The Polyglot journal
Bilingual subtitles or separate language tracks?
Two languages on screen and two languages in a menu solve different problems. Choose the right delivery, preserve the reviewed originals, and check the result where people will watch.

Choose what the viewer needs to see
Use bilingual subtitles when viewers deliberately need to compare the original and a translation at the same time. Use separate language tracks when viewers should choose one language and watch without the second taking up space. In Polyglot, a bilingual SRT or VTT combines two texts inside each timed cue. It does not create two independently selectable tracks.
This is a delivery decision for a recording and subtitles you own or have permission to process and share. Start with the recipient’s viewing situation: a comparison copy for a reviewer may need both languages, while the public version of the same video may need separate files. Decide this before naming, uploading or approving the exports.
Distinguish three different outputs
A combined bilingual sidecar is one subtitle file. Each cue contains the original plus one translation, so selecting that file displays both together. Switching that track off hides both. A line break between languages is a text boundary, not a second language selector or a guarantee of two visible lines.
Separate tracks are individual files, such as English SRT and Dutch SRT, that a compatible destination can expose as language choices. YouTube’s subtitle instructions, for example, organize additions by language and allow files with existing timing. That supports a one-file-per-language handoff. Polyglot exports the files; uploading them, labeling the languages and checking the published player are separate tasks.
Burned-in subtitles are part of the video picture. Neither a bilingual SRT nor two separate SRT files burns words into the video. If the recipient needs permanently visible text, agree on a separate video-rendering workflow. Do not promise that sending a sidecar will make it appear in every player.
Make the delivery agreement explicit
For a reviewer comparing a translation against its source, offer a clearly labeled bilingual review copy. For a public audience that needs language choice, prefer separate reviewed tracks. If a lesson genuinely requires simultaneous comparison, test the combined version on the actual teaching screen and on the smallest screen learners will use. The audience determines the choice; more text is not automatically a better experience.
Ask the recipient to confirm the media revision, accepted file format, requested languages, whether the languages should appear together, and who approves the final display. “English and Dutch subtitles” is ambiguous until those choices are settled. Keep a short written delivery note so a combined review file does not accidentally become the only public track.
A translation of dialogue is also not a complete accessibility review. W3C explains that captions include meaningful non-speech audio as well as speech. Check speaker identification and relevant sounds where needed in each intended caption track. Combining languages does not supply missing sound descriptions or establish accessibility compliance.
Prepare the pair without changing its timing
In Polyglot, begin with a reviewed original and a completed translation from the same project and media revision. Compare their meaning before combining them: correct structure does not prove that a negative, quantity or name was translated correctly. Keep single-language exports available so a later layout decision does not require extracting text from the combined file.
Open the project’s Export menu and choose Export subtitles or transcript…. Select SRT subtitles or WebVTT subtitles, choose the completed translation, and enable Include original and translation together. Choose the order consistently. For separate tracks, leave the combined option off and export the original and translation individually.
Bilingual output retains the source cue timing and pairs the corresponding original and translated text. Within each language it joins internal whitespace, then places a line break between the two texts. It is plain timed text: it does not preserve a carefully styled two-column design, set independent language positions, or lengthen a cue to fit the extra words. The player controls the final presentation.
Example: one pair, three clearly named files
The original English and Dutch sample below represents a fictional paper-folding clip. It is an editorial example, not a transcription or translation benchmark. The combined version keeps each sentence pair inside one cue. The order is English first throughout; changing that order halfway through would make comparison unnecessarily difficult.
The file set separates viewing choices from a review aid. paper-v3.en.srt contains only English, paper-v3.nl.srt contains only Dutch, and paper-v3.en-nl.bilingual.srt contains the pairs. These are suggested handoff names, not a claim about the app’s default filenames. The recipient should select the correct destination language for each single-language upload rather than expecting the filename to configure the player.
Combined review copy
1
00:00:01,000 --> 00:00:04,000
Fold the paper once.
Vouw het papier één keer.
2
00:00:05,000 --> 00:00:08,000
Leave the blue edge visible.
Laat de blauwe rand zichtbaar.
Delivery note
Video: paper-v3.mp4
English: paper-v3.en.srt
Dutch: paper-v3.nl.srt
Review pair:
paper-v3.en-nl.bilingual.srt
Order: English, then Dutch
Public tracks: English, Dutch
Pair: reviewer copy onlyJudge the rendered result, not the newline count
Load the exact exported file beside the final video in the destination player. Check the opening, closing and densest passages, plus overlaps and quick changes. Then review the full track. Confirm both languages appear in the agreed order and disappear at the right time, without an old automatic or embedded track also displaying duplicate text.
Test a narrow viewing window and the caption size the audience is likely to need. The current W3C WebVTT draft explains that players can insert additional line breaks to fit the cue width. Consequently, two text lines in a file can occupy three or four lines on screen. Check whether either language gets clipped, covers important action, or leaves too little time to read.
Use a concrete pass condition: a reviewer can read the intended language or comparison pair at ordinary playback speed, the needed image remains visible, and the language menu matches the delivery agreement. For separate tracks, select each language in turn and verify that only the intended text appears. For a combined track, confirm that the two languages remain together. Record the player and screen used; a successful desktop check does not establish every mobile player’s behavior.
If the pair is too crowded, change the delivery
First check whether both languages really need to be simultaneous. If not, deliver the separate files. If comparison is essential, have a competent reviewer make any shortening preserve the meaning, then export and inspect again. Do not delete a negation, condition, speaker label or meaningful sound just to make the block smaller. Smaller text may hide a layout problem while making reading harder.
A whole-track timing shift will not create more reading time: it moves the intervals without extending them. When cue splitting, individual retiming or a designed top-and-bottom layout is necessary, use an editor and destination that support that work and review the resulting timing. Keep those changes in a separate version. Polyglot’s plain bilingual export does not perform that layout or retiming automatically.
Hand over the approved version, then check it once more
Keep the original track, reviewed translation and combined export as distinct files. Include the media revision, language order, intended use and any unresolved issues in the delivery note. A combined review copy should be labeled as such; if an issue is still unresolved, hold that deliverable rather than calling the set approved.
After the recipient imports or uploads the files, check the final viewing page or player again. Verify language labels, selection, wrapping, opening and closing cues, and absence of duplicate tracks. If the video edit or subtitle wording changes, repeat the affected checks and regenerate the combined copy from the reviewed tracks. A useful handoff leaves the next person knowing which file to publish and which file was only there to help review it.
Sources & further reading
W3C WAI: captions, translated subtitles and meaningful audio
YouTube Help: add language-specific subtitles and upload timed files
W3C WebVTT Candidate Recommendation Draft: multiline cues and wrapping
Published by Polyglot using an AI-assisted editorial workflow. How we prepare and update our guides.
Keep exploring.
Try it on your Mac.
Find the right tools for your words and video, with processing on your Mac.
Download nowApple Silicon · macOS 14.4 or later