Google Inc. said Thursday it is introducing automatic, machine-generated captions for videos on its YouTube site. The new service, being launched this week, is intended to make online videos accessible to the deaf and hearing-impaired.
Hundreds of thousands of videos on Google sites already contain caption tracks that users have created and added manually with Google's existing captioning service. But with 20 hours of video uploaded to YouTube every minute, most videos on the site still lack captions.
So Google is tapping into the speech-recognition technology that it uses for its Google Voice call management service to make captions an automatic feature on YouTube.
Because the speech-recognition technology is still a work in progress, Google is launching the automatic captioning service on the YouTube channels of just a handful of partners, including PBS, National Geographic and a few big universities. But the company promises that the technology will improve over time - and it hopes for a much broader rollout.
In the meantime, Google is adding a new "auto-timing" feature to its existing manual captioning service to make it easier to use. Video creators will now simply have to create a text file with all the words spoken in a video and Google's speech recognition technology will take it from there - matching the text to the words as they are spoken. Google hopes this will encourage more users to add captions to their videos.