Why captions do not always match my lyrics
Captions in AITuber are timed to the audio, word by word. In narrated videos the captions match your script exactly. In music videos the AI singer can stretch, bend, or blur words, so the captions show what we hear in the song, which is not always what you wrote.
How do captions work in normal videos?
When we generate the narration, we also get the exact time each word is spoken. The captions use those times. That is why each word pops on screen at the moment the voice says it, in every language.
Why are my music video captions different from my lyrics?
Music videos work in reverse. The AI singer performs your lyrics first. Then we listen to the finished song and write down the words we hear, with their timing. This keeps the captions in sync with the vocals. But singing is not speech:
- The singer can stretch one word over several beats.
- Some words blend together or change shape in the melody.
- Loud instruments can hide parts of a word.
So a caption can differ from your written lyric. The caption is faithful to the sound; the lyric sheet is faithful to your text. We chose the sound, because captions that drift out of sync look worse than a slightly different word.
Why does my music video have no captions at all?
If the song is instrumental, there are no lyrics to show. If the vocals are too unclear for us to hear real words, we skip captions instead of showing wrong ones.
What can I do about it?
- Turn captions off for the music video: on the music video setup page, set the Captions switch to Off before you generate.
- Fix single captions after generation: open the video, click Edit, select the captions on the timeline, and adjust or delete them, then export again.
- Prefer clear vocals: simple, short lyric lines are easier to sing clearly and caption correctly.
Related
Updated on: 09/08/2026
Thank you!