Skip to main content
Participant
July 1, 2026
Open for Voting

Caption Translation matching the original Caption's timing

  • July 1, 2026
  • 1 reply
  • 16 views

Currently, when using the Caption translate option the text is translated and redistributed over the video but it creates mismatched and oddly timed captions. Sometimes the translated version merges lines from different speakers which is useless. There should be a checkbox like “keep original timing” so the text is translated line by line, maintaining the original captions’ TC.

In the screenshot below you can see the original caption in the bottom track and the translated version above it. As you can see the cuts are totally different. Above in the monitor you can see the original text (red) and the translated version. They do not match.

 

    1 reply

    Stan Jones
    Community Expert
    Community Expert
    July 4, 2026

    @Victoria22752362hoem,

     

    Upvoted. I see the problem, but the solution may be difficult.

     

    In the untranslated captions, the words fit the speaker’s rate of speech. And even that can be a challenge when they speak rapidly. But a translation often involves more characters and words to “say” the same thing. When there are fewer words, the translated captions are more likely to keep the same start and end timecode. And when there are more words in the translation, there can easily be too much for one caption.

     

    The settings for the original captions will affect this. If the captions were set to be short, single line, etc, then we’ll have more problems than if they were longer, double line. 

     

    Multiple speakers are a special problem, and it is frustrating that the speaker information from transcripts is still not used to facilitate styling multi-speaker captions.

     

    Stan