5

I figured I'd leave a question here while I did the manual labor, but what do you find is the best way to transcribe footage?

I'm an editor and the first thing I like to do is create a script of people said, then piece together parts of their takes, and THEN get to the actual footage editing. That way I'm only working with what I need ahead of time.

Basically what I do now is load up the videos in VLC, and open a TextEdit (mac), then watch it through normally or double speed until I get to a part with the interview, then I mark that timecode on the TextEdit, and go watch through the interview in triple-quadruple slowdown, so that I can keep up with what they say and not have to constantly stop and start it back up.

With this way it seems like I'm doing somewhere around 5 minutes real time to transcribe a minute, which comes to about 12 video mins a realtime hour. And I was just wondering if there was a better way you guys already know about. I'm not sure I trust auto-transcribe-type things.

Do you think this slowing the vid down and having to go back 5 seconds only a few times when I can't quite tell what he's saying because it's so slow, is saving more time than watching it in real time, pausing, writing, going back and watching the next part?

Or do you know better ways? Seems transcribing is the most time-consuming part of my workflow.

Since VLC is opensource, would it be possible to combine VLC and a Text Editor...something like VLC on the top and the text editor on the bottom, and whenever you hit a certain key, it automatically puts the timecode in the text, and you can use VLC's shortcuts while typing? Etc?

Again, non-automatic preferred. For one, it could be wrong, and two, I'd like to get used to the footage.

Thanks

Please help?

3 Answers 3

4
+50

Speech Detection with Adobe Premiere

You haven't mentioned what NLE you use to edit, however here is a method you can use to transcribe if you are using Adobe CS4 or later.

This method uses speech detection to automatically transcribe videos - a feature brought in with CS4. It then adds the the text into the metadata of the file.

Analyze speech to create text metadata

  1. Select a file or clip.
  2. At the bottom of the Metadata panel, click Analyze Speech, or Analyze (Adobe Premiere Pro).
  3. Set the Language and Quality options, and select Identify Speakers if you want to create separate speech metadata for each person. Note: Speech Search can use any of several language-specific and dialect-specific libraries, such as libraries for Spanish and UK English.
  4. Click OK.
  5. The spoken words appear in the Speech Analysis section.

To retain the speech metadata, save the project.

Obviously this depends on the quality of the audio. Good clean audio will provide the best results. Its also worth noting that Premiere uses Soundbooth to process the detection, so if your audio wasn't up to scratch you could clean it up in Soundbooth, and then perform the analysis there.


VLC Scripting

I came across this article explaining how to take the current timecode from a video and add it to a text document for transcribing. The title of the article states its a method for Linux. However I'm sure it would be easy enough to take the method and apply it to Windows or OSX.


Manual

In regards to your current method, it depends on how many times you fall behind. When I have transcribed video I have done it at full speed but I haven't bothered to fix errors while I'm typing. Usually I'll go back and fix it up later - and being that they are just for editing purposes they don't need to be totally correct.

Manual live tagging with Adobe Prelude

You might also want to have a look at Adobe Prelude CS6. It's an ingesting and meta logging program. One of its features is that it lets you add markers (and name them) in realtime as you are watching the video.

1
  • Hey thanks OrangeBox I gave you the bounty. I'll check out the VLC scripting article thanks so much! I use FCP to edit. And I have no idea how you can keep up with full-speed speech...I can barely keep up with half-speed... Commented Aug 3, 2012 at 20:53
1

Well, no need to worry about. Transcribing footage is something very easy now-these-days. What you have to make sure is that just write down all your audio in chronological order and he time codes of when each section begins and ends. Try to include punctuation and how phrases are expressed, so your are as close to the video as possible. If you are seriously very much keen, then you should go ahead and try to use the real transcription symbols, which can indicate pauses and overlaps in speech.

1

oTranscribe Web App

I've tried using oTranscribe with some success. It let's you input a local media file or a YouTube link. And it also works off-line. (!)

For a free web app I think it's quite decent:

  • Hotkey for timestamp insertion: CTRL+J
    • Clicking a timestamp jumps to that moment.
  • Hotkey for rewind: F1 (or CTRL+1)
  • Hotkey for fast-forward: F2 (or CTRL+2)
  • Hotkey for slow down: F3 (or Ctrl+3)
  • Hotkey for speed up: F4 (or Ctrl+4)
  • Hotkey for pause: ESC

However:

  • No way to enlarge the video decently (and not enlarge the transcription text at the same time). (I've tried to transcribe conference video feeds, and the projected slides are not really readable anymore.)
  • No way to un-letterbox videos. (Again with the conference videos. This conference video for example is absurdly letterboxed: top, bottom, left and right. Try to read anything there.)
  • No current timecode visible while playing. Needs mouseover action to fade in.
    • Also: Rewind via F1 will not show you what time you have jumped to.

So, while not perfect, it's way, way better than the tab-back-and-forth-between-windows approach.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.