How to transcribe only the dialogue and skip the singing
Tell Vizard Agent that the sung passages are not dialogue. Speech recognition transcribes singing as readily as speech, so without that instruction a theme song and a lyric over a montage end up in the transcript looking exactly like lines of script.
What do you type into Vizard Agent?
State the exclusion as part of the job, not as a correction afterwards. The default behaviour is reasonable and simply not what you want. Say what Vizard Agent should do with the songs too: dropping them silently and marking them as "[song]" produce very different documents, and the marker is usually what you want for translation or subtitles.
Transcribe the spoken dialogue only. Leave out the singing, but mark where the songs are.
What does Vizard Agent actually do?
It classifies before it transcribes. The audio is examined to work out which stretches are spoken and which are sung, and only then does transcription run on the spoken stretches. Deleting lyrics from a finished transcript means recognising them from the text alone, which is unreliable. In the session behind this article it:
- Ran a second pass on a drama episode after a first transcript had included everything, because the request to skip the singing came later.
- Extracted the complete audio track and checked how spoken passages differ from sung ones in that particular episode.
- Identified the speaking and singing sections across the whole runtime.
- Cross-checked the on-screen Chinese subtitles at the dialogue stretches only, zooming into frames to read small lines where the audio was hard to make out.
- Checked the closing section separately so the sign-off lines were not dropped with the end theme, since endings are where speech and song overlap most.
When does this not work well?
When someone speaks over the song. Dialogue under a music bed has to be classed as speech-with-music or as song, and the call will sometimes go wrong both ways, so Vizard Agent flags the ambiguous stretches. Chanting, recitative and rap sit between the categories, and a flattened mix with the song at speech level leaves nothing to separate.
Common questions
Will the timestamps still be right? Yes. Vizard Agent keeps the original timings, so removed songs leave gaps rather than shifting everything.
What about dialogue over background music? That counts as dialogue by default. Tell Vizard Agent if you want it excluded.
How accurate is the split? Good in the middles, less certain at boundaries, which is why Vizard Agent checks those specifically.