I was struck by the profile of Karla Ray in last week’s New York Times, where she describes the nuance of her work as a real-time broadcast captioner for Vitac a world leader in captioning and subtitles. The ability of people like Ms. Ray to accurately convey the meaning and context of the audio associated with video will likely never be matched by speech-to-text audio processing, though we are working every day to improve.
Broadcast monitoring in the U.S. relies primarily though not exclusively on the closed captions created by people like Ms. Ray and firms like Vitac. We index this text similarly to how Google or Bing indexes the text on the Web, and use it to make the spoken word searchable within our TV and radio monitoring and search application Media Monitoring Suite. Where captions don’t exist, we create search text via speech-to-text software.
Ultimately when it comes to radio and TV broadcasts, caption text alone, as good as it is, is not a perfect substitute for the way TVEyes delivers a record because while it may report what was said, it will fail to capture how it was said in terms of a speaker’s cadence, demeanor and tone. These variables are also vitally important to understanding a speaker’s true opinion and belief and can only be captured by aligning a text search with the actual video and or audio stream.
Closed captioning was not created to aid media monitoring, but to make TV accessible to those among us who have hearing deficits. The U.S. and Canada lead the rest of the world by providing nearly all TV content with captions. In Europe, the U.K, France, Holland and Nordic countries lead with respect to implementation of closed captions. Many others have little or no captioning in place. In 2011, the European Federation of Hard of Hearing People expressed its vision that 100 percent of public European TV be captioned by 2020. In Asia, South America, the Middle East and Africa, closed captioning is far less prevalent.
The global lack of closed captioning is what motivated us to create TVEyes Language Technologies this year. Our vision is to make global TV broadcasts searchable as easily as the Web is for text. Through TLT, we are advancing the effectiveness and accuracy of speech-to-text as well as adding new language models. This year we will add at least six languages and corresponding international markets for broadcast monitoring, and we’re open to special requests, most of which we can complete in three months or less.
Will we ever advance speech-to-text to be as accurate and context-aware as Ms. Ray and those like her who through closed captions make TV accessible? I'm not sure… but I am certain that we will continue to improve and apply speech-to-text to extend our lead in global TV searching and alerting.
If you'd like more information, or a free trial of the leading TV and radio broadcast search, alert and monitoring solution for PC, Mac, iPhone and iPad, please call 203-254-3600 x100 or email email@example.com.