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

# Voiceover & lip-sync

> Generate natural voiceover, localize ads, and control pronunciation.

Voice is half of a video ad. Superscale generates natural-sounding voiceover and, for talk-to-camera formats, syncs it to the creator's mouth. It can also help you move an ad into new markets, but localization is more than translation.

## What voice controls

<CardGroup cols={2}>
  <Card title="Native & studio voice">
    English Seedance ads can use native generated audio. Non-English voiced ads use post-mixed studio voiceover.
  </Card>

  <Card title="20+ languages">
    Localize voiceover-driven formats for the markets you advertise in.
  </Card>

  <Card title="Lip-sync">
    Talk-to-camera lip-sync is supported for English UGC Testimonial-style videos and template workflows that expose lip-sync controls.
  </Card>

  <Card title="Correct pronunciation">
    Add notes or upload pronunciation audio for brand names, acronyms, technical terms, and product names.
  </Card>
</CardGroup>

## Native audio versus voiceover mix

| Video case                   | What happens                                                                                           |
| ---------------------------- | ------------------------------------------------------------------------------------------------------ |
| **English UGC Testimonial**  | The spoken lines are part of the Seedance scene plan and are voiced natively.                          |
| **English narrator formats** | The narrator can be voiced natively when the format supports it.                                       |
| **Non-English voiced ads**   | The spoken script is generated as external voiceover and mixed into the video after render.            |
| **Faceless B-roll**          | Voiceover is mixed over visual scenes; the scene plan should stay visual rather than contain dialogue. |

<Note>
  Non-English talking-to-camera Seedance testimonials are not the right workflow. Use non-English UGC Lifestyle, B-Roll, or another voiceover-mix format instead.
</Note>

## Localization has four layers

| Layer   | What to check                                                       |
| ------- | ------------------------------------------------------------------- |
| Copy    | Is the message natural in the target language?                      |
| Visuals | Do the scenes, people, currency, and examples fit the market?       |
| Audio   | Does the accent, age, gender, pace, and pronunciation match the ad? |
| Context | Are local objections, competitors, and claims different?            |

<Note>
  Translating the script is not the same as localizing the ad. A native-sounding voice still needs market-appropriate examples and visuals.
</Note>

## How to get better pronunciation

<Steps>
  <Step title="Write the script" icon="file-text">
    Keep brand names, numbers, and technical terms clear.
  </Step>

  <Step title="Add pronunciation hints" icon="message-square">
    Use simple phonetics: "Amaze, rhymes with days" or "HVAC, say each letter."
  </Step>

  <Step title="Upload pronunciation audio" icon="upload">
    In Video Context, upload a short `.mp3`, `.wav`, or `.m4a` pronunciation file when text hints are not enough.
  </Step>

  <Step title="Preview the voice" icon="headphones">
    Listen before spending heavily on a full batch.
  </Step>

  <Step title="Adjust the market context" icon="globe">
    Tell Superscale the country, language variant, and cultural context, not just "translate to Spanish".
  </Step>
</Steps>

## Common voice questions

<AccordionGroup>
  <Accordion title="Can I upload my own voice?" icon="upload">
    Upload support depends on the workflow and account. If you need a specific voice, ask support or use the available voice/reference controls in the current video workflow.
  </Accordion>

  <Accordion title="Why is the accent wrong?" icon="languages">
    Specify the market, not only the language. "Mexican Spanish" and "Spanish" are different creative instructions.
  </Accordion>

  <Accordion title="Should I localize visuals too?" icon="image">
    Yes. Currency, screenshots, people, examples, and competitor references can all change how native an ad feels.
  </Accordion>
</AccordionGroup>

<Tip>
  Save pronunciation rules in [Video Context](/features/video-context) when they should apply to future video generations.
</Tip>