International Captioning Project
Time-of-air Captioning
Real-time captioning refers to captions which are created during live programs, such as news, sports or special live broadcasts. Real-time captions most often are displayed as "roll-up" captions, so named because of the way they appear on the television screen: a maximum of four lines of text may be visible at one time in a continually scrolling display of text, most often (but not exclusively) placed in the lower third of the picture. (Currently, real-time captioning is limited to the English language only; however, some experimentation with French real-time captioning has been occurred in Canada. See page 24.)
True real-time captioning takes advantage of technology developed for court reporting, called "computer-aided real-time translation." A specially trained stenographer listens to a program while it airs, and types on an electronic stenotype keyboard. Machine shorthand is a phonetic system; each key or combination of keys can stand for either a particular sound combination or phoneme, or can be an abbreviation for a phrase or proper name. Using machine shorthand, a qualified stenocaptioner can do what no ordinary typist could do: keep up with the fastest-talking news reporters, who often speak at speeds upwards of 250 words per minute. Thus, by nature, stenographically generated real-time captions are virtually verbatim.
The data output of the stenotype keyboard is connected to a PC running translation software. The translation software searches large dictionary files, which are customized to the individual stenographer, and which contain stenographic outlines and the text they represent. When the particular stenotype stroke is found in the dictionary, the software generates the text, adding captioning codes. This caption data is sent to an encoding device.
As with off-line encoding, program video is routed through the encoder, which inserts the caption data into the vertical blanking interval on line 21 of the broadcast signal. The stenocaptioner need not be on-site in order to send data to the encoding device: data can be delivered either by modem from a remote location or via a direct connection. What is important, however, is that the stenocaptioner have access to both video and audio during the entire broadcast.
Once caption data have been inserted into the broadcast signal, the video is then broadcast through the usual over-the-air, satellite or cable transmission means. Two or three seconds after the audio is transcribed by the stenocaptioner, the caption data arrive at the viewer's home, where a decoder turns the caption data into text which appears on the bottom of the television screen.
