Skip to content

Editing captions

The editor has the preview (the video with the captions drawn over it), the style panel beside it, and the timeline underneath. Every edit is saved a moment after the last one, and every edit can be undone.

ActionHow
Play or pauseSpace, the play button, or a tap on the picture
Step one frame← and → (not while typing in a field)
MuteThe speaker button
SeekClick or drag on the ruler or either track of the timeline

Pause the video and double-click a word in the preview. Type the correction and press Enter. Esc cancels. A few rules:

  • A word is edited one at a time and keeps its timing.
  • Spaces inside it stay, and it is still one word.
  • Clearing the word and pressing Enter deletes it.

While the video is paused, drag the caption block in the preview. It is pulled towards the centre lines of the picture, and a thin guide appears when it is centred on one. Where you drop it is the position for every caption.

The timeline has a ruler, a video track and a caption track, with the playhead across them. Each block of the caption track is one caption line.

  • Select a block, then drag one of its edges to change when it starts or ends. If it touches the next caption, that caption’s edge is pushed along, down to a minimum length.
  • Zoom with Ctrl or ⌘ and the mouse wheel, or pinch on a trackpad. The zoom buttons centre on the playhead, and Fit shows the whole video.

How many words share a line is the Words per line setting in the Text tab. A line also breaks where the speaker paused, so words said minutes apart never share a caption.

If the captions are all a little early or late against the audio, use Timing in the style panel. One setting moves every caption together, in milliseconds (positive is later), and both the preview and the export use it.

Ctrl or ⌘ + Z undoes, Shift + Ctrl or ⌘ + Z redoes. The buttons beside the zoom do the same. Each step puts back the transcript, the style and the timing offset together, and a run of the same change made close together (a slider, a drag of an edge) counts as one step.

For what the controls do not do yet, such as splitting or merging captions or retiming many words at once, Edit transcript as text opens the transcript as JSON, one word to a line. Applying it checks the text first and names the segment and word that are wrong.

The same dialog imports subtitles: pick an SRT or WebVTT file (with or without per-word times) or an OpenCaptions JSON export. It is turned into the transcript text for you to look over before you apply it, as one undoable step. This is how to caption a video whose subtitles you already have. The format is described in Transcript format.

There is no Save button. A change is saved 800 ms after the last one, and the indicator in the editor shows when it has landed.