use. If you need a starting file, generate an SRT file from your video and format it to those five decisions.