Video Editing & Generation — Overview

Video Editing & Generation takes a long-form (around two hours) English training video and produces Arabic versions for DJI's audience. It:

1. Transcribes the spoken English.
2. Translates it into Arabic.
3. Lets a person review and correct the Arabic.
4. Produces two outputs from the reviewed text:
   • a subtitled video, with the Arabic text on screen, and
   • a dubbed video, with native Arabic voices speaking the translation.

Why a human reviews it

Legal training has to be accurate. The machine produces a first-pass transcript and translation, but a reviewer corrects the wording and timing before anything is published — and the dub then speaks exactly the reviewed text, so the spoken Arabic matches what was approved.

Who it's for

DJI administrators and the content team preparing Arabic training material from English source videos.

A note on cost

Dubbing is the most expensive step (roughly the price of the audio length), so the system shows a cost estimate before a dub is started. Subtitling is far cheaper.