> ## Documentation Index
> Fetch the complete documentation index at: https://docs.yarn.so/llms.txt
> Use this file to discover all available pages before exploring further.

# Synthetic Voice

> Generate narration from script text and refine it for natural pacing and clarity.

## Overview

Use Synthetic Voice when you need fast narration updates without re-recording audio takes.

The key is writing script text for speech, not for reading.

\[insert screenshot of Synthetic Voice panel with script input, voice selection, and generate action]

## Before you start

* Split your script into short scene-aligned chunks.
* Decide the target tone (neutral, energetic, instructional).
* Keep project timeline open so you can validate sync immediately.

## Steps

1. Write one concise script block per scene.
2. Choose a voice profile that matches your audience and tone.
3. Generate narration for a small section first.
4. Review pronunciation, pacing, and emphasis.
5. Regenerate only weak sections instead of replacing everything.
6. Align generated audio timing with scene visuals.

\[insert screenshot of generated narration clip aligned to a scene in timeline]

## Check your result

* Narration sounds natural enough for the intended audience.
* Key terms are pronounced correctly.
* Scene transitions stay synchronized with spoken content.

## Troubleshooting

### Voice sounds robotic

Shorten sentence length and add punctuation for natural pauses.

### Pronunciation is wrong

Adjust script spelling/phonetics for the specific word and regenerate only that segment.

### Narration timing does not match visuals

Trim the audio segment or split the script block so pacing matches the scene.

## Next

Draft and organize script structure in [Camera Scripts](/talk-track/camera-scripts).