How to tune the narration

Use the Narration tab to set the project’s voice and speed, control how narration sits under the original audio, and adjust individual lines. Open a video project and select Narration in the panel beside the player.

The tab appears once the project has descriptions. It isn’t available for YouTube-sourced projects, because the preview needs an uploaded video file.

Your export uses whatever is set here — the voice, the speed, the mix, the placement, and any line you adjusted. See Export a video.

Work in this order

Each step below undoes work from the steps after it, so doing them out of order costs credits and loses adjustments:

1. Settle the descriptions themselves — rewrite, elaborate, delete. See Generate and edit audio descriptions.
2. Fix any wording you want the voice to read differently. Saving new wording resets that line’s manual placement, so doing this after step 5 throws away your timing work.
3. Choose the voice and project speed.
4. Set the placement and mix for automatically placed lines.
5. Prepare the narration preview once, when the line list is final.
6. Adjust individual lines last, now that you can hear them.

Choose the voice

Under Narration voice, pick from the female and male voices offered in your project’s language.

Select the sample button beside the picker to hear the voice. Samples are free and cost no credits.

If some lines are already rendered in another voice, you’ll be asked to confirm the switch and told what re-rendering costs. Nothing is deleted — your earlier recordings are kept, so switching back later is free.

Try voices freely; you only pay once per line per voice. Each line shows its own state: Ready, Ready, restore free, or Needs rendering. “Restore free” means you’ve already paid for that line in the current voice, so Restore (free) brings the recording back at no cost. That’s what makes auditioning a second voice on a finished project — and changing your mind — cost nothing beyond the one render.

Set the speed

Narration speed runs from 0.75× to 2× and applies to the whole project.

Changing it is free. The speed is applied during playback and again during export, so no line is re-rendered.

Decide how narration sits in the video

The How narration sits in the video section governs every line that is still placed automatically. Lines you’ve adjusted yourself keep their own placement.

When a line doesn’t fit the silent gap it was given:

• Choose Keep the video playing if the exported file has to stay exactly as long as the source. Long lines may overlap the original audio.
• Choose Pause the video when a line doesn’t fit if nothing may be talked over. The frame freezes until the line finishes, which makes the exported video longer. The panel tells you how many lines pause and how many seconds the export gains.

Speed lines up to fit the gaps lets our tool speed a line up to 2× before falling back to whichever of those two options you chose. Turn it off if faster delivery is harder to follow than an overlap or a pause.

Lower the source under narration sets how far the original audio drops while a line is read. Source volume sets the level it drops from. Both run from 0% to 100%.

For the reasoning behind these three options, see About narration placement.

Adjust a single line

Each line in the list shows its start time, whether it’s placed automatically or by you, and whether its audio has been rendered.

Select Adjust on a line to change:

• Start — type a time, or nudge with the arrow keys by 0.1 s (hold Shift for 1 s).
• Placement for this line — play over the video, or pause the video.
• Speed — 0.5× to 2×, for this line only.
• Volume — −30 dB to +10 dB, for this line only.

Changing any of these pins the line. Pinned lines are unaffected by the project-level placement settings above, and the remaining lines are re-planned around them.

To undo, choose Return to automatic placement. That also clears the line’s start time, speed, and volume.

Fix the wording of a line

Select Fix wording on a line to edit the text the voice reads.

Editing is always free. If the line has already been rendered, the recording no longer matches the text, so the line needs re-rendering — the dialog shows the cost. Putting the old wording back restores the original recording at no cost.

To change a description’s content rather than its wording, use rewrite or elaborate in the description editor instead — see Generate and edit audio descriptions.

Hear the export voice before you export

Until you render, previews use your own device’s voice, which is not the voice your export will use.

Select Prepare preview to render every line in the export voice. Playing the lines afterwards is free, and your export reuses the rendered audio instead of paying for it a second time.

To re-render one line — after a wording fix, for example — use the button on that line. A single line is priced by its length rather than at a flat rate. See Plans and credit costs.

Each line is limited to 800 characters. Over its lifetime a project can render up to three times its line count; if you reach that limit, export the project to render the remaining lines.

Preparing the preview and adjusting lines one by one require a paid plan. On the free plan you can still choose the voice and speed your export will use.

Read descriptions aloud while you edit

Editor read-aloud speaks each description as you move through the video, using your own device’s voice. It’s an editing aid: it costs nothing and is never part of a preview or an export.

Next steps

• Export a video — produce the final MP3, MP4, or SRT.
• Share a video — publish a link instead of a file, and let each viewer choose how descriptions reach them.