Appearance
Alignment
Alignment syncs a script you already have to its recording. You get the start and end time of every word, and can download subtitles or karaoke lyrics.
The page
The script fills the screen on the left, with the recording attached underneath. On the right, one panel has two tabs: Settings (language) and History. On a laptop the page itself doesn't scroll.
Align a script
- Open Alignment in the sidebar.
- Paste the exact words into the script box. Put each caption or lyric line on its own line.
- Add the recording: drop it anywhere on the script, or choose Choose file under it. An audio file plays next to its name; choose Change to pick another.
- In Settings, set Language if you know it.
- Choose Align. The price shows next to it.
The file uploads, then the alignment opens and shows Listening and aligning…. You can leave the page. It's listed in History when it's done.
Read the result
The finished alignment fits the window: the recording at the top, the lines below (they scroll on their own), and a panel on the right with Export and View tabs. Choose New alignment (top right) to start another.
- Each line shows its start time. Click the time or any word to play from there.
- Words the recording didn't say clearly are underlined with dots. Their timing is estimated.
- Match (at the top of Export) shows the share of script words heard, for example "92% of script words heard". The label says how many words are estimated, or "all heard".
- If much of the script wasn't heard, you see: "Much of the script wasn't heard in this recording. Check that the script matches the audio; estimated timings are underlined."
- Follow playback (in View) scrolls to the line being spoken.
Export
- In the Export tab, choose a format:
- Subtitles (.srt)
- Subtitles (.vtt)
- Lyrics (.lrc)
- Karaoke lyrics, word timings (.lrc)
- JSON with word timings
- Text with line times (.txt)
- For SRT and VTT, set Characters per line (32, 37 or 42) and Lines per caption (1 or 2).
- Set Timing offset (ms) if the subtitles run early or late. Positive shows words later; negative, earlier.
- Choose Download.
Options
- Language: Detect automatically or one of 31 languages (the same list as Transcribe).
- Names and brands in your script help the recording be heard as you wrote them.
Limits
- Script up to 100,000 characters.
- Files up to 2 GB.
- Recordings up to 10 hours.
- Audio or video files.
- Timing offset from −10,000 to 10,000 ms.
Credits
- 100 credits per minute of audio, rounded up. The minimum is 1 credit.
- If alignment fails, your credits are refunded.
Tips
- Paste the words exactly as spoken or sung. Extra or missing lines lower the match.
- One line per caption gives cleaner subtitles.
- Use Karaoke lyrics, word timings for players that highlight each word.
- Check the Match figure before you export. A low figure usually means the script and audio don't match.
Related
- Transcribe: you don't have the script yet.
- Music: make a song, then align its lyrics.
- API Keys: alignment is also available from the API.