Translate Video Subtitles and Export SRT

Video localization involves two separate decisions: what text to translate, and what to do with text already visible in the picture. Keeping them separate helps avoid translating the wrong source or putting a new caption on top of an old one.

Subtitle Remover can recognize visible subtitles with OCR or spoken dialogue with speech recognition, translate the result, and produce a subtitled video and SRT files. Translation changes the captions, not the spoken audio: this is not a dubbing workflow.

Choose the recognition source

OCR reads text from the picture. ASR listens to speech. They can produce different sentences if the existing captions summarize the speaker or are already translated.

For a tutorial whose on-screen instructions differ from the narration, decide which audience experience you want to reproduce. For a bilingual caption block, check which lines recognition returns before accepting the translation.

Decide whether to remove existing subtitles

If the video has no visible captions, choose no removal and enable translation. There is no reason to reconstruct a clean picture simply to add subtitles.

If captions are burned in, choose Auto for the lower-half subtitle area or Precise for defined boxes. Recognition uses the original source while removal prepares the clean video for the new text. The hardcoded subtitle guide covers region selection and quality checks.

If captions are a selectable track, see the MP4 track-removal guide. Deleting a track and erasing pixels are different operations. Keep any original subtitle file as a reference for proofreading.

Set up the translation job

1. Open the studio and upload your clip.
2. Choose the removal mode, or keep the picture unchanged.
3. Enable translated subtitles and select on-screen text or speech as the source.
4. Select the target language. Choose the original-language option if you want to recognize and restyle captions without translating them.
5. Choose the offered font, color, size and position, then review the credit estimate.

The studio currently accepts videos up to 10 minutes, 2 GB and 2K input. Output is up to 1080p. Translation adds 1 credit per second; removing subtitles has its own mode-based charge. A 30-second Auto-plus-translation job is currently 60 credits, while translation without removal is 30. Check pricing for current rates.

Proofread before delivering

After processing, review the translated lines in the editor. Correct names, numbers, technical terms and negation before polishing style. A fluent sentence can still mean the opposite of what was said.

Watch the video while checking the start and end of each cue. Captions should not disappear before the reader can finish or linger over an unrelated scene. Adjust long lines so the text fits the chosen display size, and inspect vertical videos at phone width.

You can edit wording and timing, change the available style settings, and burn the edited subtitles into the video again. Re-burning uses additional credits; review that estimate before starting. Save an SRT of the approved text as well as the rendered video.

Know which download you need

The job view provides the outputs available for that job: a clean or base video, the subtitled MP4, and recognized and translated SRT files. Use the clean version for another editing pass and the subtitled MP4 when you need captions permanently visible.

An SRT contains cue text and timestamps. It does not preserve the studio's font, colors or precise placement, and it does not translate spoken audio. If a platform accepts a separate subtitle file, upload the SRT with the clean video and test playback there.

Keep the source video and final subtitle file together. If a translation needs a correction later, an editable text file is easier to update than a caption already burned into a delivered copy.

Common questions

Can I translate a video with no subtitles?

Yes, select speech recognition as the source and enable translation. Review the transcript when sound is noisy or speakers overlap.

Can I change the language while keeping the original voice?

Yes. This workflow adds translated text while retaining the original dialogue audio; it does not synthesize a new voice.

Why did recognition miss a caption?

Small or stylized text can be difficult for OCR. Check the source quality and whether speech recognition better matches your goal. Proofread the result rather than assuming the two sources are interchangeable.

Start with a short section in the translation studio, review the text and picture together, then use the same approach for the rest of the edit.