What makes these scripts different
Several widely used scripts are written from right to left, including those used for Arabic, Hebrew, Persian and Urdu. Readers start at the right edge of a line and move leftward, and the first word of a sentence sits at the right. For a subtitle, this affects alignment, the order of mixed content and the position of punctuation. Right-to-left scripts are not exotic: they are used by a very large number of viewers. Even so, display faults are common because much subtitle software was built first for left-to-right text.
Bidirectional text
The central complication is that real subtitles are rarely purely right to left. A line in Arabic may contain a brand name written in Latin letters, a number, or an English word. Digits are normally displayed left to right even inside right-to-left text, and Latin words also run left to right. The result is a line containing both directions, called bidirectional or mixed-direction text.
Computers handle this by applying display rules that decide the visual order of each run of characters. The rules are well established, but they rely on the software knowing the base direction of the paragraph. If the base direction is wrong, the line can look scrambled even though the stored text is perfectly correct.
Typical symptoms include:
- A full stop or question mark that appears at the wrong end of the line, at the start of the text visually instead of the end.
- Brackets or quotation marks that seem reversed or on the wrong side of the quoted words.
- Numbers or Latin words that appear in the wrong position relative to the Arabic or Hebrew words around them.
- A line that is left-aligned when it should be right-aligned, or the opposite.
These problems are about display, not spelling. Editing the text to “fix” the visual result can make the underlying file wrong for other players, so it is better to identify the cause first.
Punctuation at the edges
Punctuation causes most of the visible trouble. A sentence-ending mark has no inherent direction of its own, so the display rules decide where it goes from the characters around it. If a line ends in a Latin word or number followed by a full stop, the mark may jump to the wrong side. Authors sometimes work around this with invisible direction-marker characters, which tell the software which direction a mark should follow. Such markers are part of the standard text encoding system, and they work reliably only if the player understands them. Some software shows them as odd boxes, so test before relying on them.
Style conventions also differ. Arabic uses its own question mark and comma shapes, and Hebrew and Persian have their own punctuation habits. Local style guides may prefer certain forms, so a reviewer from the target region should confirm them. General notes on subtitle punctuation apply, but the details change by language.
Script features that affect layout
Some right-to-left scripts have extra properties:
- Joined letters. In Arabic script, letters change shape depending on their position in a word and are joined together. Software must choose the right shape. A tool that cannot do so shows disconnected letters, which look wrong to readers.
- Marks above and below letters. Some text includes vowel marks or other signs. They need enough line height so that they do not collide with neighboring lines, and a thin or very small font can make them unreadable. Choosing a suitable font is part of subtitle legibility.
- Line breaking. Words in these scripts are separated by spaces, so breaking a line is similar to Latin script, but the reading flow matters. Breaking at natural phrase boundaries helps, as with any language.
Character encoding and fonts
If a file is saved with the wrong character encoding, right-to-left text can show as question marks or garbled symbols. A broadly supported universal encoding is the usual recommended choice, but players differ in what they accept. The problem is covered in character encoding.
Fonts matter too. A font that lacks the needed characters causes empty boxes or substitutes from another font. This is a particular risk with burned-in subtitles, where the text is fixed into the picture and cannot be corrected afterward. For separate files, the player or the viewer’s device chooses how to render the text, and the same file can look different between players. Details are in file formats and web players.
Positioning and layout on screen
Subtitles are usually centered at the bottom of the screen, and centered text is not affected by direction in the same way as aligned text. When the subtitles are aligned to one edge, as in some interface styles, right-to-left text should be aligned to the right. Dialogue with two speakers, shown on two lines with a dash for each, also needs attention. In left-to-right subtitles the dash comes at the start of the line on the left; in right-to-left subtitles the visual start is on the right, and the dash may appear on either side depending on the display. Check how speaker identification looks in the player.
Where a translated title or on-screen text appears in the picture, the subtitle should avoid covering it, regardless of direction.
Testing before delivery
Because faults often appear only in the final player, practical testing matters more here than for most languages.
- View the subtitles in each target player or device type, not only in the editing tool.
- Read every line that contains numbers, Latin words, brackets or end punctuation.
- Check long lines and two-line subtitles for correct alignment and order.
- Ask a fluent reader of the language to review a sample on screen.
- Keep a note of any markers or special characters used, so the next editor does not remove them by mistake.
Human reviewers are important here. Automated tools can produce text with correct words and wrong display order, and the problem may not appear until the text is on screen. Standards and practices vary by region and platform, so confirm requirements with the party that will broadcast or publish the subtitles.