Translate subtitles before editing: workflow?

[Jim Newberry]

I've got a small doc project with people speaking Thai, which I don't understand. I have someone in Thailand to translate for me. I can't edit until I get the translation. I'm trying to figure out a good workflow to post the unedited clips (Youtube, Vimeo, or wherever) to be translated, have English subtitles added, and view the subtitles in FCP X for editing and finishing.

I've heard of Amara.org which seems like it might help, but I've never used it (I've read that it works with Vimeo, and YT also I think). I only have about 5 clips, a few minutes each, not a ton of stuff. I know I could just have the translator use timecode and a text file, but I was wondering if there was a higher tech way, where he could add the text to the video clip without permanently burning it in...but maybe I'm overthinking things.

Also, I added source timecode windows to my clips, but when I add them to the timeline the windows show the program/timeline timecode...maybe it would be better to upload the clips independently to avoid timecode confusion.

[Bill Davis]

I'm confused, but it's probably just me. ; )

So the actualities on the video are spoken in Thai - you want to send those video clips to Vimeo - then you want the translator to download them, do the English VOs - then provide you with What?

Sidecar files in enghlish with timing cues to the video? or proxy video clips with english VOs? Or final resolution video files with the english VO in place?

Do you need to maintain the original Thai soundtrack and do an english version on another track?

Again, it's probably just my brain not understudying this.

But at the basic level, one of the magical things a bout X is that you can work in discrete blocks that are synced by the magnetic timeline as standalone elements. So you can conceivably do each "speech" as a separate unit, keep the Thai vocal embedded in the video. Lay the English version as a connected clip synced to that video PHRASE only. And build the entire project in snippets.

This way the timing of the Thai against the English can "float" using gap clips to extend the video - and allowing you to patch or extend the scenes via connected B-roll if the timing of the Thai and the English are too far off to match well.

Presuming I'm understanding any of this!

: )

[Jim Newberry]

So the actualities on the video are spoken in Thai - you want to send those video clips to Vimeo - then you want the translator to download them, do the English VOs - then provide you with What?

Yes, the people in the video are speaking Thai. I want to upload the clips to a private page on Vimeo or Youtube (or wherever) for translation (translator doesn't necessarily have to download clips, if there's a way to annotate at Vimeo or YT or Amara). I would like to see the English translations as text, so I can add titles in FCP X. I realize the translator could just type timecode and translations into a text file, but I was wondering if there was something where he could annotate the video with one of the services I mentioned, so the text would be wed/synced with the video; IOW, a more automated, user-friendly way to create a text or XML file with timecode that I can import into FCP.

Sidecar files in english with timing cues to the video? or proxy video clips with english VOs? Or final resolution video files with the english VO in place?

Do you need to maintain the original Thai soundtrack and do an english version on another track?

I don't know--I've never done this before...whatever file type etc that would allow me to add titles syncronised with the interviews. Not final res files; I won't even begin to edit until I get the translations. I want to keep the Thai audio; I'm not overdubbing, just want to show English subtitles. (No VO's).

But at the basic level, one of the magical things a bout X is that you can work in discrete blocks that are synced by the magnetic timeline as standalone elements. So you can conceivably do each "speech" as a separate unit, keep the Thai vocal embedded in the video. Lay the English version as a connected clip synced to that video PHRASE only. And build the entire project in snippets.

This way the timing of the Thai against the English can "float" using gap clips to extend the video - and allowing you to patch or extend the scenes via connected B-roll if the timing of the Thai and the English are too far off to match well.

Good to know, thanks.

[Jim Newberry]

Update: Looks like Amara - Caption, translate, subtitle and transcribe video. should take care of this for me, though I haven't tried exporting the subtitles yet; but they can be exported to SRT and I found a plugin that will import that into FCP X. If anyone's curious, there's a video at that URL above that shows how Amara works.