Coming to You Live, Accurate Realtime Captioning is Difficult to Achieve

You see it nearly every day. You’ve settled in comfortably with a TV program only to have the screen flash red with a banner saying “breaking news,” followed by a newsperson with an urgent expression explaining the latest crisis. They’ve got your attention. Now what? All is well and good if you’re a native English speaker without any hearing deficits. But if you require the used of closed captioning services on your television to stay informed of the latest events (and even to understand your nightly news broadcast or favorite sporting events), you may be sadly out of luck. With a growing hard-of-hearing population in the United States, as well as skyrocketing numbers of citizens who speak English as a second language, the importance of accessibility and transparency in closed captioning for broadcast media is clear, so much so that United States law now requires that closed captioning services be accessible.

Providing closed captioning services is a job for trained and skilled labor. Nowhere is this more apparent than in the world of live, real-time captioning, particularly in the case of news broadcasting and public affairs programming. Try it for yourself. Watch any live programming and hit the CC button on your remote. When a single speaker is speaking at a moderate pace and enunciating carefully, the captioning is often nearly perfect. But see what happens when multiple speakers are arguing over an issue. For a real train wreck, trying watching a live sporting event with closed captioning services. Not only is the play-by-play too fast for the operator (whether human or robot), but wait until you see the hash they make out of many of the players’ names.

Simply put, captioning or subtitling broadcasts in real time is extraordinarily difficult and leaves no margin for error. Speech-recognition software is used for live captioning and has proven to be a very poor solution. Captioning is performed as well by humans, but even with great skill and specialized equipment, real-time captioning remains a tough nut to crack. Some broadcasters even use “respeakers,” individuals who listen to the broadcast and simultaneously repeat or paraphrase the information to a transcriber or to voice recognition software.