Voiceover and subtitles

A demo with a voice explains itself. DemoTake writes one short spoken line per scene, turns it into speech, and puts it on the take like any other edit. Nothing you recorded changes: edit a line, change the voice or remove the voiceover at any time, and Undo takes it back.

#Write the script

Open the editor's Captions tab. The Voiceover group is at the top. Click Write Voiceover Script (you need at least one recorded scene).

DemoTake writes a line for each scene from its description (the intent you gave it when you planned the take). With DemoTake AI set up, DemoTake AI writes the lines, each short enough to be said in its scene at a natural pace. Without it, the lines are made from the descriptions as written. Nothing is spoken yet.

#Edit the lines

Each line has its own text field, under its scene's number and title. Type over it like any text. The label on the right tells you where the line stands:

Use the play button to hear a line, and Speak Line Again (the arrow) to make its audio again. The preview plays the voiceover along with the video at normal speed, and the Audio track in the timeline shows each line; click one to go to it.

Lines start at the beginning of their scene, so they move with the scene when you trim, cut or speed up footage before them.

The ⋯ menu has the rest:

• Rewrite Script from Scene Descriptions… writes every line again. Your edits to the lines are lost, so DemoTake asks first.
• Speak Every Line Again… makes new audio for every line. It asks first and says how much voice time that is.
• Remove Voiceover removes the lines and their audio from the take.

All three can be undone with ⌘Z.

#Speak it

Pick a voice, then click Generate Voice. When only some lines changed, the button reads Generate 2 Changed and speaks just those. When everything is spoken, it reads Voice Up to Date. You can keep editing while it works, and cancel it; lines already spoken are kept.

Voices switches between two kinds of voice:

The Voice menu lists the voices for the voiceover's Language. The Mac's best voices come first, marked Enhanced or Premium. For more natural Mac voices, add an Enhanced or Premium voice in System Settings › Accessibility › Spoken Content › System Voice › Manage Voices. If the Mac has no voice for the language, the group says so.

With DemoTake AI voices, the group shows how much AI voice time is left this month. Included minutes depend on your plan; see Pricing. When there is none left, the Mac's voices still work.

The voiceover's language starts as the take's own language, never your Mac's. To make the demo in other languages, see Languages.

#Lines that run long

DemoTake never speeds a voice up. If a line is longer than its scene, the line says how far it runs over: "Runs 1.4 s past the end of the scene. Shorten it, or hold the scene."

You can:

• Shorten the line, then speak it again.
• Click Hold Scenes to Fit. Each scene whose line runs long holds its last frame until the line ends. The holds are edits like any other, so you can undo them.

#Mix it with music and your microphone

The voiceover has its own lane in the mix. If the take has music, the music dips under the voice and comes back up after it.

Mute microphone under voiceover silences what the microphone recorded in scenes that have a voiceover, so you don't hear yourself under the voice. The app's own sound keeps playing.

#Captions that follow the voice

Turn on Captions follow voiceover to replace the captions made from scene descriptions with the voiceover itself: one caption per spoken sentence, timed with the voice. When you speak lines again, the captions are written again to match. Your other captions stay. See Captions.

#Subtitles and transcripts

The Export sheet's Subtitles section adds files and tracks to every video you export:

• Files: an SRT file, a WebVTT file and a transcript in Markdown, with a heading for each scene.
• Subtitle track: a subtitle track inside the MP4. Players that show subtitle tracks, such as QuickTime Player, let viewers turn it on.
• Words from: when a take has both a voiceover and captions, choose which one the subtitles say. Without a voiceover, they say what the captions say.

Everything is timed to the edited video: after trims, cuts, speed changes and holds. A take in several languages exports subtitles per language, and can carry every language as a track; see Languages.

When you share a video with WebVTT subtitles, the share page and embeds show them. See Embeds, calls to action and analytics.

#In DemoTake Web

The web editor has the same Voiceover group in its Captions tab, with DemoTake AI voices. A voiceover made with a Mac voice plays there too; speak changed lines again on the Mac to keep the same voice, or switch to an AI voice. Holding scenes for a long line is done on the Mac. The web Export sheet downloads SRT, WebVTT and Transcript files.

#Privacy

• Mac voices send nothing. Speech is made on your Mac, offline.
• DemoTake AI voices send the text of each line, its language, the voice you picked and a short label with the take and scene names. No pictures and no recording. The audio comes back to your Mac and is stored in the take.
• Writing the script with DemoTake AI sends the scene descriptions, captions and steps as text, after Secret Guard has replaced likely secrets. See Privacy and data.

The voiceover is part of the take, on your Mac, until you sync or share it.

#For agents

Agents write, speak and edit the voiceover with these tools:

• voiceover_script writes the lines (style: concise, tutorial or friendly; instructions for DemoTake AI).
• voiceover_update changes a line, where it starts, and the options mute_microphone, sync_captions and volume. fit: true holds scenes whose line runs long.
• voices_list lists the Mac's voices and the AI voices for a language, with the time left.
• voiceover_generate speaks the lines. engine is device (the Mac's voices, the default) or hosted (DemoTake AI voices). dry_run: true reports what would be spoken and for how long.
• voiceover_remove removes lines or the whole voiceover.
• export_subtitles writes SRT, WebVTT and transcript files on their own.

{ "tool": "voiceover_generate", "arguments": { "take_id": "take_…", "engine": "device", "fit": true } }

From the terminal, demotake export adds subtitles to an export:

demotake export take_01j… --format 16:9 --subtitles srt,vtt,md --embed-subtitles
demotake export take_01j… --format srt,vtt --subtitle-source captions

#Next

• Languages to make the demo in other languages
• Interactive demos, where each step plays its part of the voiceover
• Music
• Captions