CaptionHub Voiceover
Learn about our text-to-speech service that allows you to create synthetic voiceovers.
Requirements: Enterprise subscription
CaptionHub Voiceover is a text-to-speech service that lets you create synthetic voiceovers directly within CaptionHub. It reads aloud the text attached to a set of captions.
With the Voiceover Editor integrated into the Caption Editor, you can generate and refine synthetic voiceovers seamlessly. This guide walks you through the process of creating and editing voiceovers for optimal results.
Generating a voiceover
You can generate a synthetic voiceover from directly within the Caption Editor
- Navigate into a project
- Open the Caption Editor for original captions or the language version you wish to create synthetic voice for.
- If you have multiple speakers, make sure that they’re labelled correctly via caption metadata (within your original captions).
- Use paragraph markers to indicate the start of a new paragraph, as they help distinguish sentences and structure voiceovers.
- Ensure that original captions and translations are correct, and synchronised to the audio before proceeding to the next step.
- Go to Display Options.
- Under Editor type, select Voiceover.
- Click Generate Voiceover.
- This presents a modal where you will need to:
- Select a voice provider
- Select a voice for your default speaker (required)
- Map voices for all speakers
- Please note that speaker metadata in this view is automatically carried over from the speaker data assigned within the Caption Editor. To learn more about managing speakers, including how to edit or update speaker assignments, please refer to this guide.
- Application of a voice dictionary if applicable
- The option to enforce strict synchronisation with captions
- Check this if you'd like to keep everything in sync. If a sentence of voiceover output is longer than the caption allows, CaptionHub will adjust the speed of the voiceover to make it fit. However, if the voiceover is too short to match the captions, the system will also slow down the voice to ensure the duration matches as closely as possible. When this is enabled, we attempt to match the duration exactly.
If you prefer a more natural voiceover pace, disable this option. After the initial creation, you can adjust the tempo of the voice as needed.
- Once your selections have been made, choose ‘create audio’ to produce a translated synthetic voiceover.
- Learn how to edit your voiceover output here.
- Learn how to create voiceover deliverables here.
Note:
- In order to ensure a natural speech pattern, we may combine captions into a single voiceover sentence block.
- By default, the Voiceover Editor is visible only to superusers, producers assigned to a project, and language supervisors, who can view subtitles matching their assigned language proficiency.
- For linguists or reviewers to access the Voiceover Editor, a Superuser must enable "All users can format captions" in Team Settings > Roles & Permissions. Once enabled, they can be assigned a captionset as usual and access the Voiceover Editor from the Display options dropdown.
