Google announced last week that they are introducing an automatic captioning facility for YouTube videos. In the official Google blog they state that every minute 20 hours of video are uploaded. The chances of each individual video owner taking the time and effort to caption each of their videos is unlikely. It means that most user generated content on the video sharing site is inaccessible to people who are Deaf or hearing impaired.
It uses their own automatic speech recognition (ASR) technology combined with the existing YouTube captioning system to automatically produce captions – auto-caps for short. It is initially only available on a number of mainly educational channels. Although ASR has been around for a while in the GoogleVoice application it has never been deployed on such a large-scale before.
An additional facility that has also been introduced is automatic timing. If a video owner already has a transcript, they can upload this as a text file with the video and the software will automatically link the words in the transcript with the sounds on the video.
Both of these features are not only suitable for viewers with hearing difficulties but make content available in more languages. The facility is only available in English but using the Google translation tool they can be translated into a number of languages.
Obviously this is not an exact science and mistakes are bound to occur but even with the problems, auto-caps is a major step forward in making all videos accessible.