Updated · Clover Blog
Quick answer
To create offline subtitles in Clover, open a recording’s subtitle panel or import an audio file, choose the spoken language and a compatible bundled model, and start transcription. Review and save the text and timing, then export TXT, SRT, or a subtitled video.
Choose between a transcript, subtitle file, and finished video
Use a transcript when you need notes or searchable text. Choose timed subtitles when the destination player or editor accepts a separate subtitle track. A video with visible subtitles is useful when you want the wording to remain in the picture.
Review the final video edit before generating subtitles. Removing an interval later changes the timing; subtitles made for an earlier cut may no longer line up with speech.
| Output | Contains | Useful for |
|---|---|---|
| TXT | Plain text | Notes, quotations, editing text |
| SRT | Text plus start and end times | An editor or player with subtitle-file support |
| Subtitled video | Video with visible subtitles | Sharing a finished recording |
Check words and timing against the audio
Listen to names, numbers, acronyms, and technical terms instead of accepting the first recognition result. Check the first and last words of each subtitle segment as well as the text itself. A correct sentence at the wrong time is still difficult to follow.
The sample below is hand-written instructional content, not an accuracy demonstration or the output of a recorded transcription. It illustrates two numbered subtitle cues with start and end times.
Illustrative SRT file
1
00:00:01,000 --> 00:00:03,500
Choose the window you want to record.
2
00:00:04,000 --> 00:00:06,500
Check your microphone before you begin.
Download the illustrative SRTBefore you start
- Use a published release that includes the local speech models for your platform.
- For a recording, choose the microphone or system-audio track containing the speech. For an existing file, use MP3, WAV, or M4A.
- Finish any active recording first: transcription tasks wait while recording is active.
Open the recording or import audio
In Clover Studio, open the subtitle panel for the recording and choose the audio source containing speech. Alternatively, import one MP3, WAV, or M4A file in the transcription workspace.
Select the language and compatible model
Choose the language spoken in the audio. Automatic recommendation uses SenseVoice-Small for simplified or traditional Chinese tasks and Whisper Small for other supported language tasks. Chinese tasks can also explicitly use Whisper Small. Incompatible choices are unavailable.
Start local transcription
Start the task and follow preparation, queue, and transcription progress. Keep the source media available. Processing time depends on audio length and device performance; a queued task may be waiting for active recording to finish.
Review text and subtitle timing
Check names, technical terms, punctuation, and segment boundaries. Edit incorrect text and adjust subtitle timing where needed. Save your changes before exporting or switching to another file.
Export the result
Choose TXT for a transcript without timing, SRT for a separate timed subtitle file, or a subtitled video for a finished recording. Open the exported text or preview the video to confirm the result before sharing.
Troubleshooting
- The task stays queued
- Check whether a recording is still active. Transcription waits during recording. Finish the recording, then check the task state again.
- A language or model is unavailable
- Choose a compatible bundled model. The current SenseVoice workflow is limited to Chinese; other supported languages use Whisper Small. The model selector does not download arbitrary custom models.
- The words or timing are wrong
- Confirm the spoken language and that you selected the track containing speech. Review the source audio, edit the result, and check subtitle timing. Supported language selection does not guarantee recognition accuracy.
Check the result
You should have a saved, reviewed transcript or subtitle document and an exported file in your chosen format. Local speech processing does not itself publish the audio or create a shareable viewing link.
Continue reading
How to cut out the middle of a screen recordingHow to record your screen with microphone and system audioQuestions about Clover?
For product questions or help with a recording issue, contact us.