Customize translation projects
Use the customization stage to turn the initial project into the delivery you want. Some choices are made before the first translation run, and others are refined while you review the project page.What changes before and after translation
| Stage | What you usually decide there |
|---|---|
| Before translation | Transcript quality, voice strategy, background audio, subtitles, lip-sync plan, glossary context, and publishing setup |
| After translation | Segment review, SmartSync output quality, manual keeps or regenerations, exports, and publishing |
Recommended review flow
Check the transcript and speakers
Confirm that the transcription is usable before you generate the translated output. Review any transcription issue warnings, speaker
labels, and segment boundaries first. If the source structure is wrong, later steps are harder to correct cleanly.
Open dubbing settings right after project creation
This is usually the best next step after project creation. Set the background audio behavior, subtitle behavior, voice isolation method, voice
mode, and translation instructions before you commit to a full translated run.
Choose and preview voices
Match the delivery to your goal. Voice cloning keeps the original identity closer, while voice library and custom voice flows give you more control over consistency and reuse. Use available voice previews to compare options before you commit to a final run.
Set output options
Decide on subtitles, background audio handling, Smart Publish, dynamic subtitles, and any lip sync work before you finalize the run.
Launch and review the translation run
Once the settings are in place, VoiceCheap generates the translated text, dubbed audio, and related assets for review. SmartSync is part of this output, so you review its result instead of treating it as a separate export step.
Transcription issue analysis
After project creation, VoiceCheap can flag original transcript segments that look suspicious and may need a human review before translation. The analysis uses the transcript, project context, surrounding segments, and confidence signals to highlight places where a speech-to-text mistake could affect the meaning. These warnings are review hints, not automatic corrections. A segment without a warning can still contain an error, and a flagged segment should be checked against the source audio before you change it.| What you see | How to use it |
|---|---|
| Highlighted original segment | Prioritize this segment while proofreading the source transcript |
| Warning icon | Open the analysis details for the possible issue |
| Explanation or suggestion | Compare it with the source audio, then edit the transcript only when the suggestion is correct |
- Open each flagged original segment before launching translation.
- Listen to the source audio around the segment.
- Edit the transcript if the warning points to a real mistake.
- Use Replace all occurrences when the same transcription mistake appears many times.
Important distinctions
- Brand vocabulary helps source transcription. Glossary rules help preserve translation wording across projects.
- Transcription issue analysis helps you decide what to proofread. It does not replace listening to important or unclear segments.
- SmartSync is part of the translated output. You review or adjust that result after generation.
- Lip sync is optional and usually makes the most sense only after the voice and timing already sound natural.
- Team guests can collaborate inside projects, but some team-level settings, such as glossary management or billing, stay owner-controlled.
What you can export after review
- Original imported video when a downloadable source file is available
- Translated video
- Dubbed audio
- Original or translated SRT files
- JSON exports
- Hosted output links
- Lip sync output when generated
- Dynamic subtitles output when generated
- Publication to connected channels when Smart Publish is configured