sxsphinxstack

Skills / Working / Localize a video

Working Video & audio skill

Localize a video

Localize video through translated meaning, timing, voice, captions, graphics, metadata, and cultural context.

Preserve intended meaning and audience experience rather than translating isolated words.

Procedure

  1. Define target locale, audience, distribution, dialect, accessibility, rights, and approval owner.
  2. Lock and transcribe the source with timecodes, speakers, terminology, and on-screen text.
  3. Create a context-rich translation with glossary, names, numbers, humor, idioms, and disputed terms.
  4. Choose subtitles, dubbing, voice-over, graphics replacement, or a combination.
  5. Adapt line length, timing, lip constraints, reading speed, layout, and safe areas.
  6. Record and mix voices with consent, casting, pronunciation, and performance direction.
  7. Replace graphics, dates, units, URLs, legal text, credits, and platform metadata.
  8. Review linguistically, culturally, technically, and in the final player.
  9. Deliver locale-labeled masters, captions, transcript, and a localization report.

Guardrails

  • Never imitate a person's voice without clear authority and disclosure.
  • Do not erase culturally meaningful context to force literal timing.
  • Keep source and localized claims aligned and fact-checked.