How to Make English and Chinese Bilingual SRT Subtitles
A bilingual SRT places two languages inside each timed cue. The timecodes remain unchanged; only the cue text gains a second line. This is useful for review copies, festival screeners, international client approvals and social videos aimed at mixed-language audiences.
FrameMath’s subtitle translator and bilingual SRT maker can combine an existing English SRT with translated Chinese text entirely in the browser. Automatic cloud translation is currently paused, but TXT/DOCX matching, cue editing, bilingual export and delivery checks are available without uploading the video or subtitle file.
The bilingual SRT structure
A normal English cue looks like this:
1
00:00:01,000 --> 00:00:03,400
The city wakes before sunrise.
A bilingual English-first cue adds the Chinese translation below it:
1
00:00:01,000 --> 00:00:03,400
The city wakes before sunrise.
城市在日出前苏醒。
You can download a bilingual English-Chinese SRT sample to test it in VLC, Premiere Pro, DaVinci Resolve or another subtitle-capable player.
Create bilingual subtitles step by step
- Load the source English
.srtor.vttin the bilingual subtitle maker. - Import the translator’s
.txtor.docx, or paste one translated line per cue. - Run Auto-match translation to map the translated lines to the original cue order.
- Review every cue, especially where the number of source and translated lines differs.
- Choose
Source + translationorTranslation + source. - Export SRT or VTT and test the result over the actual picture.
The source subtitle provides all timecodes. The translated document provides the second language. The tool does not need the video.
English first or Chinese first?
Use English first when the primary reviewer reads English and the Chinese line is a reference translation. Use Chinese first when the intended audience reads Chinese and the English line is retained for quality control or international approval.
Do not alternate the order between cues. A stable visual pattern reduces reading effort and makes missing translations easier to spot.
One translated paragraph per cue
Automatic matching works best when the translator keeps the same cue order and returns one non-empty paragraph per source cue. They may rewrite the sentence naturally, but they should not merge several cues into one paragraph or split one cue across multiple paragraphs unless the subtitle will be manually conformed afterward.
If the document has 120 translated paragraphs and the SRT has 124 cues, stop and inspect the mismatch. Common causes include:
- Music or sound-effect captions omitted by the translator
- Two short source cues merged into one sentence
- Blank paragraphs in DOCX
- Speaker labels removed or added
- A source cue duplicated during editing
FrameMath reports the line-count difference before export so you can correct it instead of silently shifting every later translation onto the wrong timecode.
Keep bilingual subtitles readable
Two languages need more screen space than one. After matching, check:
- Line length: avoid long English and Chinese lines competing across the full width.
- Reading speed: a short cue may not allow enough time for both languages.
- Line count: many players handle two lines well; four visual lines can cover faces or graphics.
- Punctuation: use the punctuation conventions of each language.
- Names and numbers: compare proper nouns, prices, dates and units manually.
The local quality check flags empty translations, overlapping cues, long lines and high characters-per-second values. These checks do not judge translation quality, but they catch many delivery failures before export.
SRT or VTT?
Choose SRT for editing software, desktop players and broad compatibility. Choose VTT for HTML5 players and web delivery. If you receive ASS, SSA, SBV or LRC instead, convert it with the subtitle format converter first. For a detailed comparison, read SRT vs VTT.
The safest workflow is simple: preserve the original timecodes, keep translation order aligned, review mismatches, export, and watch the full result once over picture.