PC Magazine – March 4, 2010
————
–
YouTube Makes Captioning Available to All
by Mark Hachman
Google’s YouTube on Thursday announced that it has moved its automatic
speech-recognition and closed-captioning technology out of beta and have now
made it available to the YouTube community at large.
Most, if not all, YouTube videos now include a “CC” button that, if pressed,
will automatically generate the closed-captioning technology. The technology
processes the audio feed, using the speech-recognition technology used in
the core voice search feature that has also built into the Android voice
search feature, the GOOG-411 phone search, and other products.
Twenty hours of videos are uploaded to YouTube every minute; Google could
not provide a timetable on when all of the videos would be captioned, but it
will be “very soon,” executives said.
Google rolled out the closed-captioning technology in beta form back in
November, along with a technology to help synchronize the captions to the
actual video.
The technology still isn’t perfect, but Google is also using YouTube, with
its vast array of accents, languages, background noise, and other
distractions to improve its technology.
But the closed-captioning also means that millions of users who are hard of
hearing, deaf, or simply speak another language as their primary language
can now have a greater access to YouTube videos, the company said. “A core
part of YouTube’s DNA is access to content,” said Hunter Walk, the product
team lead for YouTube. “From day one, that’s what we were hoping to do with
video.”
At the present, YouTube can only “understand” spoken English to
auto-generate the caption track. However, Google’s translation services can
also take the English machine-transcription track and translate it to up to
fifty languages. Adding more languages that Google understands “is a
priority,” said Mike Cohen, a speech technologist at Google.
“This is huge,” said Ken Harrenstien, a software engineer at Google and one
of the project’s leaders. “You can go to any video online and now you can
see some captions. It’s not perfect; we know it’s not perfect. Sometimes
it’s funny. But I love it.”
Users can also edit and assist in the transcription. Any text editor or any
tool that’s used to edit caption files can edit the machine transcription,
Google executives said, and then re-upload it for others to use.
On older videos, a button may appear that allows users to request a machine
transcription of the audio. When pressed, an automated request is generated
to the process that generates the transcription. With enough requests, the
video will be pushed upward in the queue. Google executives said that a
typical request may take a few hours to a day or two to be transcribed. As
the process improves, already-transcribed videos may be re-transcribed,
Cohen said.
Google has also partnered with a number of content partners, including UC
Berkeley, Stanford, Duke University, MIT, PBS, National Geographics, and
Demand Media.
Editor’s Note: This story was updated at 11:33 AM PT with additional
details.
