The captions will be machine generated and automatically timed to match the video (and still be readable, I hope). Just like any dictation program, I suppose, they’ll listen to the audio from the clips and transcribe it. Now, normally, you know I’d be in there making fun of all the hilarious errors a machine makes when transcribing the participating partners’ clips (because, hey, even humans don’t always do a good job of captioning)—but as I write this, they’re actually not doing it yet. Sigh.
Google rightly points out that this will enhance search within YouTube. Previously, only the accompanying text of a video would be searchable; now the actual content of the video would be indexable as well.
Google has actually been indexing the content of selected videos’ audio for over a year through its GAudi project. However, even after the elections of last year, they continue to focus on political news only in that project.
Also, the new automatic caption timing feature will make it easier to manually caption videos, for nonparticipating partners. You know, the rest of us.
What do you think? Are you excited for captions on videos from partners like Duke (IN YOUR FACE, FRANK AND ANDY—welcome to basketball season!)?