Voiceover QA FAQ & Help

Supported workflow

Use one matching script and single-speaker MP3, WAV, or M4A file. Production limits are 100 MB and 30 minutes. English is primary; Japanese post-generation comparison is beta. Output includes timestamped wording differences, long silence, clipping, level changes, and a CSV repair list.

Recoverable failures

  • Empty script or file: add a non-empty source pair.
  • Unsupported or invalid format: re-export as MP3, WAV, or M4A.
  • Oversized or overlong audio: compress or split the file.
  • Provider failure: wait and retry; failed-job reserved time is released.
  • Browser timeout: reopen the same server job from Recent analyses instead of uploading a duplicate.
  • Insufficient checked time: upload a shorter file or buy one credit pack.
  • Payment failure: start one new checkout; contact human support if PayPal shows a completed charge but the balance does not change.

Sources and boundaries

The configured speech-to-text provider is documented by Volcengine. The payment flow follows the official PayPal Orders API and PayPal webhook documentation.

Provider transcripts may normalize or omit speech. Alignment and signal rules are review heuristics, not guarantees of script accuracy, native pronunciation, voice quality, publication readiness, approval, or business outcomes.

Open Voiceover QA · Email human support