First Experience with YouTube Caption Uploads

Yesterday, Google put out a post about the new beta caption upload feature in YouTube. This morning, I decided to put it to the test with all of the videos that we’ve uploaded to our USGS YouTube Channels.

Overall, I’m pretty impressed with how the captions have been synced. The files I uploaded were text-only transcripts. In most cases they were transcripts that had a lot of whitespaces and paragraph breaks. They weren’t just a long line of text. That’s a good thing because I didn’t have to worry about formatting my text to make it nice and need. It appears that the caption upload ignores all the whitespacing except for what’s in between the words as regular single spaces.

One thing that helps, obviously, is a clear speaker. Almost every video had a good recording volume. What surprised me is how well the conversion process ignored background noise. One video on earthquake preparedness contains a lot of Hollywood type music and explosions. The captions turned out nicely considering all the sounds. It also seemed to do well with accents. A couple videos had a woman speaking with a thick (not really thick) accent and the captions, again, synced up nicely.

The syncing process also appeared to be relatively fast.

I’m looking forward to when Google opens up the auto-captioning for all videos that get uploaded. Considering my experience so far with the current options, it’ll be interesting to see how much a time saver the auto-captions provide.

Oh, and once uploaded, you can download your captions in the Subversion format if you care to use them in other applications.

Note: This post is of my own personal opinion and does not constitute and endorsement of any product mentioned by any Federal, State, or Local government organization.