Azure Video Translator API: How to specify voiceName in initial request to avoid wrong accent and double billing?

We are building an AI video localization platform using Azure Video Translator and have hit a critical blocker that makes the service impractical for most non-English/non-Mandarin languages.

Problem Description

In the initial POST /video-translations request, the API deliberately blocks the voiceName parameter. This forces the first translation generation to use a default voice (often English or Mandarin), which is almost always incorrect for our target locales like Cantonese (yue-HK), Japanese (ja-JP), Korean (ko-KR), and Thai (th-TH).

Then do any one knows How to specify voiceName in initial request to avoid wrong accent and double billing?