The quest to get access to captioned videos online has gotten a bit easier. Google has added automatic-captioning capability to YouTube videos. Ken Harrenstien, a Deaf software engineer, led the captioning project, but has insisted that it represents a team effort, and has praised Google for its Deaf-welcoming environment, its actively soliciting input from its staff, and for striving to make the Internet accessible to all users.
Harrenstien's from Kansas. He has a long, impressive "techie" background, having worked with several renowned Internet pioneers such as Vinton G. Cerf, from the early days of MIT's AI computer lab, SRI, and the ARPANET. The term "hackers" was originally applied to these pioneers—a word that has, sadly, changed meanings with time. It originally meant a skilled, imaginative person, a computer expert who could solve problems by improvising and make connections where none had previously existed. Nowadays it refers to a criminal who breaks into other people's networks, destroys data, and spreads viruses.
Google demonstrated its new captioning capability at its Washington, D.C. branch on November 19, 2009. Cerf, Chief Internet Evangelist at Google, gave a delightful introduction to Harrenstien's demo. Harrenstien noted that while 20 hours of video stream are uploaded to YouTube every minute, "the majority of user-generated video content online is still inaccessible to people like me." Only a handful of videos have had captions added to them by their downers before they're downloaded.
The new captioning capability uses Automatic Speech Recognition technology, which has been around for a while. The new development is "auto-timing," which uses ASR to generate an editable transcript, then synchronizes it to the narrative or dialogue. The caption-generating isn't perfect, but the transcript can be edited to eliminate goofs. This auto-timing feature makes it easier for downloaders to add captions to their videos.
The automatically-generated captions are currently available in English on just over a dozen educational "partner channels" on YouTube, but they can be automatically translated into any of the 51 languages supported by Google. Availability will expand as the technology is refined. The ultimate goal is universal access—to have all YouTube videos caption-accessible.
Of the the new auto-captioning, auto-timing system, Harrenstien said, "This is something that I have dreamt of for many years. To see it happening is amazing."
Just a word: Thanks, Ken, and the other members of the team!