A few weeks ago, YouTube unveiled a new feature that allows you to upload a transcript of your video and have YouTube automatically analyze it to figure out how to turn it into timed captions. I had the opportunity to test this feature (which is still in beta, and only available in English right now), and I have to say I am extremely impressed with the results. I had a handful of Powerpoint presentations that I converted to videos using Camtasia Studio. I then added the narration tracks and produced the videos as MP4 files.
I uploaded the new videos to YouTube, adjusted my settings accordingly and then went to the “Captions and Subtitles” tab for each video. To my surprise, I noticed that the “transcript” feature was now available, so I decided to test it. I opened the script for each narrated presentation, saved it in txt format (for best results, if you have your transcript saved as a Word document, save it in MS-DOS txt format and enable character substitution so that it automatically replaces the fancy quotes, dashes, etc. with the standard versions) and decided to test the transcript feature.
I chose the appropriate transcript for each video, clicked the “Transcript” radio button and told it to upload the file. Unfortunately, the first time I did that, I was told that a file called “English” already exists. So, I gave my transcript a name (even though YouTube says that step is optional) and tried again. This time, it started processing the transcript and went on its merry way.
Once it was finished processing, I watched the video and was pleasantly surprised to see just how accurately YouTube was able to figure out the appropriate timing for the script. Some of the transcripts were rather inaccurate, but YouTube’s processor was able to handle that, guessing at the timing for the inaccurate portions, and then picking up exactly where it should when the transcript became accurate again.
Another great new feature (as part of this update, I’m guessing) is that you can now download copies of the captions for your videos. YouTube automatically converts your plain text transcript into a YouTube subtitle file and allows you to download that file for use with other video players. You can also edit the captions within YouTube, now, fixing any potential errors or timing issues without having to reformat everything again.
YouTube will also be unveiling (they may have already, but I didn’t see it) an automatic transcription feature that will use the voice transcription algorithm from Google Voice to automatically transcribe your video into captions. You won’t need to upload captions or a transcript at all; YouTube will simply listen to the video and attempt to create a transcript for it.
These are great steps forward for accessibility on YouTube, and I look forward to using them more heavily in the future.